**Implementing Lung Cancer Screening and Prevention in Academic Centers, A**ffi**liated Network O**ffi**ces and Collaborating Care Sites**

**Cary A. Presant 1,2,\*, Ravi Salgia 1,2, Prakash Kulkarni 1,2, Brian L. Tiep 1,2, Shamel Sanani 1,2, Benjamin Leach 1,2, Kimlin Ashing 1,2, Jossie Sandoval 1,2, Mina S. Sedrak 1,2, Shana Landau 1,2, Sophia Yeung 1,2, Dan Raz 1,2 and Shanmugga Subbiah 1,2**


Received: 18 April 2020; Accepted: 9 June 2020; Published: 11 June 2020

**Abstract:** Lung cancer is one of the deadliest and yet largely preventable neoplasms. Smoking cessation and lung cancer screening are effective yet underutilized lung cancer interventions. City of Hope Medical Center, a National Cancer Institute (NCI)- designated comprehensive cancer center, has 27 community cancer centers and has prioritized tobacco control and lung cancer screening throughout its network. Despite challenges, we are implementing and monitoring the City of Hope Tobacco Control Initiative including (1) a Planning and Implementation Committee; (2) integration of IT, e.g., medical records and clinician notification/prompts to facilitate screening, cessation referral, and digital health, e.g., telehealth and social media; (3) clinician training and endorsing national guidelines; (4) providing clinical champions at all sites for site leadership; (5) Coverage and Payment reform and aids to facilitate patient access and reduce cost barriers; (6) increasing tobacco exposure screening for all patients; (7) smoking cessation intervention and evaluation—patient-centered recommendations for smoking cessation for all current and recent quitters along with including QuitLine referral for current smokers and smoking care-givers; and (8) establishing a Tobacco Registry for advancing science and discoveries including team science for basic, translation and clinical studies. These strategies are intended to inform screening, prevention and treatment research and patient-centered care.

**Keywords:** cancer center; lung cancer; lung cancer screening; low-dose CT scans; cancer prevention; smoking cessation; tobacco control; national guidelines for screening and prevention; pharmaceutical aids to smoking cessation

#### **1. Introduction**

Lung cancer is the leading cause of cancer-related deaths. However, lung cancer is one of the most preventable human malignancies. A screening program is able to detect lung cancer at an earlier and more successfully treatable stage, and thus improves survival. In order to control lung cancer incidence, morbidity and mortality, it is important for a health care delivery system to provide and monitor screening and prevention programs including tobacco cessation. In this report, we will review the experience of the City of Hope Medical Center at its academic campus in Duarte, CA, as well as its 27 community practice sites throughout much of Southern California. This is an observational descriptive study, and broader results of our experience are under study and will be published separately.

#### *1.1. City of Hope's Experience in Lung Cancer Screening and Prevention*

City of Hope is an NCI-designated comprehensive cancer center and a member of the National Comprehensive Cancer Network (NCCN) network. City of Hope utilizes the NCCN guidelines for lung cancer screening and prevention. In addition, all clinical sites (the academic center in Duarte and all 27 community centers) follow the NCCN lung cancer care and screening guidelines and comply with Via Oncology Pathways (as modified by City of Hope) for evaluation, anti-tumor treatment and surveillance after treatments. A map of these sites indicates the broad coverage of communities in the greater Los Angeles area (Figure 1). Compliance with these pathways is monitored for every physician and advanced practice practitioner (APP). Pathway compliance is very strongly encouraged. We follow the NCCN guidelines on smoking cessation, since smoking cessation is conceptually an integral component of cancer care.

**Figure 1.** Map of City of Hope Locations in Southern California. Blue Square, Duarte National Medical Center Academic Site; Blue Teardrops, Community Cancer Center Practice Sites; Orange Triangles, City of Hope Oncologist-Staffed Community Hospitals; Yellow Square, Orange County Campus (under construction). Map designed at City of Hope Medical Center.

Lung cancer team leadership at City of Hope is provided by Dr. Ravi Salgia, Dr. Dan Raz and Dr. Erminia Massarelli. Under their direction as champions, the lung cancer disease team promotes lung cancer screening and prevention in Duarte and in all network community centers. The relationships between these components of the program are pictured in Figure 2. This model following the team leadership with advice of the Department of Population Science allows interactive design of the initiative and modifications, and implementation in Duarte and community sites and a synergistic effect implementing screening and prevention.

**Figure 2.** Lung Cancer Prevention and Screening Model. The City of Hope Lung Cancer Tobacco Control and Cancer Screening Initiative Is Implemented in the above Model. Model was designed by City of Hope Medical Center.

#### *1.2. Academic Center Experience*

At the academic center, City of Hope conducts preclinical laboratory investigations into genetic factors associated with lung cancer. This research helps to define which patients are at risk for developing lung cancer. Genomic alterations are identified which may help in screening and development of drugs targeted for cancer prevention. Population science research further aids in cancer control activities. Clinical trials, developed from these research activities, are initiated at the academic campus in Duarte as well as at one or more of our 27 affiliated network community cancer centers.

At the academic center, cancer program members (67 physicians, 67 APPs, 12 research faculty, 14 graduate students and fellows and 10 additional data analysts, and clinical trial coordinators) and specifically the lung cancer team conduct weekly disease team-based meetings/tumor boards/research seminars and invite participation of physicians at the network community cancer center offices. At the academic center, screening with low-dose CT (LDCT), smoking cessation, and patient education are provided.

At the academic center, 8623 total new patients including 3255 new cancer patients are seen annually (data from 2019), with 252 new lung cancer patients evaluated per year. Racial/ethnic diversity exists, with 61.4% Caucasian, 19.9% Hispanic, 14.7% Asian/Pacific Islander, and 3.9% African-American patients. Overall, 4.2% of all patients at the academic center self-report as current smokers, 27.8% are former smokers, and 61% of patients are never smokers. Medicare is the payer in 37.9%, Medicare Advantage in 1.7%, Medicaid in 9.3%, PPO or commercial in 29.1%, HMO or managed care in 17.9%, and other payers in 3.4% of patients. This patient population is different compared to the 27 community centers, which has implications for screening and tobacco control. Actual practices in tobacco control vary by treatment site. At the academic center, there is a formal smoking cessation program, while there are no formal smoking cessation resources at network community offices. At all

sites, patients are referred to the California Smokers Helpline 800 No Butts program (800-No-Butts) or to the California QuitLine operated out of the University of California San Diego (800 Quit Now).

At the academic center, LDCT screenings were ordered in 434 patients over a 4 year period, and 424 (98%) completed the procedure. We universally use the Lung-RADS system of evaluation of LDCT scans [1]. Of those patients undergoing LDCT, Lung-RADS results showed suspicious lung cancer category 4A results in 3.3% and highly suspicious category 4B results in 1.7%. Of the six patients who had biopsies performed at the academic center, four (67%) had stage 1 lung cancer diagnosed, and two (33%) had stage 4 lung cancer. Other patients were referred back to their primary care physicians for diagnostic biopsies. Long-term follow up of patients diagnosed through our LDCT screening program will determine reductions in risk of death and will be reported in the future. Any cost savings associated with earlier diagnosis depend on the health system, and different health payers are involved with our patients (see above).

One of the issues in lung cancer screening is the excessive use of scans. At City of Hope, Shared Decision-Making Consults and Lung-RADS are some of our tools to address the issue of reducing excessive imaging, false positives, excessive harm as a result of the additional testing, patient anxiety, and cost. During the Shared Decision-Making Consults, ordering clinicians use checklist and visual tools to discuss the risk and benefits with the patients, address patients' concerns, and assist patients to determine whether the lung cancer screening is suitable. The clinician also educates patients on what is low-dose CT, and advises postponing the screening if they have had a chest/PET-CT in the past 12 months. In the Duarte cancer center LCS program, we also coordinate with the attending clinicians to prevent excessive imaging. Lung-RADS is a widely accepted and logical nodule reporting and management tool used by the radiologists to minimize false-positive findings, excessive imaging, and unnecessary procedures. As part of our lung cancer screening initiative, we are developing educational tools and informational aids to help patients understand these issues before screening.

#### *1.3. Community Centers Experience*

Our community cancer centers are staffed with 43 medical oncologists, 40 radiation oncologists, 7 APPs and a clinical trials coordinator. In order to improve lung cancer control, physicians and APPs at the 27 network community cancer centers advise screening and prevention at local tumor boards and medical lectures, and depend on collaborating/referring community primary care physicians, pulmonologists, radiologists and hospitals for lung cancer control. Primary care physicians who are the referring doctors of patients to City of Hope, as well as City of Hope providers themselves (to varying degrees), make assessments of lung cancer risk, order LDCT screening exams, advise or provide smoking cessation to current smokers or recently quit patients, refer patients to other local smoking cessation programs if available, prescribe smoking cessation medications, and provide behavior intervention and counseling with follow-up visits to assess compliance.

In only one-half of the network community practices do the oncologists prescribe medications to assist in smoking cessation, while the other half simply refer the patients back to their primary care physicians with recommendations. In a current survey of community practice sites, almost all of the network community center physicians discuss screening and prevention at tumor boards or hospital meetings where primary physicians and pulmonologists are in attendance, but a few still do not, even today.

At the 27 community centers, 22,000 new patients are seen annually (data from 2019). In total, 19,201 patients are seen by medical oncology/hematology specialists. The incidence of current smokers is 7.2%. Of the patients, 49.4% are Caucasian, 20.3% are Hispanic, 4.5% are African-American, and 6% are Asian/Pacific Islander. Among our patients, 43.7% are insured by government programs, 22.9% are PPO/commercially insured, and 22.7% are insured by HMOs or IPA medical groups.

#### *1.4. Institutional Vision*

Even in an integrated oncology network, there remain challenges with opportunities for improvement. In our vision, there is a focus on team medicine. The academic center uses its knowledge, skills and experience to set standards and create pathways so as to encourage its network community centers to implement the science and evidence base (Table 1). The community center encourages the local collaborating physicians, medical groups, IPAs, and hospitals to utilize their resources to implement the screening and prevention activities as well as providing smoking cessation advice and medications and/or nicotine replacement treatments for some patients. Naturally, these necessitate that the primary care physician, pulmonologist and/or oncologist order the interventions necessary to reduce the risks of lung cancer. The component activities or interventions seen in Table 1 are most appropriately (indicated by +++) taking place at the academic center or network community cancer center.


**Table 1.** Sites for Lung Cancer Screening and Prevention.

To address these, we have developed the City of Hope Tobacco Control Initiative. Elements of the initiative were begun in 2013. As the program elements matured and new information informed needed improvements and coordination, the initiative was revised. The current initiative was formulated in January 2020. It includes (1) a Planning and Implementation Committee; (2) integration of IT including medical records and clinician notification/prompts to facilitate screening, cessation referral; (3) physician and clinician training and supervision that endorses national guidelines and refer and encourage lung cancer screening and smoking cessation interventions; (4) providing a clinical champion at the community sites, and at the academic center for program leadership; (5) Coverage and Payment reform and aide to facilitate patient access and utilization and reduce cost barriers; (6) implementing systemic tobacco exposure screen for new and existing patients; (7) smoking cessation intervention and evaluation—patient-centered recommendation for smoking cessation for all current and recent quitters along with QuitLine referral for all current smokers and the smoking care-givers; and (8) establishing a Tobacco Registry for advancing science and discoveries including team science for basic, translation and clinical studies.

The implementation of the initiative is in progress at Duarte and in community sites, as led by the lung cancer disease team, clinical departments, and the network, regional and individual site leaders, and community center lung cancer champions. Evaluation methods include a review of electronic health records for LDCT screening orders in patients with higher risk based on smoking history and other clinical and epidemiologic factors, referral for smoking cessation, continuing visits for monitoring cessation efforts, and frequency of prescribing anti-smoking medications and/or nicotine replacement therapy medications. Compliance measures are thus multifactorial. At present, we have demonstrated

that all 27 community cancer centers can operate together, in a coordinated well-integrated effort with the academic site.

#### **2. Focus on Lung Cancer Screening**

#### *The Science*

Unfortunately, LDCT is a lung cancer screening method which is underused [2]. The frequency of performing lung cancer screening in patients who are eligible for screening by nationally accepted criteria is 3.9% as of 2015 [2]. However, this may be improving, since a recent study suggests that in a survey of 10 states in 2017, 12.7% of people 55–80 years old were eligible for screening by national criteria, and 12.5% actually had received screening [3]. There was a variation by state ranging from 9.7% to 16%.

As summarized in NCCN guidelines [4], patients who are generally eligible for LDCT screening are those at high risk, defined as 55 to 77 years old, have a 30 or more pack year history of smoking, and are active smokers or have stopped smoking within the last 15 years. Although not defined in the screening trials as high risk, other patients who are also at higher risk are those who are age 50 years or over, with a 20 or more pack year smoking history, and have other risk factors. These risk factors can include prior lung or head and neck cancer, a family history of lung cancer, prior chest radiotherapy, asbestos or radon exposure, presence of HIV or COPD or pulmonary fibrosis, or significant second-hand smoke exposure.

Criteria for evaluation are also nationally standardized as outlined in the NCCN guidelines [4]. They cover not only which findings require immediate biopsy or excision, but also which require close follow up versus those who require a repeat LDCT in 1 year.

LDCT has been shown to detect earlier state lung cancer. In fact, 63% of lung cancers which were found by LDCT were stage IA or IB [5]. Further, this earlier detection has been associated with reduced lung cancer deaths by 20% worldwide [6].

It is important to identify patient symptoms which might be associated with lung cancer. Persistent cough, hemoptysis, shortness of breath, chest pain, weight loss, and/or wheezing may be symptoms of lung cancer. Someone who has any of these symptoms should not get a low-dose CT scan, but they should get a diagnostic CT scan with and without contrast. Payment for diagnostic CT in the presence of symptoms is covered by insurance even in managed care plans. Requests for prior authorization should include diagnostic codes for the specific symptoms. For patients with prior lung cancer, annual diagnostic CT chest scans are indicated, with details included in NCCN guidelines. Clinicians both at Duarte and in community sites are committed to compliance with those guidelines.

If the LDCT is negative, patients should be counseled regarding prevention. This is addressed in the following section. Since the prevalence of LDCT screening is low among patients who meet requirements to receive it, there remains considerable need for medical leadership in local communities to encourage more widespread use.

#### **3. Smoking Prevention**

#### *The Science*

The most prevalent causative factor for lung cancer is smoking—an estimated 85–90% of patients [7]. Thus, smoking cessation is crucial. Smoking cessation programs include a recommendation from a physician to a patient to stop smoking, followed by referral to a smoking cessation program. In order to prevent withdrawal symptoms and facilitate cessation, the physician can prescribe pharmaceuticals such as varenicline, nicotine replacement therapy (NRT), or buproprion with or without NRT [8]. The success of cessation in achieving complete abstinence is poor with a 1 year abstinence rate of 20% using buproprion with NRT, and a 24 week abstinence rate of 26% with varenicline [8]. Behavioral interventions have been shown to improve odds of success over medications alone. Therefore, follow up is most necessary to continue to promote cessation on the individual patient level. Referral to smoking cessation is best performed by the primary care physician or APP, since their relationships with the patient are generally closer than relationships with the oncologists. However, cessation is also often prescribed and facilitated by an oncologist as part of standard cancer care. Further, the oncologist has a role in promoting smoking control at the collaborating hospital or medical staff meetings. The immediate and long-term benefits of smoking cessation include not only better survival, but also critical positive impact on cancer care, improved success of treatment, reduced recurrence of the primary cancer, less frequent development of new cancers, and less progression of comorbid conditions.

If smoking cessation is achieved, lung cancer incidence rate is reduced by 83% in totally abstinent non-smokers, and by 55% in partially abstinent smokers [9]. Recidivism rate, with patients again starting to smoke, are unfortunately high. Continued motivational support for the patient is necessary. Many hospitals including the City of Hope academic center in Duarte have smoking cessation support programs. Alternative programs are provided by such organizations as the American Cancer Society California Division and state of California via phone or internet-based support program called 1-800-No-Butts.

There is high and growing use of E-cigarettes for either nicotine replacement or combustible smoking cessation. The exact impact of E-cigarettes in transitioning from combustible smoking to non-smoking remains controversial. Recent studies cast doubt on the true efficacy of substituting vaping for combustible cigarette smoking. For many individuals who turn to vaping, dual use is common. This leads to an additive effect of toxicities from smoking and vaping, which is counter to the overall goal of tobacco cessation.

City of Hope has helped to develop and adheres to NCCN guidelines on smoking cessation. We do not recommend E-cigarettes or vaping as a smoking cessation or smoking replacement tool either for single use or in combination with FDA-approved smoking cessation medication. If a patient has quit smoking using E-cigarettes, we support continuation of non-smoking status, encourage transition to approved smoking cessation medications, encourage behavior support, and provide support to reduce the risk of lapses or relapses. Of special concerns are the dual use of tobacco and E-cigarettes and E-cigarettes in children where lifetime addiction to nicotine products are of particular concern.

In addition to smoking cessation, other interventions are known to be effective in reducing lung cancer rates and are advised by the National Institutes of Health [10,11]. Patients should avoid second-hand smoke, avoid excessive radiologic scans, reduce exposure to radon (present in 1 of every 15 homes), not be exposed to asbestos (and also chromium, nickel, beryllium, cadmium, tar, uranium, coal products and diesel fuel), and avoid beta carotene use in smokers.

Other interventions to reduce lung cancer risk which have evidence to support them but not yet accepted in national guidelines include eating a Mediterranean-style diet with increased fruits and vegetables, possibly leading to a 33–50% reduction in lung cancer [12]; increasing exercise, with possibly a 25% reduction [13]; or using green tea, with possibly a 22% reduction [14].

A number of nutritional interventions have been recommended without clear evidence of possible improvement. These interventions include red wine, flax seed, garlic, ginger, selenium, turmeric and vitamin D.

#### **4. Translating Screening and Prevention from the Academic Center**

It is difficult to successfully promote lung cancer screening and prevention at the local community since insurance often does not adequately reimburse for these programs or for pharmaceuticals. The evidence base is clear that screening and prevention should be strongly encouraged. Our survey of primary care physicians in our catchment area of greater Los Angeles indicated that these physicians were aware of LDCT screening guidelines—only 12% of primary care physicians in the community (not at academic centers) had referred patients for lung cancer screening in the past year [15]. This indicates a need for community leadership by City of Hope oncologists at local hospital meetings and tumor boards to promote appropriate LDCT screening.

There are many factors that help to promote these programs (Table 2). Guidelines exist for screening and prevention, but they must be more widely accepted by physicians, medical groups, IPAs, HMOs, and hospitals. This must be based on a collaborative relationship, which must be fostered by network community cancer centers and collaborating local physicians and APPs. Promotion can be most effective at tumor boards, at continuing medical education presentations, and even in lunchrooms. Hospitals often need to differentiate themselves from competing institutions, and controlling lung cancer can be an important factor. Having champions in the hospital and on the medical staff can help, and often philanthropy can foster a successful program of smoking cessation. Coordination with voluntary health care organizations can provide educational and patient-support resources not otherwise available in the community.


**Table 2.** Factors in Translating Screening and Prevention Science from Academic Centers to Affiliated Network Community Centers and Collaborating Community Sites.

However, many obstacles exist in successfully controlling lung cancer. Pressure to reduce health care spending is part of the national debate, and motivates lower authorization of screening and prevention efforts at the insurer, IPA, HMO, alternative payment model (APM) and medical group administrative level. Contracts between those organizations and individual physicians and APPs may limit ability to refer for screening and prevention interventions. Even with approval for LDCT screening or smoking cessation, many hospitals do not have resources available. Since it takes time to evaluate lung cancer risk, and order tests and preventive interventions, there is often a lack of reimbursement for those individual patient visits.

Importantly, compliance surveys on which contracts and payments are based only infrequently assess lung cancer screening and prevention activities. There is poor coordination of patient care in the area of lung cancer screening and prevention. Even when a health plan approves screening or smoking cessation, copayments by patients may still be high for those interventions as well as for medications to suppress withdrawal symptoms.

Just to reemphasize, physicians and APPs should always distinguish between screening LDCT, versus the need for diagnostic chest CT. A diagnostic CT is appropriate and indicated if patients have any symptoms that could be associated with lung cancer. Ordering the appropriate test is most important.

Another opportunity for enhancing local community hospital implementation of lung cancer screening is the evolution of algorithms for interpreting LDCT scans. As software and analytics improves, we expect to see new criteria for identifying lung nodules likely to be neoplastic. Although sometimes difficult for a local community radiologist to remain current in updates, the City of Hope Department of Radiology can provide educational updates for local community hospital radiologists to be immediately aware of the new algorithms necessary for earliest diagnosis of lung cancers more amenable to curative interventions.

#### **5. Recommendations for Translating Screening and Prevention into Local Medical Communities**

Although there are obstacles for lung cancer control, advances in the science of lung cancer screening, prevention and treatment can favorably impact cancer care. The success of the City of Hope programs across the entire enterprise of academic center and 27 affiliated network community sites has led us to be able to recommend steps that other institutions can take to improve lung cancer care (Table 3). We recommend adoption of the NCCN guidelines for lung cancer LDCT screening and prevention by physicians, health plans, medical groups, physicians, IPAs, HMOs, APMs and hospitals. To provide leadership, a clinical champion at the community site and also a champion at the academic center are important. The electronic medical record systems used should be updated to separately identify LDCT screening, prevention counseling and smoking cessation referrals and to enable lung cancer risk assessments. Compliance with guidelines should be measured, with reports back to physicians of their success rate or impact. High compliance should be promoted and rewarded. Additionally, the electronic health record can be better programmed to prompt or "nudge" health care providers to implement appropriate cancer screening [16]. Nudge interventions, such as automated alerts directed to the medical team to discuss breast cancer and colorectal cancer screening with patients, showed a dramatic increase in the number of ordered screening tests [17], but it did not affect how likely the patient was to actually complete the test within a year. Once cancer screening is ordered, the patient still has to take the right steps to complete it, such as scheduling an appointment and then going to the appointment. Hence, interventions should test methods to nudge both providers and patients to complete lung cancer screening, while also attempting to eliminate or mitigate some of the hurdles to these tests.


**Table 3.** Recommendations for Improving Integration of Lung Cancer Screening and Prevention in Clinical Networks.

Medical community and public promotion of lung cancer screening and prevention should be implemented. This will only be successful if each of the elements of screening and prevention are covered by insurance with reasonable copayments by patients.

As science improves, there continues to be a need for clinical trials of the possibly beneficial improvements. These trials should be funded by government programs and/or voluntary health care agencies. As leaders in our communities, we should support these trials.

#### **6. Limitations of This Study**

This study is at present observational and descriptive, and results will be determined prospectively with additional implementation of the initiative. The short follow up at present precludes detailed

reporting of results on cessation success, reductions in lung cancer mortality, and cost savings. Results will be analyzed and published in future communications. Comparisons of lung cancer staging results with LDCT compared to staging resulting from diagnosis when symptoms were present will be evaluated when more patients have been diagnosed at Duarte and in the community network.

#### **7. Conclusions**

Lung cancer continues to be a challenge to patients, providers, and health care systems despite dramatic improvements in treatment. We can use successful screening and prevention programs to reduce these life-threatening diagnoses and give patients more hope for the future. Smoking cessation in particular must become an integral component of cancer care and prevention. Reporting on the experience at City of Hope, the recommendations of the translational medical system of City of Hope, and the City of Hope Tobacco Control Initiative is designed to promote broader lung cancer screening and prevention efforts at other institutions and in other communities. We hope our experience will help contribute to improved patient outcomes and increase tobacco control.

**Author Contributions:** Conceptualization, C.A.P., S.S. (Shamel Sanani), B.L., K.A., B.L.T., R.S., P.K., J.S., M.S.S., S.L., and S.S. (Shanmugga Subbiah); Methodology, C.A.P., S.S. (Shamel Sanani), B.L., K.A., B.L.T., R.S., P.K., J.S., M.S.S., S.L., S.Y., D.R., and S.S. (Shanmugga Subbiah); Investigation, C.A.P., S.S. (Shamel Sanani), B.L., K.A., B.L.T., R.S., P.K., J.S., M.S.S., S.L., D.R., and S.S. (Shanmugga Subbiah); Data Evaluation, C.A.P., S.S. (Shamel Sanani), B.L., K.A., B.L.T., R.S., P.K., and S.S. (Shanmugga Subbiah); Writing—Initial Draft, C.A.P.; Writing—Review and Editing, C.A.P., S.S. (Shamel Sanani), B.L., K.A., B.L.T., R.S., P.K., J.S., M.S.S., S.L., S.Y., D.R., and S.S. (Shanmugga Subbiah); Supervision, C.A.P. and R.S.; Project Administration, C.A.P., B.T., R.S., P.K., and S.S. (Shanmugga Subbiah). All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded in part by NIH Grant P30 CA033572

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **List of Abbreviations**

APM—alternative payment model organization; APP—advanced practice provider; COH—City of Hope; COPD chronic obstructive pulmonary disease; HMO—health maintenance organization; IPA—independent practice association; IT—information technology; LDCT—low-dose computed chest tomography; L-RADS—lung imaging reporting and data system; NCI—National Cancer Institute; NCCN—National Comprehensive Cancer Network; NRT—nicotine replacement therapy; PPO—preferred provider organization; Via—Via Oncology Clinical Pathways, now also known as ClinicalPath (Elsevier)

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### **Advancing the Science and Management of Renal Cell Carcinoma: Bridging the Divide between Academic and Community Practices**

**Nicholas J. Salgia 1, Errol J. Philip 2, Mohammadbagher Ziari 3, Kelly Yap <sup>4</sup> and Sumanta Kumar Pal 1,\***


Received: 29 April 2020; Accepted: 14 May 2020; Published: 17 May 2020

**Abstract:** The treatment of metastatic renal cell carcinoma (mRCC) has rapidly evolved; however, the progress made in the field is heavily contingent upon timely and efficient accrual to clinical trials. While a substantial proportion of accrual occurs at tertiary care centers, community sites are playing an increasing role in patient recruitment. In this article, we discuss strategies to optimize collaborations between academic and community sites to facilitate clinical research. Further, as the role of biomarker discovery has become increasingly important in tailoring therapy, we will discuss opportunities to bridge diverse accrual sites for the purpose of translational research.

**Keywords:** renal cell carcinoma; team medicine; translational research; community practice; clinical trials

#### **1. Introduction**

Two decades ago, the outlook for patients with metastatic renal cell carcinoma (mRCC) was bleak. Treatments such as interferon-α (IFN-α) and interleukin-2 (IL-2) yielded a median survival of approximately one year [1]. The landscape changed tremendously with the introduction of targeted therapies from 2005 onwards, and, once again, with the introduction of checkpoint blockade a decade later [2–5]. In 2020, the current standard for front-line treatment of mRCC remains either dual checkpoint inhibition or a combination of targeted therapy with checkpoint blockade [6–8]. Survival estimates have essentially tripled what was anticipated in the cytokine era.

The rapid developments in mRCC therapy have been made possible by the timely and efficient completion of both early and late phase clinical trials. In addition, the discovery of novel therapies has been fostered by translational research efforts—the intersect of the bench and bedside. This is especially true in RCC, which has directly benefitted from research discoveries that have led to both the 2018 and 2019 Nobel Prizes in Physiology or Medicine. While tertiary centers (e.g., cancer centers and academic hospitals) serve as a common ground for early phase clinical trials and translational work, there is no doubt that their efforts can be bolstered through the participation of community oncology practices. Further, late-phase clinical trials, which often require less stringent data collection and less intensive evaluation schedules, can perhaps be done equally well at community and academic sites.

Whereas the prognosis of mRCC has improved, it is important to note that most patients with this disease are not cured, and thus, further research is imperative. In the current manuscript, we review the clinical landscape for renal cell carcinoma and delve into opportunities to maximize collaborations between community-based practices and academic sites in the hope of promoting more efficient and effective research endeavors. We approach the latter objective in the specific context of City of Hope—a tertiary cancer center that serves the greater Los Angeles area, a catchment that includes approximately 20,000,000 people. City of Hope has within its network 30 community-based satellite clinics, stretching over a radius of 100 miles [9]. With this vast network of community sites and the robust patient population that comes with it, multiple opportunities for collaborative research exist. However, selecting the right research for community versus academic practices is no small feat. Whereas our academic practice at City of Hope is well-staffed, with 5 research nurses and 7 clinical research associates devoted to genitourinary cancers, community sites often have a limited number of research staff who are responsible for managing studies across multiple disease types. These limitations notwithstanding, the integration of academic and community centers has grown into a successful model for conducting clinical and translational research.

#### **2. Renal Cell Carcinoma Incidence and Histology**

Renal cell carcinomas are a broad classification referring to tumors originating from the renal pelvis and medulla. The incidence of RCC is increasing, with an estimated 73,750 newly diagnosed cases in 2020 in the United States [10]. The disease now accounts for 5% of cancer diagnoses in men and 3% in women. The rising incidence of RCC can be partially attributed to an increase in the incidental detection of small renal masses during the performance of abdominal imaging for non-specific complaints [11]. Although the majority of incidental findings are small, early-stage masses, these diagnoses have not led to a substantial change in the mortality rate attributed to RCC; in 2020, it is estimated that RCC will account for approximately 15,000 deaths in the United States [10].

The histological variations in RCC have been well-documented [12]. Clear cell RCC (ccRCC) represents the most common subtype, accounting for 80% of diagnoses. ccRCC is pathologically characterized by its clear cell histology and acinar growth patterns. The hallmark molecular alterations of ccRCC include chromosome 3p loss and inactivation of the Von-Hippel Lindau (VHL) gene, which lies at the 3p25 locus [13]. VHL inactivation can occur either somatically or in the germline. Inactivation of *VHL* has been reported in up to 90% of sporadic ccRCC cases [14,15]. Germline inactivations of *VHL* are associated with Von-Hippel Lindau syndrome, which predisposes individuals to bilateral renal masses, as well as hemangioblastomas and retinal angiomas [16]. Other mutations associated with ccRCC are *PBRM1*, *SETD2*, and *BAP1*, all of which also lie on chromosome 3p [15,17].

Non-clear cell RCC (nccRCC) is a heterogenous classification that encompasses the other 20% of RCC diagnoses. Papillary RCC (pRCC) composes the majority of nccRCCs (~15% of total RCC diagnoses). pRCC has historically been subdivided into two classifications based on histology: type 1 involves low-grade nuclei arranged in a single layer, and type 2 encompasses a wide range of morphological presentations that can be further divided into at least three additional subtypes. Type 1 pRCC has been associated with sporadic mutations in the MET protooncogene (10–13% of patients) [18]. Other nccRCCs include chromophobe RCC, MiT translocation RCCs, collecting duct carcinoma, and renal medullary carcinoma. Each of the above-listed subtypes accounts for less than 10% of all RCC diagnoses [19].

As ccRCC represents the overwhelming majority of RCC diagnoses, the remainder of this review will focus primarily on the management of clear cell disease and use RCC interchangeably for ccRCC.

#### **3. Managing Localized Renal Cell Carcinoma**

Localized RCC comprises the majority of RCC cases, and in turn, is typically cured with definitive treatment. The standard of care for most localized tumors has been nephrectomy, whether it be partial or radical. Cancer-specific survival rates for localized RCC patients have been estimated to be 84% at five years and 76% at ten years after surgical resection, varying widely with stage. Although these estimates are overwhelmingly positive compared to patients with metastatic disease, one area of research interest has been the potential utility of adjuvant or neoadjuvant therapies to increase disease-free and overall survival (OS). A variety of agents, including IFN-α, IL-2, and chemotherapies, have been trialed as adjuvant therapies for localized RCC, but none demonstrated an improvement in disease-free survival. As the treatment landscape for mRCC has evolved, so too have the investigatory agents for adjuvant and neoadjuvant therapy.

The ASSURE trial and S-TRAC trial both compared adjuvant sunitinib to placebo in patients with high-risk localized RCC [20,21]. ASSURE also contained a third arm (sorafenib). No difference in disease-free survival (5.8 versus 6.1 versus 6.6 years for sunitinib, sorafenib, and placebo, respectively) nor OS was seen in ASSURE, and substantial toxicities were associated with both sunitinib and sorafenib. However, S-TRAC reported a statistically significant increase in disease-free survival with sunitinib versus placebo (6.8 versus 5.6 years) on independent review, but still no difference in OS. Based on the results of S-TRAC, sunitinib was approved for adjuvant use. Although approved, sunitinib is not often leveraged in the adjuvant setting due to the substantial toxicities associated with its use. Various other trials have investigated small molecule inhibitors as adjuvant therapy: PROTECT compared pazopanib to placebo, ATLAS compared axitinib to placebo, and EVEREST compared everolimus to placebo [22–24]. Both PROTECT and ATLAS have reported out and did not demonstrate improved survival with adjuvant therapy, while EVEREST has completed accrual, but results have not yet been reported.

Additional investigation of adjuvant therapy for localized disease has incorporated immune checkpoint inhibitors (CPIs). The recently opened PROSPER RCC trial randomizes patients to receive surgery alone or surgery with perioperative nivolumab (Figure 1) [25]. This study continues to accrue patients. IMmotion 010 also has investigated the role of adjuvant CPI therapy for resected RCC, in this case, comparing atezolizumab to placebo for patients with high-risk of disease recurrence [26]. This trial has not reported results to date. As the disclosed trials have reported minimal or no improvement with adjuvant therapy for localized RCC, systemic therapy is primarily isolated to patients with metastatic disease.

**Figure 1.** Schema for the phase-III ECOG-ACRIN PROSPER study.

#### **4. The Evolving Management of Metastatic Renal Cell Carcinoma**

mRCC carries a relatively poor long-term survival (12% five-year OS) when compared to localized disease (93% five-year OS)—of course, these estimates are changing with emerging therapies. The systemic therapy landscape for mRCC has rapidly evolved within the last twenty years [27]. At the turn of the century, cytokine-based approaches with high-dose IL-2 and IFN-α represented the best systemic therapeutic option for patients with advanced disease. Both cytokines have been shown in animal and human models to have anti-cancer properties: IL-2 enhances the growth and activation of natural killer cells and CD-8+ T-cells, while IFN-α enhances tumor antigen presentation, among other characteristics [28,29]. While such cytokine approaches promote tumor regression mechanistically, IL-2 and IFN-α both provided limited clinical efficacy, with a median OS of approximately one year and durable responses in the range of only 5–7% [1,30,31]. These agents also posed substantial toxicity profiles, further limiting successful therapeutic outcomes.

#### *4.1. Risk Stratification: MSKCC and IMDC Criteria*

Among the most important non-therapeutic advances in the management of mRCC have been the development and implementation of risk classification algorithms. The Memorial Sloan Kettering Cancer Center (MSKCC) Score for renal cell carcinoma represents the first widely-adopted calculation for risk stratification of mRCC patients [31]. The initial MSKCC risk score included five prognostic factors: low Karnofsky performance status, high serum lactate dehydrogenase, low hemoglobin, high serum calcium, and absence of prior nephrectomy. These were used as risk factors to stratify patients into three groups: favorable risk (0 risk factors), intermediate risk (1–2 risk factors), and poor risk (3+ risk factors). Patients with MSKCC favorable risk disease were reported to have a median OS of 20 months, those with intermediate risk had a median OS of 10 months, and poor risk disease carried a median OS of 4 months. The MSKCC model has incorporated various additions into its algorithm over time. The Mekhail extension added two variables: prior radiotherapy and 2+ sites of metastasis [32]. In 2018, the MSKCC risk model was updated to incorporate genomic characteristics, specifically the mutational status of *BAP1*, *PBRM1*, and *TP53* in the stratification algorithm [33]. The MSKCC risk model has been used as an important inclusion criterion and stratification variable for patient randomization in various RCC therapeutic trials [5].

A second prognostic model has been developed and validated for mRCC: the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) Risk Score. Like the MSKCC model, the IMDC score utilizes clinical variables as prognostic markers to define patient risk. The IMDC model incorporates six prognostic variables: less than one year from the time of diagnosis to onset of systemic therapy, low Karnofsky performance status, low hemoglobin, high calcium, high neutrophil, and high platelet levels [34]. Similar to the MSKCC model, patients are classified into one of three risk groups. Favorable risk (0 factors) carries a median OS of 43.2 months, intermediate risk (1–2 factors) carries a median OS of 22.5 months, and poor risk (3+ factors) carries a median OS of 7.8 months [35]. Like the MSKCC model, the IMDC score has been adopted as a key inclusion criterion and stratification metric for many pivotal RCC trials [6].

#### *4.2. VEGF Tyrosine Kinase Inhibitors, Multi-Kinase Inhibitors, and mTOR Inhibitors*

The first shift in the therapeutic approach to mRCC occurred with the advent of small-molecule inhibitors that bind to and inhibit the activity of membranous receptors and intracellular proteins. The primary class of small molecule inhibitors in the mRCC treatment armamentarium are vascular endothelial growth factor–tyrosine kinase inhibitors (VEGF–TKIs). Mechanistically, *VHL* mutations lead to decreased ubiquitinylation of hypoxia-inducible factor (HIF) and upregulation of the circulating VEGF molecule which then binds the VEGF receptor, promoting angiogenesis [36,37]. VEGF–TKIs inhibit the tyrosine kinase domain of the VEGF receptor, and in turn, block the intracellular signaling cascade that promotes angiogenesis and cell division

Multiple VEGF–TKIs have been approved for mRCC. One benchmark phase-III study compared the VEGF–TKI sunitinib to IFN-α in patients with previously untreated mRCC [2]. Sunitinib was demonstrated to prolong progression-free survival (PFS) compared to IFN-α (11 versus 4 months) and was associated with a higher objective response rate (ORR) (31% vs. 6%). The results of this

study led to the approval of sunitinib for first-line mRCC patients. Other VEGF–TKI agents approved in mRCC include sorafenib, pazopanib, and axitinib. While these agents have prolonged PFS and produced improved response rates compared to the previous standard-of-care cytokine therapies, VEGF–TKIs are not curative, and patients are susceptible to disease progression upon the development of drug resistance.

An additional class of targeted therapies, so-called multi-kinase inhibitors have also been approved in mRCC. These agents not only act as VEGF–TKIs but also inhibit the tyrosine kinase domains of additional protooncogenes [38]. Cabozantinib is a multi-kinase inhibitor with activity as a VEGF–TKI and also as an inhibitor of MET and AXL, both of which are associated with resistance to VEGF–TKIs. Cabozantinib was first approved for mRCC patients with treatment-refractory disease but was soon trialed as first-line therapy. The phase-II CABOSUN trial compared cabozantinib to sunitinib in the front-line setting [39]. This study met its primary endpoint of improvement in PFS with cabozantinib (8.6 versus 5.3 months) and demonstrated a higher ORR with cabozantinib based on an independent review (20% versus 9%). These results led to the approval of cabozantinib across all lines of therapy for patients with mRCC. However, like sunitinib and other VEGF–TKIs, cabozantinib has limited curative potential. As such, the approach of managing mRCC in the front-line with VEGF–TKIs and multi-kinase inhibitors has been replaced by the recent introduction of immune checkpoint inhibitors.

The mammalian target of rapamycin (mTOR) represents a highly-important intracellular target of mRCC therapy. mTOR is an enzymatic intermediate in the PI3K/AKT/mTOR signaling pathway that regulates the cell cycle [40]. Dysregulation of this pathway is a metabolic feature of many RCC tumors, making the components of its signaling cascade viable targets for pharmacologic inhibition. Two therapies approved for the management of mRCC act in this manner on mTOR: everolimus and temsirolimus. Everolimus is an oral agent that has been approved in combination for mRCC with the VEGF–TKI lenvatinib, following the results of a phase II study comparing the combination to each respective monotherapy [41]. Everolimus with lenvatinib resulted in PFS benefit compared to everolimus alone (14.6 versus 5.5 months) for patients who had received at least one prior VEGF–directed therapy. Temsirolimus, an inhibitory ester analog of mTOR that is applied intravenously, is also approved for mRCC [42]. A phase III trial demonstrated an OS benefit for temsirolimus over IFN-α (10.9 versus 7.3 months) with fewer incidences of serious adverse events [43]. Although no longer a mainstay of early-line therapy, the mTOR inhibitors everolimus and temsirolimus remain important interventions for the management of mRCC. Like VEGF–TKIs and multikinase inhibitors, however, mTOR inhibitors have given way to immune checkpoint inhibitors for the early-line treatment of mRCC.

#### *4.3. Immune Checkpoint Inhibitors*

Immune checkpoint inhibition is a relatively recent development in solid tumors, first explored in the context of melanoma [44]. CPIs are synthesized monoclonal antibodies capable of overcoming T-cell inactivation to elicit an anti-tumor response [45]. Approved CPIs exhibit activity on one of two immunoregulatory axes: programmed-death-1/programmed-death-ligand-1 (PD-1/PD-L1) and the cytotoxic-T-lymphocyte-associated protein 4 (CTLA-4). The treatment paradigm for mRCC has advanced greatly since the advent of PD-(L)1 and CTLA-4 inhibitors.

CheckMate-025 was the first study to investigate nivolumab, an anti-PD-1 CPI, in mRCC [5]. This study enrolled patients who had progressed on one or two previous VEGF–TKIs and randomized them to receive nivolumab or everolimus. Nivolumab outperformed everolimus in OS (25 versus 19.6 months) and ORR (25% versus 5%), leading to the approval of nivolumab monotherapy for mRCC patients who had progressed on previous therapy.

The ensuing CheckMate-214 trial compared the combined regimen of nivolumab with ipilimumab, an anti-CTLA-4 CPI, to sunitinib in previously-untreated mRCC [6]. Patients were stratified based on the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC) risk score. In patients with intermediate or poor risk disease, nivolumab + ipilimumab greatly outperformed sunitinib, demonstrating greater ORR (42% versus 27%) and a rate of complete response totaling 9%. However, among patients with favorable risk disease by IMDC criteria, sunitinib was shown to result in prolonged PFS (25.1 versus 15.3 months) and a greater ORR (52% versus 29%). The results from CheckMate-214 established a new treatment algorithm for front-line mRCC, in which dual checkpoint inhibition with nivolumab + ipilimumab form the backbone of therapy for intermediate/poor risk patients but VEGF–TKIs remain a viable first-line option for patients with favorable risk.

#### *4.4. Combination Therapies*

Further endeavors have clinically investigated combinations of CPIs with VEGF–TKIs. A myriad of large phase III clinical trials have been undertaken, with three having reported out in early 2019. The IMmotion 151 trial compared the investigatory combination of atezolizumab (an anti PD-L1 CPI) with bevacizumab (an anti-VEGF monoclonal antibody) against sunitinib [46]. This trial met its primary endpoint of improvement in PFS with bevacizumab + atezolizumab (11.2 versus 7.7 months) in the PD-L1+ population. In the intention-to-treat cohort that included all patients regardless of PD-L1 status, PFS favored the combination therapy over sunitinib (11.2 versus 8.4 months). JAVELIN-101 followed a similar design model, comparing the combination of avelumab (an anti-PD-L1 CPI) and axitinib against sunitinib in patients with mRCC who had received no prior therapy [8]. The primary endpoint of improved PFS with axitinib + avelumab in patients who had PD-L1+ disease was met (13.8 versus 7.2 months). Likewise, irrespective of PD-L1 status, axitinib + avelumab was again associated with improved PFS (13.8 versus 8.4 months).

The KEYNOTE-426 trial utilized an experimental arm of pembrolizumab (an anti-PD-1 CPI) with axitinib, which was compared against sunitinib [7]. The axitinib + pembrolizumab combination resulted in a prolonged PFS compared to sunitinib (15.1 versus 11.1 months), and, additionally, ORR was greater in the axitinib + pembrolizumab arm (59% versus 36%). These three trials all reported within the span of two months in 2019. Table 1 summarizes results from trials testing CPIs (in combination with an additional CPI or with VEGF–TKIs) for the first-line treatment of mRCC. A recent press release in April 2020 has also indicated that the CheckMate-9ER trial (NCT03141177) of nivolumab + cabozantinib versus sunitinib for front-line mRCC has met its primary endpoint of improved PFS and secondary endpoints of ORR and OS. The combination of axitinib + pembrolizumab, in particular, has already established a role in the mRCC treatment algorithm for front-line intervention in patients with IMDC good-risk disease [47]. As data from these trials continue to mature and additional trials are undertaken, it is reasonable to believe the approval and usage of combination therapies will expand.



Abbreviations: ORR = overall response rate, CI = confidence interval, PFS = progression-free survival, HR = hazard ratio, OS = overall survival. 99.1% confidence interval;†99.8% confidence interval.

#### *4.5. Cytoreductive Nephrectomy for Metastatic Disease: A Continuing Discussion*

The utility of cytoreductive nephrectomy as a component of care for mRCC remains a highly-investigated topic. Cytoreductive nephrectomy for removal of the primary tumor in patients with metastatic disease was a standard-of-care for twenty years. However, two recent trials, SURTIME and CARMENA, have changed the paradigm for nephrectomy in mRCC. SURTIME compared immediate nephrectomy (followed by sunitinib therapy) versus deferred nephrectomy (preceded by sunitinib therapy) [48]. The results of this study indicated that deferred nephrectomy did not improve the progression-free rate at 28 weeks compared to immediate nephrectomy (43% versus 42%), but OS was higher in the deferred arm (32.4 versus 15.0 months). These results suggested that initial systemic therapy followed by a potential debulking nephrectomy may be more efficacious for mRCC patients than the former standard of upfront nephrectomy.

CARMENA randomized patients with mRCC to undergo nephrectomy followed by sunitinib or to receive sunitinib alone [49]. Results from this study indicated that sunitinib alone offered a non-inferior alternative to nephrectomy followed by sunitinib, with OS favoring the former (18.4 versus 13.9 months). The results of CARMENA and SURTIME have led to a shift away from cytoreductive nephrectomy as a standard-of-care for mRCC, except in carefully planned cases. However, ongoing trials continue to study the relevance of cytoreductive nephrectomy. The NORDIC-SUN trial (NCT03977571) investigates nephrectomy following therapy with nivolumab + ipilimumab. Patients with IMDC favorableor intermediate-risk disease are randomized to receive nephrectomy or maintenance nivolumab. Patients who present with poor-risk disease can be randomized within this study as well, so long as their risk classification is adjusted to intermediate-risk following initial therapy with nivolumab + ipilimumab or while on maintenance nivolumab. An additional trial, CYTOSHRINK (NCT04090710), uses stereotactic body radiation therapy (SBRT) as an investigational definitive treatment of the primary tumor site for patients with mRCC. In this trial, patients are randomized to receive the standard-of-care nivolumab + ipilimumab for four cycles, followed by maintenance nivolumab or the experimental arm. The experimental arm consists of one cycle of nivolumab + ipilimumab followed by five fractions of SBRT to the kidney lesion before resuming nivolumab + ipilimumab on the standard schedule. Both these trials are currently open to accrual.

#### **5. Integrating Community and Academic Practices for Renal Cell Carcinoma Research**

As RCC incidence continues to grow, the role of the community oncologist has grown increasingly important in the network of care for RCC patients. While clinical and translational research has primarily been led by academic oncologists, a new approach incorporating community practitioners into the research paradigm is now viable. It is important that medical oncologists at academic centers integrate community practitioners, particularly urologists and oncologists, into a collaborative model that promotes exceptional, multi-disciplinary care alongside cutting-edge clinical and translational research. Through the City of Hope network, which encompasses a central academic center and a diffuse network of community sites, we have developed a collaborative model for conducting translational research and clinical trial accrual that integrates academic and community oncologists. Below, we have highlighted our experiences utilizing this integrated structure to advance clinical and translational research and discovery in RCC.

#### **6. Collaborations in Translational Research**

As detailed above, the therapies currently in use for mRCC reflect two biological principles, namely, that (1) RCC is driven by angiogenesis and that (2) defects in the immune system can propagate the disease. One of the genomic hallmarks of RCC are defects in the *VHL* gene [15,50]. Mutation of *VHL* (largely sporadic, but possibly hereditary) leads to decreased ubiquitinylation of HIF [51–54]. This, in turn, leads to the upregulation of vascular endothelial growth factor (VEGF), a potent mediator of angiogenesis [55,56]. The interplay of RCC with the immune system is much more complex. Multiple

studies have shown varying levels of expression of the immune checkpoint programmed death-ligand 1 (PD-L1) in RCC tissues [7,57,58]. Interaction of PD-L1 with its cognate receptor, programmed death-1 (PD-1), leads to T-cell anergy, which can be overcome with antibodies targeting either entity [59–61].

Interestingly, while VEGF–pathway inhibitors and immune checkpoint inhibitors both reflect "targeted therapies", both are applied in a biomarker-agnostic fashion in patients with mRCC. At City of Hope, a partnership has been forged with TGen, Inc., a company focused on translational genomics. Scientists and clinicians at TGen have developed the Ashion GEM ExTra® platform that allows for both whole-exome sequencing and RNA sequencing [62]. A sample report is shown in Figure 2. In addition to requiring tissue for tumor sequencing, the platform uses blood for germline correction.



**Figure 2.** A sample report from the Ashion GEM ExTra test.

A sequencing platform such as Ashion GEM ExTra® offers two benefits in terms of translational research. First and foremost, collecting this data and pooling it across academic and community sites may allow for the development of predictive and prognostic biomarkers. In a recent effort, Salgia et al. reviewed data from 90 patients with mRCC (Salgia et al. ASCO 2020) and examined the association between treatment type and response among those that had received immunotherapy or targeted therapy. Notably, DNA level data suggested that *TERT* promoter alterations were a negative predictor of immunotherapy response, with RNA level data from the study identifying multiple *TERT*-associated pathways that could play a role in immunotherapy resistance.

A further benefit of amassing translational data across community and academic sites is the ability to identify novel opportunities for clinical research. Our pooled data indicate, for instance, a preponderance of mutations in *VHL*, *PBRM1*, *SETD2*, and other mutations in the mammalian target of rapamycin (mTOR) pathway. We have also observed infrequent but clinically significant mutations in genes such as *EGFR*, *MET*, and *ALK* [63,64]. Many phase I and II trials in development focus specifically on these mutations but do so in a tumor agnostic fashion. The genitourinary group physicians at our institution have partnered with community sites and other disease teams to categorize cumulative

mutational frequency. By pooling this data across our sites, we can make logical decisions regarding the most appropriate clinical trials to bring into our portfolio.

Biomarker discovery and validation are not isolated to genomics, however. Recent work has elucidated associations between microbiome composition and clinical response to CPIs in mRCC [65]. Additional work from our group and collaborators at TGen, Inc. has investigated the enteric microbiome as a potential biomarker for responses to the above-mentioned therapy regimens in RCC [66]. We have utilized whole genome shotgun metagenomic sequencing to correlate the stool microbiome profile and microbiome diversity of mRCC patients with clinical benefit, both for patients receiving VEGF–TKIs and for patients receiving CPIs (Dizman et al. ASCO 2020). In these studies, we have identified a variety of microbial species (such as *B. Intestihominis* and *B. adolescentis*) that are correlated with clinical benefit in patients receiving systemic therapy for mRCC. The diversification of biomarker pursuits provides another avenue for research collaboration between academic and community partners. Protocols for the collection, storage, and shipment of stool, for example, can be implemented at any site, even those with limited research infrastructure. The clinical volume of community practice sites poses an untapped resource for furthering biomarker discovery and other translational studies in RCC.

#### **7. Collaborations in Clinical Research**

#### *7.1. Phase I and II Clinical Trials*

Phase I clinical trials have morphed in recent years, moving away from the classical 3 + 3 design [67]. More recent iterations of phase I clinical trials have started with a dose-escalation phase, expanding quickly into phase Ib trial designs, all while seeking validation in disease- or mutation-specific settings. We propose that phase I studies in dose escalation are perhaps most appropriate to remain at academic centers. The dose-escalation phase is one where rigorous oversight is necessary, with multiple visits and correlative studies. Furthermore, this is often the phase in which novel toxicities associated with an agent or combination are recognized and documented. We propose that highly selected phase Ib studies could be conducted in the community (Figure 3). Criteria for site selection in these studies must account for the frequency of study visits and the rigor of correlatives. For instance, if a proposed study requires pharmacokinetic and pharmacodynamic blood assessments on hourly intervals (not unusual in a phase-I protocol), the trial may be challenging in a busy community practice. Similarly, if a study requires repeat biopsies or other invasive assessments, academic centers may be better suited.

Indeed, there are real-world examples of phase I studies that may be too challenging to envision in a community practice setting at this time. For instance, our genitourinary cancers group is heavily invested in studies of chimeric antigen receptor (CAR)–T-cell therapies (Dorff et al. ASCO GU 2020). In addition to the complexities associated with the manufacturing and storage of CAR–T-cells, patients must be kept in an inpatient setting for prompt recognition and treatment of unique side effects of this treatment, such as cytokine release syndrome [68]. At present, commercially available CAR–T-cell therapies are used in a limited subset of patients with hematologic malignancies [69]. Until these therapies are adopted in a much more widespread fashion in the community (with appropriate infrastructure and monitoring), we recommend these studies remain within academic centers.

The phase II study is a vanishing entity, as companies often now move quickly from a large phase Ib effort to phase III. This is true in mRCC space as well—as one example, COSMIC-021 (NCT03170960) is an international study chaired by investigators at our site. This phase Ib study of cabozantinib with atezolizumab enrolled cohorts in RCC, lung cancer, and prostate cancer, as well as more than a dozen other histologies. Based on significant activity in the three noted cohorts, phase III clinical trials have been rapidly launched, including CONTACT-03 (NCT04338269) for mRCC [70]. In general, if phase II studies are to be considered in the community setting, the same factors (visit frequency, correlative studies, and so on) should aid in deciding whether studies are appropriate and should proceed.

**Figure 3.** Schema outlining appropriate trial selection for community and academic settings.

#### *7.2. Phase III Clinical Trials*

Phase III clinical trials represent an area where collaborations are imperative. These studies often include hundreds (if not thousands) of patients, and to accrue in a rapid fashion, a partnership between academic and community sites is essential. There are two general categories of phase III studies that we consider in our program—one falls under the umbrella of pharmaceutical research, and the other relies upon funding from the National Cancer Institute (NCI), typically through the cooperative group mechanism. The cooperative groups have recently been consolidated—several groups (SWOG, Alliance, NRG, ECOG-ACRIN, and COG) retain members at both academic and community sites.

In general, cooperative group studies have more recently adopted easier thresholds for enrollment and less intensive patient follow-up. An example of this is the phase III PROSPER study, run by ECOG-ACRIN, which we have previously highlighted [25]. This phase III trial compares surgery alone for localized or oligometastatic RCC to surgery with both preoperative and postoperative nivolumab, a PD-1 inhibitor (Figure 1). The study has relaxed eligibility criteria with recent amendments—with randomization possible prior to a histologic diagnosis, and a pretreatment biopsy only required if patients are randomized to perioperative therapy. The study also does not mandate central review.

Such eligibility requirements differ widely from comparable pharmaceutical trials. For example, investigators at our institution led the phase III Immotion010 clinical trial, a study comparing atezolizumab to placebo in patients with completely resected RCC [26]. The study mandated pre-registration assessment of PD-L1 status, and further required central submission of scans. Not only do such measures increase the potential rate of screen failures, but they also increase the complexity of trial administration. Submitting tissue and scans in the pre-registration period (typically 2–3 weeks) is a wieldy task, and community sites must establish whether they have adequate resources to devote to such tasks.

Further, factors such as the study population, in particular tumor histology, must also be carefully considered in the conduct of trials. For example, studies investigating rare tumor histologies (nccRCCs) are challenging to conduct in the community setting due to the infrequency of disease presentation. One such example is SWOG 1500 (NCT02761057), a study of selective MET-kinase inhibitors for patients with papillary renal cell carcinoma that was chaired by physicians in our genitourinary oncology disease team. Although the study was sponsored by an NCI cooperative group, the relative rarity of papillary histology (~10–15% of metastatic renal cell carcinomas) dictated that the study only be available at the academic center [71]. As patients with rare histologies are often referred from the community to experts at academic centers, consolidation of trial sites in these instances should be considered.

#### **8. Optimizing Partnerships between Community and Academic Sites**

As noted previously, the academic cancer center is typically well staffed with multiple research nurses and clinical research associates who work within a single disease space. Each has ownership over a selected number of protocols (typically 2–3), and there are often regular meetings with the principal investigator. These face to face meetings with principal investigators offer an opportunity for education regarding the protocol and, furthermore, allow for a closer degree of oversight of protocol-related activities. At City of Hope, investigators in the genitourinary disease team frequently host didactic sessions for research staff, offering them insight into the current standard management of RCC, bladder, and prostate cancer. As research staff are often charged with screening for enrollment or suggesting toxicity attributions, understanding the underlying disease and associated treatment landscape is critically important.

Models of research in community practices vary but commonly require research staff to manage a much broader swath of clinical protocols across multiple cancer types. Furthermore, the principal investigator for a study may not reside at the community practice site, and thus is unable to engage with clinical and administrative staff on a routine basis. At City of Hope, the vast majority of principal investigators reside at the main campus; thus, elements of study oversight (filing significant adverse events, reporting study deviations) may be somewhat delayed at community sites. In addition, services readily available at the main campus may be less accessible in the community practices; for example, labs may be obtained off-site through third-party vendors. This can create delays in lab reporting, and complicate study protocols that require same-day lab results to issue therapy. Many protocols also embed correlates that require additional processing methods (e.g., centrifugation for separation of plasma aliquots).

Similarly, access to appropriate and timely scans can vary by recruitment location, with an academic center generally able to accommodate access to both standard and novel imaging tests (e.g., fluciclovine or PSMA PET scans for prostate cancer) and not have to rely on third parties to transfer images so that relevant tumor measurements can be obtained. Whereas these may be trivial issues on a main academic campus, access to a centrifuge, processing equipment, and appropriate imaging modalities may be limited at community sites.

Despite these noted limitations, community practices can serve as valuable contributors to clinical research endeavors housed at a main campus. To optimize this relationship and promote successful collaborations, we have implemented several strategies. Firstly, self-nominated community disease team leads are identified to serve as liaisons between the research team on the main campus and oncologists practicing in the community. These individuals also play a role in determining the feasibility of specific clinical trials in the community, taking into account many of the considerations that we have outlined throughout. In addition, these individuals are invited to bimonthly genitourinary disease team meetings, where the group conducts a thorough review of all existing protocols and evaluates new clinical trial proposals. Finally, in recognition of the limited time permitted to community oncologists to learn about existing research endeavors, our genitourinary cancers group has established a newsletter that highlights ongoing and planned clinical trials with only the most salient information—basic eligibility requirements and study schema, for instance. These newsletters have triggered a number of direct referrals for clinical trial enrollment.

One relatively recent development among our community-based satellite clinics is the implementation of a standardized electronic health record (EHR) mirroring that at the main campus, optimizing the process by which physicians can evaluate and triage patients who may be suitable clinical trial candidates.

#### **9. Conclusions**

The clinical approach to RCC has drastically changed over the past two decades. The novel therapeutic options introduced to the medical oncologist's armamentarium have leveraged ground-breaking work on angiogenesis and induced hypoxia in addition to the principle of immune checkpoint blockade to overcome T-cell anergy. To maintain the momentum of RCC drug development, rigorous translational research must continue advancing our understanding of RCC biology and clinical trials must continue to accrue patients at a rapid pace. As we have highlighted herein, this can happen effectively and efficiently if community practices are integrated with academic centers in the conduct of translational research and clinical trials. We have proposed a number of strategies to optimize this relationship, with the common theme being enhanced communication. When oncologists at community and academic practices frequently dialogue, gaining a deeper understanding of each other's practice patterns, a portfolio of trials can be developed that optimizes accrual and advances scientific findings more rapidly.

The positive effects of an academic and community collaboration on trial accrual have been evidenced in our network—with our genitourinary cancers group recognized as the top accrual site for the adjuvant EVEREST study, a trial comparing everolimus and placebo in patients with resected localized RCC [24]. Our enrollment was driven in large part by patients seen at our community-based satellite offices. The rapid success of this collaborative effort was used as a foundation for the enrollment of patients into further adjuvant trials (e.g., E2810 and Immotion010), with similar positive outcomes. In sum, a strong collaborative approach between our academic center and affiliated community practices has resulted in a model conducive to high accrual rates across a broad range of clinical trials. This feat, along with the opportunities for collaboration in the translational sciences, underscores the benefit of these partnerships.

**Author Contributions:** Conceptualization, N.J.S., S.K.P.; Methodology, N.J.S., S.K.P.; Resources, S.K.P.; Data Curation, N.J.S., S.K.P.; Writing—Original Draft Preparation, N.J.S., S.K.P.; Writing—Review & Editing, N.J.S., E.J.P., K.Y., M.Z., S.K.P.; Visualization, N.J.S., S.K.P.; Supervision, S.K.P.; Project Administration, N.J.S., E.J.P., K.Y., M.Z., S.K.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** Nicholas Salgia, Errol Philip, Mohammadbagher Ziari, and Kelly Yap have no conflicts of interest that might be relevant to the contents of this manuscript. Sumanta K. Pal, MD—honoraria: Novartis, Medivation, Astellas Pharma; consulting or advisory role: Pfizer, Novartis, Aveo, Myriad; pharmaceuticals: Genentech, Exelixis, Bristol-Myers Squibb, Astellas Pharma; research funding: Medivation.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Integrating Academic and Community Cancer Care and Research through Multidisciplinary Oncology Pathways for Value-Based Care: A Review and the City of Hope Experience**

**Linda D. Bosserman 1,\*, Mary Cianfrocca 1, Bertram Yuh 2, Christina Yeon 3, Helen Chen 4, Stephen Sentovich 2, Amy Polverini 5, Finly Zachariah 6, Debbie Deaville 7, Ashley B. Lee 8, Mina S. Sedrak 1, Elisabeth King 9, Stacy Gray 9, Denise Morse 10, Scott Glaser 11, Geetika Bhatt 12, Camille Adeimy 13, TingTing Tan 14, Joseph Chao 1, Arin Nam 1, Isaac B. Paz 5, Laura Kruper 2, Poornima Rao 5, Karen Sokolov 15, Prakash Kulkarni 1, Ravi Salgia 1, Jonathan Yamzon <sup>2</sup> and Deron Johnson <sup>16</sup>**


Received: 10 November 2020; Accepted: 29 December 2020; Published: 7 January 2021

**Abstract:** As the US transitions from volume- to value-based cancer care, many cancer centers and community groups have joined to share resources to deliver measurable, high-quality cancer care and clinical research with the associated high patient satisfaction, provider satisfaction, and practice health at optimal costs that are the hallmarks of value-based care. Multidisciplinary oncology care pathways are essential components of value-based care and their payment metrics. Oncology pathways are evidence-based, standardized but personalizable care plans to guide cancer care. Pathways have been developed and studied for the major medical, surgical, radiation, and supportive oncology disciplines to support decision-making, streamline care, and optimize outcomes. Implementing multidisciplinary oncology pathways can facilitate comprehensive care plans for each cancer patient throughout their cancer journey and across large multisite delivery systems. Outcomes from the delivered pathway-based care can then be evaluated against individual and population benchmarks. The complexity of adoption, implementation, and assessment of multidisciplinary oncology pathways, however, presents many challenges. We review the development and components of value-based cancer care and detail City of Hope's (COH) academic and community-team-based approaches for implementing multidisciplinary pathways. We also describe supportive components with available results towards enterprise-wide value-based care delivery.

**Keywords:** value-based care; value-based cancer care; oncology pathways; Early Recovery After Surgery (ERAS); team-based care; oncology medical home; integrated cancer care; supportive care pathways; surgical pathways; cancer care plans

#### **1. Introduction: Multidisciplinary Oncology Pathways Are a Foundation of Value-Based Cancer Care**

While the majority of cancer care in the US is provided in the community (non-academic) setting, over the past 12 years, the organizational structure of oncology care delivery has shifted to networks of community oncologists partnered with academic centers, hospital systems, or other community practices. Aligned enterprises have evolved from the well-documented closing, merging, or acquisition of community oncology practices since 2008 [1,2]. This change has been driven by two major factors: the move from volume- to value-based payments, requiring more infrastructure management, resources, and market share as well as the rapidly increasing complexity of information needed to provide each cancer patient with the highest quality care while maintaining high patient and staff satisfaction, institutional health, and delivering these at optimal costs [3–5]. Meeting these goals has required innovations and teamwork that was pioneered first in several medical oncology home pilots from practices on the east coast by Dr. Sprandio and on the west coast by Dr. Bosserman as well as other practices across the US [6–11]. In the southwest, Dr. McAneny expanded on her New Mexico medical home model to engage six other US oncology practices in the successful Center for Medicare and Medicaid Innovation (CMMI)-funded COME HOME project [8]. Oncology pathways are a major component of these programs, with studies of implemented pathway programs showing equal or improved outcomes with lower costs in medical oncology [12–17], as well as in other disciplines including surgery, radiation oncology, supportive care, and end of life [18–23]. The evaluation and adoption of oncology pathways were further stimulated by Center for Medicare and Medicaid Services (CMS)'s oncology care MIPPs and Oncology Care Model (OCM) alternate payment programs and private payer pilots with academic and community-based networks and various private payors [24–33].

While multidisciplinary oncology pathway programs have evolved as essential components for evidence-based cancer care delivery within value-based cancer care (VBCC) programs, challenges remain to fully integrate and measure outcomes from multidisciplinary oncology pathways, including when they are combined into care plans across a patient's cancer journey [34–40]. As a practical point, multidisciplinary pathways first must be adopted, implemented, and understood for each specialty before their combined impacts can be evaluated. The processes, teams, and tools to develop, implement, analyze, and iterate multidisciplinary pathway projects throughout a large enterprise remain a work in progress. We detail here the work to date by the City of Hope (COH) academic community network enterprise as an example of the methodology and tools, some well-established and some being pioneered, to implement multidisciplinary oncology pathways and outline many supporting projects that facilitate and optimize their impacts. We share available analytics which continue to evolve and expand in the hope that our work can help others in their transformation to VBCC.

#### **2. Prioritizing Care Model Redesign for Value-Based Cancer Care (VBCC) at City of Hope with Multidisciplinary Oncology Pathways**

In response to the City of Hope Enterprise leadership prioritizing the delivery of reportable value-based oncology care, a Value Realization Project (VRP), led by a multidisciplinary group of clinicians, pharmacists, nurses, administrators, informaticists, program managers, quality, data analysts, risk and outcome experts, and our chief medical and administrative officers, was formed. This VRP group developed a Value Framework (VF) with specific projects identified under each of the three pillars of evidence-based care, care management, and care after cancer, as shown in Figure 1. These pillars are derived from the key components identified in the medical oncology home and OCM models as well as several value-based payer initiatives which have shown value [6–8,11,23–33]. A full discussion of our value-based care initiative is beyond the scope of this paper; however, we present our value framework and highlights of several other key projects within the framework to show the interdependence of those projects in advancing the implementation and evolving outcome analytics of multidisciplinary oncology pathways. Interdependent projects include the capture of complete staging data in the electronic health record (EHR), systemic therapy regimen orders built into the EHR, integration of clinical trials and pathway decision support into the EHR, and incorporation of precision medicine, among others. These companion projects support measurable evidenced-based cancer care delivery. We discuss some of the pathways and projects in the other two pillars of VBCC: management of care while on active treatments, and care after active cancer, whether in survivorship or at end of life, as part of our COH supportive care programs. Not shown in Figure 1 are the enterprise-wide initiatives to engage and measure patient and staff satisfaction, institutional health as well as the administrative support needed in informatics, precision medicine, EHR enhancements, care models, and payer contracting to fully meet VBCC goals. As we build each of the identified projects supporting multidisciplinary oncology pathways, we are working on measuring individual pathway use and impacts while working toward measurement of the larger clinical, financial, clinical trial, and quality outcomes. We also continue to iterate methodologies for team-based care and patient engagement trough academic and community staff teams in the evolution toward a mature VBCC model.

**Figure 1.** Value-based care framework projects for City of Hope Enterprise: Green represents the three pillars of value-based care: evidence-based care, care management, and care after cancer. The boxes represent categories of projects to facilitate measurable delivery of value-based care. Purple indicates clinician-led projects. Blue indicates patient-focused projects. Orange indicates projects after active therapy or curative therapy, whether in survivorship or end of life. Within the category of complete discrete patient diagnosis are the projects that will help to accomplish this. Omics refers to genomics, proteomics, metabolomics, and microbiome information that impacts a patient's treatment episode choices and potential outcomes. SDH refers to social determinants of health. Med Onc—medical oncology, Hem—hematology, Rad Onc—radiation oncology. ClinPath is the pathway system by Elsevier (formerly called VIA Pathways). EPIC refers to the EHR. IM—intramuscular, IV—intravenous. EHR—electronic health record. ERAS—early recovery after surgery pathways. SOP—standard operating procedures. Televisits—telephone and televideo visits. The purple arrow indicates that goals of care pathways are incorporated during active therapy as well during in advanced end of life care.

#### **3. NCCN Guidelines and Growing Complexity of Evidence-Based Cancer Care Also Stimulated Multidisciplinary Oncology Pathway Development at COH**

City of Hope (COH) is a National Cancer Institute (NCI)-designated comprehensive care center and one of the original 13 member institutions that formed the National Comprehensive Cancer Center Network (NCCN) in January of 1995. Within the NCCN, our faculty have participated in and led the development of guidelines for the treatment of cancers for the past 25 years. The COH enterprise now serves patients with cancer and metabolic diagnoses in a region of 25 million residents living in five counties in Southern California as well as more distantly based national and international patients. The COH enterprise includes a central academic, clinical research, hospital and clinical care campus in Duarte, California and a clinically integrated, community practice network of 31 community practices. Approximately 30% of COH physicians practice in the community (either solely or in a mixed hybrid of community/academic practice). Clinicians in the community are mostly general oncology practitioners, while the academic clinicians are disease and disease subtype specialists. Over the past year, we have identified community doctors who have a special interest in a specific cancer and will see patients with that disease preferentially (referred to as disease-leaning), such as breast, thoracic, gastrointestinal, and genitourinary cancers. A single foundation faculty model employs both academic center and community clinicians. The mission of the organization is to develop and bring cutting-edge precision cancer care and clinical research to every patient that is also personalized and delivered with compassion, high quality, and cost consciousness within the value-based cancer care initiative.

Although COH faculty participated in the development of the original eight disease guidelines published by the NCCN in November of 1996 and continue to participate and lead disease guideline committees at NCCN, until the more recent partnership between NCCN and US Oncology and their development of Level 1 Pathways powered by NCCN, the NCCN guidelines were broad and without the detailed pathway approach to better guide individual patient care [32]. COH disease teams regularly discussed narrowing the NCCN guidelines into pathways for specific patient issues and cancer subtypes and some, such as the breast, lung, and renal teams, develop and review pathways that are shared with COH faculty, as described in companion articles in this series [41–43]. With the growth of the COH enterprise, however, the need to more formally establish not only pathways for each disease but the ability to prompt for those choices, keep the pathways up to date, incorporate clinical trials as well as track the use and variations of treatments for each patient (by disease, disease subtype, clinicians, site, and payer), a formal pathway program became a priority. As part of the VBCC initiative, available pathway programs and options were evaluated by a multi-stakeholder group of faculty, staff, nurses, and administrators with campus and community network representatives over a 2-year period. This was led early on by an oncology surgeon who practiced at both the academic center and in the community and understood the practical needs for a tool to guide standardized care for similar patients. Work initially focused on the development of multimodality clinical care pathways, defined as standardized, evidence-based interdisciplinary care management plans, which would identify an appropriate sequence of clinical interventions, timeframes, milestones, and expected outcomes for a comparable patient group, i.e., by diagnosis or surgical procedure. As the work evolved, however, it became clear that the starting point for these larger multimodality care plans should be pathway programs starting with the specialties of medical, radiation, and surgical oncology.

A formal pathway tool became a priority not only to support the measurable delivery of care across the growing COH enterprise but to bring expert input to the point of clinical care for the growing number and complexity of cancer therapies. The continued growth in new Food and Drug Administration (FDA) hematology–oncology drug approvals since 2000 is a telling example of the need to provide real-time expert decision support across an enterprise to aid the delivery of high-quality, high-value cancer care.

The FDA website for oncology–hematology drugs reports that from 2016 through June of 2020, 230 new drug approvals were issued, up from 22 in 2016 to 58 in 2017, 63 in 2018, 46 in 2019, and 41 in the first half of 2020 [44]. These approvals are also increasingly complex and include not only new drugs but new indications for existing drugs. The approval can be under the expedited review process so that the approval could also change with further data. Approvals can include whether the use is approved in early disease, as neoadjuvant, adjuvant therapy, or extended adjuvant or for advanced disease. The approvals can include use only with specific biomarker results from the patient or their tumor, and sometimes only after biomarker testing by specific laboratories. The approvals often specify use only in specific lines of therapy, after previous general or specific therapies, in combination with specific other drugs or drug categories. Biosimilar drugs are also listed in the FDA approvals with their specific uses referenced to the reference drug [44]. Thus, keeping up with the details of the almost weekly new FDA approvals, which individually can impact one or many kinds of cancers, requires increasingly sophisticated and constantly upgraded decision support pathways not only for clinicians but also for patient education to support their participation in shared decision-making and to ensure appropriate use and documentation requirements for authorizations and payer coverage. Financial toxicity has been identified as another growing barrier to patients receiving appropriate cancer therapies, especially targeted, molecular, and immunotherapy-based treatments [45]. Thus, building the workflow infrastructure to support the use of pathway tools which include increasingly higher cost therapies can also support more timely authorization of treatments for patients and can help to engage early patient assistance, when needed, so patients can receive timely care which improves outcomes [46,47]. Several examples of the growing complexity of high-quality care are described in companion articles on colorectal and non-small cell lung cancer by colleagues in this series [48,49].

The development of surgical pathways was driven by the benefits of standardizing surgical type and processes but also by the benefits of reducing morbidity, mortality, and costs by coordinating education and care processes with the many providers involved in a patient's care before, during, and after surgery [50– 53]. These comprehensive care plans have come to be known as Early Recovery after Surgery (ERAS) pathways [54–57]. At City of Hope, ERAS pathway work began in the early 2000s by the urology faculty with the development and implementation of a successful ERAS pathway for cystectomies showing lower length of stay, complications, and readmission rates that were confirmed by outside groups as well [19,58,59]. This stimulated the development of further urology ERAS pathways as well as ongoing work on ERAS pathways for other surgical subspecialty departments for the COH academic campus and at regional hospitals where COH faculty practice.

Given these many drivers toward standardizing care in all specialties that could provide better care, better outcomes, and potentially lower costs, City of Hope's Value Realization Project identified a Value-Based Core (VBC) team of experts to meet regularly and work with a multidisciplinary physician VBC group team, using the Value Framework projects noted in Figure 1 above, to engage disease leaders in medical oncology, surgery, and radiation oncology along with administrators, information technology (IT) and informatics and EHR experts to evaluate, implement, and iterate oncology pathways for the enterprise. Insights and outcomes from this work follow and add to the work of other academic community network enterprises working together to improve the quality of care for cancer patients [60,61].

#### **4. Medical Oncology, Radiation Oncology, and Hematology Subspecialty Pathways**

#### *4.1. Evaluation and Adoption Process for Pathway Processes and Tool for the Enterprise*

Initial meetings identified medical oncology pathways as the starting point for a formal oncology pathway tool. Available oncology pathway programs, their implementation, and integration abilities were evaluated during regular multi-stakeholder meetings. In 2016, a decision was made to use the VIA Oncology Pathways developed at the University of Pittsburg Medical Center (UPMC) Hillman Cancer Center (now ClinPath by Elsevier and will be referred to as such going forward) which met the criteria of the January 2015 ASCO Policy Statement on Clinical Pathways in Oncology as well as their criteria for high-quality pathway programs shown in Figure 2 [62,63]. The ClinPath pathways were then evaluated and approved by each medical oncology disease team for a 1 January 2017 go live date. The ClinPath system was chosen because it addressed the most common tumor types with readily available evidence summaries for clinician review and prioritized pathways by efficacy, toxicity, and cost [64]. The pathways were being recognized by payers as part of growing value-based payer initiatives [16,17,30–32]. The ClinPath Pathway program also welcomed our faculty to actively participate and, when appropriate, lead pathway disease committees. The regular disease committee meetings, made up of national pathway disease experts and interested users, work to keep pathways updated regularly. In addition, pathways were available for hematology and radiation oncology, with plans for further expansion of pathways for additional solid tumors, hematology, radiation, and possibly surgical pathways. A computer icon tool could be launched from campus and community computers on different EHR platforms for the January 2017 go live with the capability to be later integrated into the planned enterprise wide EHR change to the EPIC system (EPIC systems, Verona WI) in December of 2017.


**Figure 2.** ASCO's criteria for high-quality oncology pathway programs under pathway development, implementation and use, and analytics as described by Zon et al., J Oncol Prac 13:207–210, 2017 [63].

#### *4.2. Initial Medical Oncology ClinPath Adoption Processes and Definitions*

Following the decision in 2016 to implement ClinPath pathways, starting with medical oncology, teaching decks were built by the COH pathway and education teams to share the rationale, clarify the diseases for initial navigation at go live 1 January 2017, and share use and development expectations. On-pathway vs. off-pathway treatment decisions were defined. On-pathway decisions were decisions that adhered to the pathway's decision algorithm for that specific stage of disease. Clinical trials and secondary treatments for alternate patient scenarios were still considered to be on-pathway, as were any decisions to not treat or to take a patient off active treatment. Off-pathway decisions that did not align with pathway recommendations on review were typically driven by new data not yet incorporated into the pathway

(indicated as physician disagrees with pathway choice or a free texted comment to that effect, reports not shown), unique patient presentations, specific patient preferences, and on occasion by physician discretion. It was emphasized that while it was expected that for most cancers, on-pathway choices would be in the range of 80% or more, there would never be an expectation of a specific pathway compliance percentage by doctor or by disease type. It was also recognized that in these times of rapid new drug discovery and molecular targeting, the best therapy for an individual patient may not yet be in the pathway tool. All clinicians were encouraged to join one or more pathway disease committees to participate in the ongoing development process, review the latest evidence, and share best practices.

#### *4.3. Key Milestones and Practical Time Implementations for ClinPath Pathways across the COH Enterprise*

January 2017: Medical Oncology go live for six disease types: Medical oncologists at the academic campus and community sites were trained and instructed to use the pathway tool over 6–8 weeks prior to go live for all new medical oncology therapies including both initial therapy and subsequent therapies in six disease categories: breast, lung, genitourinary (GU), gastrointestinal (GI), gynecologic (GYN), and head and neck. Educational tip sheets and training team personnel were available on an ongoing basis for individual support as well. Other solid tumor types and some hematology diseases were available in the tool but optional for use and not assessed as part of our initial metrics. COH had two EHR systems when the VIA pathways went live: Allscripts on campus and Touch Works in the community. Both were interfaced to pull patient demographic and physician schedule information into the pathway tool. After that, however, clinicians would navigate pathways by entering discrete staging and prognostic or clinical care feature elements to reach the recommended pathway choice or choices. Physicians were required to enter a reason when an off-pathway decision was chosen. After making a therapy choice, clinicians then went back to their EHR to order the chosen therapy. Clinicians were made aware of plans for future EPIC integration, addition of clinical trials into the pathways, and expansion of tracking and reporting for other disease types and specialties. Although formal surveys were not done, clinicians expressed frustrations from the double entry of data from their EHR notes into the pathway system, having to use the system when they were confident and familiar with the appropriate pathway choice as well as skepticism on future upgrades and benefits to their disease programs (personal communications).

2 December 2017: Enterprise-wide transition to Epic: Treatment protocols were built in Beacon by the disease teams based on the regimens used in the previous systems and a review of those in the pathway system before go live. Additional COH-specific regimen preferences and many clinical trials were also built. Oral chemotherapy regimens were not prioritized. Standardized nausea regimens based on NCCN emetogenic levels were added for each regimen as well as hypersensitivity premedication regimens and they were integrated when needed to avoid overuse of steroid medications. All documentation, data entry, and ordering were then done in the EPIC EHR across the enterprise.

January 2018: OnCore clinical trial management system went live. Its benefit to the pathway program is discussed below.

August 2018 and ongoing: Clinical trials were formally added into the ClinPath pathway tool by disease teams starting with the most commonly seen cancers in the community. Additional solid tumor clinical trials have been added and closed trials removed as noted below.

March 2019: Radiation oncology clinicians on campus and community sites went live on ClinPath radiation pathways with integration of COH radiation oncology clinical trials as described below.

20 July 2019: The one-way integration between the Epic EHR and the pathway tool was started and is described below.

August 2020: Hematology pathways had been available in the ClinPath system but their formal evaluation by disease leads began as well as identification of clinical trials to be added to the ClinPath

pathways. As reviews are being completed, trials are being added and formal launch of enterprise-wide use and measurement of ClinPath for the common hematology diagnoses is scheduled to start December 2020 as described below.

6 October 2020: The two-way integration between the EPIC EHR and the ClinPath tool went live in the EHR for 11 disease subtypes (breast, bladder, colorectal, gastroesophageal, melanoma, mesothelioma, ovary, prostate, testicular, small cell lung, and thyroid). With two-way integration, we expect that most if not all discrete staging elements from the EPIC staging forms will be interfaced to auto-populate into the ClinPath tool over time and as elements are added to the EPIC system. Any additional or missing discrete data elements in the EHR system would still have to be entered into the pathway tool to trigger a pathway prompt. At go live, tumor types fell into three categories from this pioneering work: fully mapped tumors (breast, colorectal, gastroesophageal, melanoma, mesothelioma, bladder, ovary, sarcoma, small-cell lung cancer, testicular, thyroid, and uterus), partially mapped tumors (head and neck, prostate), and pending mapped tumors (anal, non-small cell lung, neuroendocrine, pancreatic, renal). When clinicians navigate within the EHR to the ClinPath system, the data elements entered into the EPIC staging forms are then shown on a field next to the ClinPath data entry choices. For fully mapped tumors, one click of an APPLY button automates the entry of the EPIC data into the ClinPath system. The doctor then continues navigating to the pathway choices. For a partially mapped tumor, those elements that are mapped will auto-populate in ClinPath and the others will have to be manually entered. For non-mapped tumors, the EPIC data are there to see but each element has to be manually clicked before continuing the navigation. Whether or not a tumor is fully, partially, or not yet mapped, however, after a pathway choice is made, whether for a clinical trial or a standard regimen, the user is then taken back to the EPIC Beacon regimen for completing the order or ordering a clinical trial team evaluation. This two-way integration will continue to expand by matching additional data elements captured in EPIC to the pathway decision trees to more efficiently guide pathway choices. In addition, the EHR system is working to include data elements that more fully describe a patient's disease status as required by decision support tools. It is expected that ongoing work will continuously reduce duplicate data entry requirements, including for the growing number of actionable genomic results, while ensuring that the primary clinical information remains in the enterprise EHR [65]. A survey of provider experience, time savings (if any), and satisfaction was built and, pending team approval, will be sent to all faculty who use the ClinPath system in November of 2020 to further inform this work.

#### *4.4. Clinical Trials Incorporated into ClinPath System and Integrated with Clinical Trials Management System Can Prompt for Available Trials and Track Adoption*

Studies have shown improved clinical trial assessment and accruals for clinical trials incorporated into pathway tools that are routinely used, including by multisite organizations [66,67]. The ability to add clinical trials available at and through our COH enterprise was another key reason the ClinPath pathway system was chosen in 2016 [68]. Prior to incorporating clinical trials into our pathway tool, though, COH implemented the OnCore clinical trials management system (CTMS) in January 2018. This system provides faculty and clinical research teams with a comprehensive integrated system that supports virtually all aspects of clinical trial offerings including features for managing studies, electronically capturing protocol and patient data, creating custom reports, and supporting financial activities. Having it fully linked into the decision support pathway tool enhances clinician notification of available clinical trial options, especially timely notices of pending, open, on hold, or closed trials to maximize clinical efficiencies.

The OnCore tool integrated with the EHR will hopefully enhance the value of the pathways tool through bi-directional data flows indicating protocol status and patient demographics. Phase I of the project included the decommissioning of the legacy clinical trials management system (MIDAS) and migration of protocol status and data. Phase II of the project linked detailed subject management activities

enabling COH to track patient visits, enhance charge segregation, and payment reconciliation. OnCore was also integrated with City of Hope's regulatory committee management platform, iMEDRIS to provide greater efficiency and data integrity. Phase III, which is going live Q3 of 2020, involves migrating and making available to investigators the IRB-approved versions of the protocol-related documents and informed consent forms in OnCore and decommissioning our current Clinical Trials On Line (CTOL) system, where these documents are currently. This will reduce the number of systems which investigators will need to navigate.

A major goal of an integrated decision support pathway tool is the incorporation of clinical trials prompted by the patient's disease information. After the initial pathway tool launched in January of 2017, training was reinforced, navigations increased but both campus and community faculty were eager to have clinical trials placed into the pathways. The pathway team PharmD (DJ) collected a list of clinical trials from each disease lead, worked with them to place each trial in the appropriate pathway branch points, then submitted the information to the pathway company for inclusion in the tool for all faculty at all COH sites. This process took 4–5 weeks for each disease. The schedule and disease types for availability of the clinical trials in the pathway tool went as follows: GI: 8/2018, 15 trials; Lung: 2/2019, 15 trials; GYN: 5/2019, 5 trials; Breast: 7/2019, 18 trials; GU: 9/2019, 20 trials'; Head and Neck: 4/2020, 12 trials.

Still pending disease types for adding clinical trials in solid tumors are melanoma, sarcoma, and neurologic tumors. With the hematology division now actively engaged in finalizing their adoption of the ClinPath hematology disease pathways, disease leads are working with the pathway PharmD lead to have their clinical trials added into the pathway tool for the planned launch of enterprise-wide use of the hematology pathways in December of 2020. Doctors, APPs, and disease leads have informally but almost universally expressed anticipated value in improving clinical trial identification for their patients as they navigate their therapies in the pathway tool.

Pathway reports show 4.5% of patients by individual medical record number were on a past or present clinical trial from the start of our pathway use in January of 2017 until June of 2020 for the six originally monitored cancers (GI, Lung, Breast, GU, GYN, and Head and Neck). The per quarter percent of patients on a trial varied from 2% to a high of 7–8% in Q4 of 2018 and Q1 of 2019 after the GI and lung cancer trials were added to the pathways. The impact of the full integration of clinical trials into the pathway tool will not be fully assessable until additional programs are matured during 2021–2022. These include the ongoing expansion of clinical trial hubs into regional community sites, which started in 2020 and will cover most of the network regions in the next 2 years. Figure 3 shows that our overall accrual numbers remained similar overall in the community sites and the academic center between 2017 and 2019. The academic accruals include community patients referred to the center who entered a clinical trial. The overall numbers for 2020 reflect only 9 months of data which are promising, given the reported major decreases in clinical trial accruals across the US since the COVID-19 pandemic began in March 2020. As the community network clinical trials programs expand with regional trial directors and trial hubs, plans to open more trials for eligible patients in the community, and the addition of hematology trials for the common hematology diseases frequently seen in the community sites, we expect that the inclusion of clinical trials in the ClinPath pathways will stimulate increased assessments and accruals for community network patients over the next 2 years. The expansion of the discrete data capture in the EHR system and pathway navigation reports will also be fed back regularly to disease teams and the clinical trial leadership so that they can better track the impact of assessments and accruals for clinical trials across the enterprise. The ability to track every patient who goes on a systemic therapy in hematology and medical oncology as well as a radiation oncology treatment plan will allow trial leads to more quickly identify gaps in referrals that can then be addressed. Additional issues are reviewed by colleagues in an accompanying article in this series [69].

**Figure 3.** Clinical trial accrual at Duarte and community sites 2017 through September 2020. Light blue bars are the number of patients accrued to treatment trials at community sites where some trials were available. Dark blue bars are number of patients accrued to treatment trial at the Duarte academic campus.

#### *4.5. Oversight and Insights from ClinPath System Pathways Use in Medical, Hematology, and Radiation Oncology at COH*

The ClinPath pathways program at COH was and is overseen by an interdisciplinary team which continues to meet two or more times per month. This team reviews and directs data analytics, gathers feedback from clinicians and administrators, directs dashboard and interface development, sharing of reports to individual clinicians, site and disease leads as well as administrative and contracting leadership. They oversee and encourage ClinPath disease committee and leadership participation as well as ongoing clinical trial and EHR integration with expansion of pathway adoption across diseases and disciplines. The group reviews monthly pathway compliance rates to identify outliers, leading to further study of clinical issues, patient issues, or individual faculty issues that alter compliance rates, most often because new data have not yet been incorporated or there is not a pathway for an individual's episode of care need. COH faculty's participation and, for some, leadership of disease committees has been essential in supporting timely updates of practice changing information as well as sharing back information and the national perspectives on care standards.

Utilization of the pathways by COH clinicians has increased over time and serves as one of the performance metrics for members in medical oncology. These metrics are reviewed by the leadership (including the chairs, senior medical director of community practices, and regional medical directors of community practices) as well as the COH pathway and the value-based leadership teams to maximize standardization, quality, and value of clinical care across the enterprise.

Clinicians, especially those in the community, report that the availability of the pathways has improved clinician efficiency and time management. With the knowledge that these are standardized pathways, agreed upon by the entire institution, with the added benefit of integrated clinical trials, there is less need to contact disease team academic leader(s) for many opinions, even in these times of increasing availability of newer therapies and clinical trials.

While doctors are navigating new therapy starts consistently in the pathway tool, the COH pathway team was asked to track the correlation between indicating a pathway choice in the tool and ordering

the regimen in the EHR as a quality control measure. A Variation of Care report was designed in 2019 and reports the monthly rates of any difference between what was navigated in the ClinPath system and what was ordered in the EPIC EHR through the Beacon module. The data from May to December of 2019 showed that the difference varied from a low of 0 variations from 871 navigations in 11/2019 to a maximum of 18 variations from 688 (2.6%) navigations in October of 2019. Overall, discrepancies were rarely over 3%. A random chart evaluation noted that these discrepancies occurred when patients or providers changed their treatment plan based on additional work-up, most often when additional disease was found. They then ordered the therapy in the EPIC EHR Beacon treatment orders but did not go back into the ClinPath tool to change the navigation information in the pathway tool. This also happened for rare patient preference for an alternate therapy after the initial ClinPath navigation was done.

Since our ClinPath go live in January of 2017 until early Q3 2020 when data were reported for this manuscript, clinicians had done 28,271 total navigations, including 16,034 navigations from the academic clinicians and 12,190 navigations from the community clinicians, including 47 at our newest Newport Beach site, opened in January of 2020. The percentage of therapies ordered in Beacon that were navigated in the pathway tool by enterprise, community, and campus doctors since implementation is shown in Figure 4. Pathway navigation was higher on the academic campus except for two quarters. The reasons are under further evaluation.

**Figure 4.** Percentage of new EPIC Beacon starts that were navigated in ClinPath by medical oncologists for solid tumors since go live for enterprise overall then by Duarte campus site and all community sites. Note Q3 data were incomplete at time of manuscript data collection in May 2020, so the final% of patients in Q3 who were navigated in ClinPath is likely higher, as is seen in other recent quarters.

Simply having a pathway system does not ensure the use, usefulness, or production of informative analytics. A multidisciplinary pathway team of clinicians, analytic, quality, her, and informatics experts meets every 2 weeks to evaluate pathway use, compliance by disease type and clinician, track low compliance issues to understand if a new therapy not yet incorporated into the pathway or other reason is increasing off-pathway choices, use of clinical trials, and whether the indicated therapy in the pathway system is, in fact, the ordered therapy in the EHR. An enterprise-wide incentive program to capture pathway use and choices for medical oncology was launched in 2020 and final percent navigations in the pathway for all medical oncology therapies in the EHR trends have risen significantly. Disease leads have also requested regular reporting of every patient who meets specific disease, stage, and other criteria for an available trial as to whether or not the trial was considered and if a trial evaluation and ultimate

accrual occurred. Several leads are asking for real-time reports so that they can more proactively reach out to network clinicians to resolve any barriers for patient accrual to an available clinical trial. Ultimately, by tracking specific clinical trial eligibility and accrual by site, disease, and trial type across the enterprise, the trial team can optimize which types of trials are opened at which sites to maximize clinical trial resources.

Having a pathway tool enables the tracking of the percentage of patients who are treated on-pathway (including clinical trials) and off-pathway across the enterprise and by academic vs. community network sites as shown in Figure 5. Academic clinicians consistently see more patients treated off-pathway, which, on reviews, has been explained by the higher number of patients seen where a pathway does not exist for the patient presentation or where a different treatment is recommended due to new molecular findings or emerging data are not yet incorporated into the pathway. After reviewing data, the pathway committee has identified and recommended to the pathway vendor that they add that additional reason as an option in the dropdown list when entering a reason for choosing an off-pathway therapy. Such a new choice for going off-pathway, "new information not yet incorporated into the pathway" could then be tracked for each tumor type and stage during six month intervals over twelve to twenty four month periods to see if and when such a therapy choice is added to the pathway. This could also better inform clinicians about how rapidly new, practice-changing therapies that experts agree should be used become available for prompting in the ClinPath decision support pathway tool.

**Figure 5.** Pathway compliance percentage over time for medical oncology patients navigated in pathway tool (ClinPath) for enterprise, Duarte academic campus, and community sites for new therapy regimens ordered in the EPIC-BEACON EHR from January 2019 through June of 2020. Green bars are on-pathway choices (including clinical trials) and yellow bars are off-pathway choices. Note that the scale of the x-axis representing the numbers of patients is different for each group to show the comparison percentages in one graph. The combined number of patients navigated in Duarte and community sites is reflected in the number totals described on the enterprise graph.

#### *4.6. Pathways for Radiation Oncology Using ClinPath Program*

As with medical oncology pathways, radiation therapy pathways were piloted by the University of Pittsburgh Medical Center (UPMC) group, who led the development of the ClinPath (then called VIA) pathways for radiation oncology based on the same principles used for the development of medical oncology pathways to standardize care for the most common radiation treatments, prioritizing efficacy then toxicity then costs. Studies by their group showed that using the pathways with a peer review process appeared to encourage compliance with clinical pathway recommendations [70]. Specific studies included one showing increased community practice physicians' adoption of hypofractionated regimens for whole breast radiation [71].

This is significant as several studies have shown slow and or poor adoption of evidence-based use of shorter course radiation treatments which have lower costs and shorter treatment times for patients with equivalent outcomes. Hypofractionation for breast cancer is a good example. While the majority of women undergoing radiation therapy for early breast cancer have equal disease control with equal or better cosmetics and a shorter course for patients has been adopted by UK and Canadian clinicians for the majority of patients, it has been much slower to be adopted in the US despite positive long-term outcomes in studies and expanded guidelines for use by the American Society for Radiation Therapy (ASTRO) [72–75]. As in medical oncology and hematology, the growing options for the use of different radiation techniques such as 3D conformal, intensity-modulated radiation therapy (IMRT), proton beam, and other methods of localized radiation therapies along with shorter treatment courses have led to increasing use of radiation therapy pathways to support busy clinicians through their prioritizing efficacy, toxicity, and safety as another component of value-based care [76,77].

As the COH enterprise expanded with many satellite sites delivering convenient local radiation therapy treatments, the radiation oncology department chose to implement and use ClinPath radiation pathways at all sites in March 2019. COH has the largest single institution radiation oncology network in California, comprised of 46 physicians practicing at 20 satellite sites. ClinPath pathways were integrated into the day-to-day clinical practice for 35 radiation oncologists at 16 sites as four outside hospital site doctors are not yet on the system. Radiation oncologists with specific disease interests contributed to the development of some of the pathways and some faculty now co-lead some radiation oncology disease pathway committees. Clinical trial options were also integrated into the pathway options customized to our institution.

The radiation oncologists use the EPIC EHR for all care, except the specific radiation therapy planning, which is done in the radiation-specific modules of MOSAIQ or ARIA EHRs at different sites. The pathways have had a launch button integrated in the EPIC EHR and the radiation oncologists are looking forward to the two-way integration in progress so that the staging in EPIC will be pulled into the pathway system which is on the planning agenda for 2021–2022.

The pathways, besides giving case-by-case feedback regarding treatment options to clinicians, also give detailed recommendations regarding the technical aspects of radiation (i.e., dose constraints, fractionation options), which is useful at the time of actual radiation treatment planning. The pathways cover the most common disease sites and scenarios. There are still complex or rare cases that do not fit into any of the existing pathways, where the radiation oncologist notes "no pathway", and there is an option to add a narrative comment.

In general, the pathways program has been well received. There has been >90% participation rate by COH radiation oncologists in using the pathway tool to capture their treatment decisions, with most physicians using it to indicate 100% of their decisions. Physicians are also sent auto-generated weekly reminders listing any patients where pathways data about their treatment choice were not entered. COH does not require strict adherence to the pathway's preferred choice, but the analytic team gives physicians

feedback which can show clinicians where their choices fall and how they compare to their peers. Onand off-pathway rates are tracked for feedback to individual clinicians, disease, and regional site leads. Pathway compliance remains high after initial launch and has been at or above the generally accepted 80% pathway compliance benchmark, which is an accepted standard in the radiation oncology quality assurance program as defined by the quality assurance review center (QARC), which established an 80% rate based on clinical trial standards [78]. The COH on-pathway rates for the combined enterprise and the separate Duarte and community sites are available but we share in Figure 6 how available additional reporting can further detail the types of therapy or trial as well as who has just completed a therapy plan or who is off therapy by quarter for all patients seen. This same graphic can be constructed separately for Duarte campus radiation therapy, for all as well as for individual sites and by doctor.

**Figure 6.** Individual patients seen by radiation oncologists by the enterprise per quarter since ClinPath pathways were initiated for radiation therapy in Q1 of 2019. The numbers represent the number of individuals with a decision made each quarter who are: on therapy with a non-pathway diagnosis, shown as "other trial" (purple), were on a pathway but went off treatment (yellow) that quarter, on a pathway treatment (green), are off treatment (light blue), are on an off-pathway treatment, (red) or on a clinical trial (dark blue) for an on-pathway disease.

Individual radiation oncologists informally report at leadership meetings that they have found the pathway information regarding various fractionation options helpful. Feedback and questions regarding pathways content are welcomed at all levels but formal feedback from radiation oncologists has not been solicited. COH faculty working at the main medical center and satellites engage via many regularly scheduled virtual meetings both before and exclusively after the COVID-19 pandemic from March 2020 forward. They have weekly peer review of patient new starts, monthly didactic sessions regarding a specific disease site where a COH attending reviews a case and asks residents questions, anyone can contribute, and is extremely well-attended. They also have weekly breast and head and neck tumor boards and include community physicians on disease teams. The entire radiation oncology department

has previously gotten together for a yearly in-person one-day science and clinical care retreat, which will be held virtually during the pandemic. Although no single physician has time to participate in every discussion, the regular meetings are available with technology support. Clinical leaders note that it remains challenging to engage and include new and distant satellites as they continue to be added to the COH network, but this is an identified priority.

The COH radiation oncology department has also created an internal Microsoft Teams messaging system where anyone can share case questions and message the entire department. It has yet to be regularly utilized but doctors are in the process of being trained on the Microsoft Teams functionalities. E-mail, therefore, is still the most utilized form of communication department-wide. Overall, pathway compliance is high for the enterprise, as shown in Figure 6 above. Reasons for off-pathway choices are captured as for other specialties, reported and reviewed by the disease teams. Any recommendations for pathway updates are communicated to the appropriate ClinPath disease committees.

#### *4.7. Additional Multidisciplinary Pathways for VBCC: Geriatric Oncology and Genetic Counseling/Risk Assessment*

Geriatric oncology is an established field pioneered by the late COH faculty member, Dr. Arti Hurria. Dr. Hurria's goal was to bring the principles of geriatric oncology into practice through standards and pathways to improve the care of older adults with cancer. Given that one in three Americans aged 60 years or older will be diagnosed with cancer in their lifetime and the population of older adults is growing rapidly with the aging of the U.S. and worldwide populations, cancer incidence is expected to rise by 67% for individuals aged 65 or older between 2010 and 2030. Although older adults represent the majority of individuals with cancer, they are severely under-represented in clinical trials and research. Hence, the majority of evidence that sets the standards of care for oncology treatment pathways is derived from younger individuals.

To address these unmet needs, COH formed the Cancer and Aging program in 2006, under the leadership of Dr. Hurria, which ultimately led to the formation of the Center in Cancer and Aging (CCA) in 2017. The mission of the CCA is to join investigators from all cancer disciplines to study the biology, treatment, and survivorship issues of older adults with cancer. Over the last decade, COH staff have led over 30 studies in cancer and aging, enrolled more than 5000 patients on studies, disseminated research findings through more than 250 publications, and received competitive grants from the NIH as well as philanthropic funding. CCA is one of the elite centers in cancer and aging in the nation.

With this background work, COH's CCA is now leading the way in expanding pilot pathway projects into community sites, starting with standardizing geriatric assessments (GA), which can help to guide personalized care for our increasingly aging population while achieving their best health outcome. For example, an ongoing implementation study of an APP-led multidisciplinary team telehealth intervention is being conducted in one of COH's 31 community cancer sites. This Antelope Valley regional site is in a remote area of high need and limited resources. Learnings from this pilot will be used to expand geriatric pathways for assessments and care planning across our enterprise.

Colleagues at COH have also examined the multifaceted barriers hindering participation of older adults in cancer trials. They offer strategies to improve the participation of older adults in clinical trials in an accompanying article in this series [79]. It is expected that many of the projects outlined in COH's Value Framework, such as the complete discrete staging and collection of social determinants of health along with geriatric assessments with clinical trial integrations into the pathway navigation tool for systemic therapies in medical oncology, hematology, and radiation oncology, will help to support the piloting of these strategies.

Clinical cancer genomics has been an expertise at COH since the department was established in 1996 which has grown to provide training, research, and consultative services overseen now by the division of

clinical cancer genomics (CCG). A counseling, research, and professional teaching program, internationally recognized for its onsite and telehealth training of healthcare providers worldwide, continues with ongoing professional conferences open to trainees and the COH faculty at all sites. At the academic center, the department provides consultations for individual cancer risk assessments, genetic testing and result interpretation, risk reduction counseling, and clinical trial enrollment. The CCG division is also engaged in the support of COH's enterprise precision medicine initiative to expand somatic and germline testing with associated counseling on results, integration of results into care pathways and into research teams to power new discoveries. Work to standardize these processes through discrete data entry in EPIC on the upgraded staging forms is in progress, along with ways to present clinically actionable results of expanding somatic and germ line genomic testing to clinicians within the data for each patient's episode of cancer throughout their journey. Having relevant clinically actionable data for integration into decision support pathway tools and molecular tumor boards while having the larger dataset easily accessible for analytics and discovery by the research teams is a current priority of the precision medicine, genomics, informatics, value-based care, and disease specialty teams informed by input from faculty at both the academic and community sites.

Genomics pathway work is in the earlier stages of development and will be informed by the expansion of CCG services to the enterprise through academic community faculty and staff teamwork. Pathway work has started with one day per week onsite genetic counseling consultations at one local network site. This has now expanded to one day per week of genetic counseling services being available now by telehealth to patients at all community sites. MD geneticist counseling is offered 2 days per month for the community sites after an initial pilot at one community site. Another pilot pathway for genetic assessments is underway with two breast surgeons, trained and supported by the academic CCG team to provide genetic cancer risk assessment (GCRA) to their breast cancer patients with referral to campus for additional genetic counseling or genetics MD for complicated cases as needed. A breast surgeon at another community site is being trained and will provide GCRA services to her patients as well as to other affected and unaffected patients seen by faculty at that site as well as any external referrals. That surgeon also plans to develop a high-risk screening clinic for all cancer types at that community site. Two additional medical geneticists and two additional genetic counselors are being recruited for our second academic campus site in Orange County to provide precision medicine and clinical cancer genomics initiatives.

Standardizing the collection and validated data entry of discrete multigenerational family histories and incorporating these data into individual patients' cancer or cancer risk data is another project in the planning stages in the value-based framework. It is being worked on in sequential pilots and projects toward enabling our EHR database to capture relevant family and patient histories that can prompt the development of genomics pathways to trigger precision medicine and clinical cancer genomic services as well as personalized therapies within the ClinPath pathway tool. The opportunity to leverage the expertise of the genomics division, working as a team with the community faculty and COH informaticists, holds the promise of developing more pathways for automating risk assessments, genomic counseling, testing, and interpretation as well as clinical trials as part of each patient's comprehensive care plan. Pathways at COH in other disciplines as described in this manuscript are serving as models to build these to better serve patients across our enterprise.

#### *4.8. Expansion of ClinPath Pathways Use for Hematology Diseases*

While our acute leukemia teams are working to develop comprehensive care plans and pathways to include diagnosis, therapy, trials, and follow-up for acute myelogenous leukemia (AML) and acute lymphoblastic leukemia (ALL), which is primarily treated on the campus, they have developed referral trees to ease access to campus leukemia experts and their pathways by community faculty and outside

providers. For the other major hematology diseases, lymphomas, myeloma, myelodysplastic diseases, and benign idiopathic thrombocytopenic purpura (ITP), which are commonly treated in the community and can involve years of sequenced therapies with growing costs, the ClinPath system has expert hematology disease committees and pathways available in the ClinPath tool alongside the medical and radiation oncology pathways.

The growth of value-based contracting, where payers track metrics on compliance with disease pathways for commonly treated diseases, made adding hematology disease team engagement and enterprise-wide clinical use of the ClinPath hematology pathways a priority for Q4 2020. Since many of these diseases are seen and treated by community faculty, the pathway oversight team reviewed and found that 983 navigations had been done for five types of hematology diseases using the ClinPath tool without any requirement or specific training for their use. While a formal survey was not done, members of the pathway team talked to community faculty and several reported that when they have a question, they use the pathway for hematologic diseases to rapidly inform them of the latest therapies and regimen details for patients with these diseases, especially for the most common diagnoses. They report saving more complex cases for individual campus faculty expert discussions, tumor boards, or second opinion referrals. On the campus side, informal discussion with disease experts raised concerns that clinical trials be more widely disseminated as well as nuanced strategies adopted by campus disease experts who meet weekly to standardize their approaches. While academic faculty welcome calls, referrals, and interactions with the growing community network faculty, they felt that hematology care could be enhanced by hematology pathway adoption.

With the March 2020 California-wide COVID-19 pandemic slowdown in onsite clinical care, the value, staging, and pathway leadership engaged the department chair of hematology in demonstrations of the current EHR, ClinPath pathway tool, and pending interfacing capabilities for the hematologic diseases with clinical trials available in the ClinPath system. Five initial academic and community-based disease leads for diffuse large B-cell lymphoma (DLBCL), chronic lymphocytic leukemia (CLL), Hodgkin's disease, multiple myeloma, and immune thrombocytopenia (ITP) were identified and those for chronic myelogenous leukemia (CML), myelodysplastic syndromes (MDS), and the other lymphoma subtypes followed. For each disease or subtype of the lymphomas, a video meeting was scheduled, ClinPath pathways and EHR staging data element capture on the current EHR staging forms were shared, and feedback solicited. Disease leads identified additional discrete data elements that track to pathway prompts and a list of those recommended for addition as part of an EHR staging upgrade program was captured and shared with the staging form upgrade build team. Hematology faculty were signed up on the pathway system, made aware of the disease committee meeting timing, and encouraged to participate in adding COH expertise to the pathway regimen discussions. A major drawing point is the ability to add the clinical trials to the flow sheet nodes in the pathway tool as we did for medical and radiation oncology. The disease leads anticipate that this will expand awareness, screening, and ultimately accrual to clinical trials in hematology. The leads have also requested real-time in-basket messaging for patients with specific disease features who would be eligible for a trial and reporting on whether or not a trial was chosen so they could reach out to their colleagues in real time to gain a deeper understanding and to address any barriers to accruals. Having the expanded datasets of disease data and therapy choices will also better inform disease leads on populations served at various sites to better inform opening of targeted clinical trials. After 9 months of meetings, trainings, and engagement, the hematology pathways in ClinPath will be tracked for use and on-pathway compliance starting in December of 2020 with planned ongoing engagement with disease leads and faculty who see hematology patients and build out of more comprehensive analytic reports.

#### **5. Oncology Surgery Pathways with ERAS and COH Experience**

#### *5.1. ERAS Pathways for Value-Based Surgical Care*

While pathway work in medical oncology expanded during the early 2000s, surgical pathways, referred to as "early recovery after surgery", or ERAS pathways were developed out of early work by Professor Henrik Kehlet in Denmark on multidisciplinary care approaches to standardize colorectal cancer surgical care that improved outcomes [51,56]. An international ERAS study group was formed in 2001 to "develop perioperative care and to improve recovery through research, education, audit and implementation of evidence-based practice." The group later formed an ERAS society in 2010 and a US branch, ERAS@USA in 2016, as data accumulated that orders alone were not enough to improve outcomes but comprehensive approaches to pre, day of, and postoperative (post op) care could improve outcomes in growing numbers of cancer surgeries [80]. The goals of ERAS pathways are to decrease practice variability, lessen morbidity and mortality, and lower costs through education, care coordination, and specific orders before, during, and after cancer surgeries. These are built into care plans on paper or in EHRs to prompt for all agreed upon steps across the surgical journey. Clinical, quality, clinical trial, and financial outcomes can then be analyzed and benchmarked as in other oncology disciplines.

Depending on the surgery, the preoperative (pre-op) processes might include performing an American College of Surgeons developed National Surgical Quality Improvement Program (NSQIP) risk score. This tool predicts the risk of post-op complications "using data from a large number of patients who had a surgical procedure similar to the one the patient may have" [80,81]. This tool is available online and is allowed to be opened from EHRs. Geriatric risk indicators based on data from patients >age 65 were added in 2019 [82]. Using such calculations, for example, post-op nausea and vomiting risks can inform tailored nausea and vomiting prevention education and mediation orders to improve outcomes. The pre-op orders can include breathing training, drain training, smoking cessation, and education on limiting alcohol. They include a review of home medications and any modifications recommended during the pre and postoperative periods. Other goals include improving pain control with limited opioid use through education and standardized orders as well as increasing early mobility to minimize length of stay and enhance the patient experience. The clinical, quality, clinical trial, and financial outcomes of these standardized approaches for specific cancer patients undergoing specific cancer surgeries can then be benchmarked. Teams can use these data for continuous improvement for cancer surgery patients and to better prepare and care for patients at multiple sites. Another key opportunity is standardizing direction of specific high-risk surgeries to high-volume centers where extensive subspecialty surgical expertise has been shown to lower morbidity and mortality and often improve survival across multiple cancer types, systems, and countries despite ongoing controversies on methodologies and the retrospective data reviews [83–90]. Having teams of academic center and community network clinicians working closely together can ease the transitions and the teamwork between academic- and community-based surgical specialists, help to ensure that patients can be easily referred to the academic center for appropriate surgeries per the group pathways while providing continuity and standardization of care using the general and disease-specific ERAS order sets.

City of Hope's urologic oncologic surgeons started adopting ERAS pathways for cystectomy back in 2008 and garnered a large database on outcome benefits published in 2016 [19]. City of Hope's academic-based surgeons have the most substantial experience worldwide in performing robotic cystectomy. These patients were an ideal population to design further care improvements around due to the high-risk patient demographic (elderly, many comorbid conditions), complexity of surgery (5–8-h surgeries mixing together multiple organ systems), and challenges of baseline post-surgery recovery (>80% complication rates, >30% major complication rates, and >30% readmission rates). The evolution of a

pathway that became the cystectomy ERAS pathway started when the campus urologists began using almivopan to assist with bowel recovery as bowel resection is a key recovery factor in cystectomies. The academic urologists oversaw their high volume of surgeries while integrating teams to deliver standardized care at COH's academic specialty cancer hospital. They adopted standard orders for pre-op preparation, operative intervention, and post-op management to facilitate recovery for patients. In collaboration with the supportive medicine department, they developed a multidisciplinary rounding team specifically for cystectomy patients that provided an organized plan of care for each patient daily and kept patients actively involved in understanding and participating in their care. Their pathway details the order components for medical management, symptom management, patient education, supportive care, and case management from the operative day through the day of discharge and specifies the follow-up visit order. Adopting a coordinated care program led to several improvements: the hospital length of stay after surgery decreased from 8 days to 6 days, the 30-day complication rate decreased from 68% to 50%, multiple late-stage patients were primarily referred by urology to hospice (in contrast to prior practice where all poor prognosis patients were referred to medical oncology), and some patients and caregivers reported in Press Ganey surveys that they felt much more informed and empowered in their care [19].

#### *5.2. Expanding Cystectomy ERAS Pathways to Regional Hospitals and for Prostate Cancer Surgeries*

After establishing a successful ERAS program at COH, urology teams rolled out the cystectomy pathway with similar principles to surrounding community hospitals where faculty practiced. Although outcome data show that surgeries of this complexity ideally should be conducted at specialized care centers [59,91], the realities of where our patients reside, their insurance company restrictions, and distributed care delivery necessitate being able to deliver the same standard of care outside COH. Thus, the academic community team faculty partnership built a general ERAS pathway in the order system of the Cerner EHR for a regional hospital where faculty practiced. Over the course of 6 months, they conducted numerous meetings with hospital leadership, nursing leadership, pharmacy, social work, case management, operating room, and nutrition to replicate the model created for the COH campus. At the central academic campus and community sites, teams continue to make iterative implementation improvements using the dedicated pathway functionalities within the EPIC EHR system that incorporates standardized care to reduce variation and cost, organize day-by-day care coordination and documentation, and prompt for discrete capture of outcomes to build robust reporting of outcome and quality measurements.

The urology group next adopted ERAS pathways to drive same day discharges after robotic prostatectomy. Historically, there was a disincentive in fact to discharge these patients the same day as they were considered inpatient only surgeries. However, CMS rule changes have now considered this an outpatient surgery. The group adopted prehabilitation, usage of regional transverse abdominis plane (TAP) anesthesia blocks, early ambulation and feeding, and use of non-narcotic pain medication. This significantly reduced the need for patients to use narcotics, which is beneficial to them individually and in combatting the opioid crisis. Internal data review showed that patients returned home sooner to recover in the comfort of their family and returned to baseline functional and dietary levels much sooner as well.

#### *5.3. Expanding ERAS Pathways for Breast Cancer, Colorectal, GYN, Thoracic, and Other Cancer Surgeries*

Other surgical oncology teams at COH have worked to develop ERAS pathways for on-campus surgeries as well as for many regional hospitals where our faculty work to achieve similar goals of standardizing surgical care, improving outcomes, and lowering healthcare costs while more actively engaging patients [91–94]. In one regional hospital, a generalized ERAS pathway was developed and built into their hospital's EHR to fast-track postoperative mastectomy recovery and minimize narcotic use. At COH, the EPIC EHR has an ERAS breast surgery pathway that incorporates pre-op education, a visit

with occupational therapy, and a tailored medication order set. The educational component involves a one-on-one session with a breast team nurse that is performed during the patient's pre-anesthesia testing visit. Patients receive information on nutrition, exercise, and alcohol/smoking cessation. The information is aimed at preparation for surgery, as well as post-op and long-term recovery strategies. The patients then have an appointment with an occupational therapist to learn about recommended post-op exercises and lymphedema precautions. This "prehabilitation" visit has been found to expedite return to baseline function for patients following breast cancer surgery. Finally, the dedicated breast surgery order set includes medications shown to improve postoperative pain control and minimize intractable postoperative nausea and vomiting [95,96]. The importance of non-narcotic pain control is twofold, as opioids lead to a host of troublesome side effects and COH is focused on combatting the current opioid epidemic [97]. Additionally, the medications given to minimize post-op nausea and vomiting are especially important in breast surgery, as these patients are more likely to develop significant post-op symptoms that prohibit early hospital discharge [98].

The COH ERAS breast surgery pathway can also be tailored to fit an individual surgeon's preferences or patient needs. The COH breast team has an internal (available on request) detailed mastectomy ERAS pathway order set which details the orders on pre-op education, patient preparation, prehabilitation with physical therapy (PT), or occupational therapy (OT) if needed, educate and order specific medicine prescriptions for post-op care before surgery, and ensures that patients understand the post-op caregiving needs before surgery. On the day of surgery, the anesthesia team is actively involved both at campus and at any of the regional hospitals in their part of optimizing enhanced recovery strategies including as appropriate: intraoperative administration of anti-emetics, intravenous (as opposed to inhalational) anesthetic agents, maintenance of euvolemia, and minimization of opioids. These orders have been aligned with studies showing lessened nausea and vomiting and minimizing post-op opioid requirements, which facilitates faster patient recovery and lessens length of stay [99]. Following surgery, post-op visits, nursing care, and care coordination have helped to lower readmission rates while reducing suffering and promoting faster recoveries. We have surgical specialty teams developing, piloting, or using ERAS pathways in the breast, colorectal, thoracic, and gynecologic surgery programs. As they are developed and implemented with structured orders in the EHR from paper orders, the informatics, analytic, and finance teams are getting engaged to develop outcome reports.

Although expanding comprehensive ERAS pathways to every site our patients are seen in remains in process, other departments at COH and various regional hospitals have reported having benefited from involvement in the ERAS pathway work and order set development. Faculty and staff in the anesthesia department were engaged to align perioperative and intraoperative patient management which, once adopted, were embraced for other surgeries. Surgical faculty have reported informally that some outside anesthesia groups have also eagerly participated in ERAS approaches in our community hospital network to improve the care for patients on and off ERAS pathways.

#### *5.4. Challenges with ERAS Pathway Use*

While COH surgical specialists uniformly report that they are very successful in getting together and developing clinical agreement and alignment with each other and with other community oncology surgeons at outside hospitals when needed, full implementation of the pathways for all cancer surgeries at COH campus and in all community hospital sites remains a work in progress. A review of the example of the process steps for our mastectomy ERAS pathway is shown in Table 1 and illustrates the complexity of ERAS pathways. To achieve optimal surgical outcomes for cancer surgery patients, ERAS pathways have shown that many different participants in different locations with specific timing, educational, procedural, medication, or care-related tasks need to be facilitated and remain an active focus of our care

and quality improvement efforts [92–99]. A model template showing the temporal and detailed types of order considerations for ERAS pathways for cancer surgeries demonstrating this is shown in Table 1.


**Table 1.** Detailed timing and order categories to address in ERAS pathways cancer surgeries.

Our ERAS pathway leaders report an easy process in determining alignment for patient and nursing educational needs and process steps and tracking goals for ERAS pathways among team members. Operationalizing multidisciplinary at home, in clinic, in hospital, and post-hospital steps flawlessly, however, remains challenging. Coordinating the different disciplines with staff internally is slightly easier but made more difficult with the many hospital outpatient and freestanding hospital facilities where our patients receive care. Our surgical lead for the value-based care ERAS development reported that the most challenging aspects of implementation have been related to nursing education, engagement, and data tracking. While there was consensus on the content of both preoperative and postoperative education, identifying the resources and a workflow that could be seamlessly incorporated into each patients' journey for different cancer surgeries at different sites and in different EHRs outside of our enterprise EPIC system remains a work in progress. With respect to tracking compliance and patients' clinical outcomes related to the different ERAS pathways, one group plans to pilot a third-party application with an EPIC EHR incorporation. Others are working on various internal EHR tools and the patient portal tools to improve scheduling and prompting of patient education and follow-up care as well as early symptom reporting and triage. In addition to compliance with the pathways, other endpoints being evaluated are length of stay, post-op pain, complications, and survival. Formal measurement of staff satisfaction has not been

done but is discussed at surgical department meetings. Patient satisfaction is measured for their overall care by standardized Press Ganey surveys, which consistently report high satisfaction scores but do not evaluate the impact of ERAS pathways alone.

#### **6. Supportive Care Pathways**

As with other oncology care pathways, those in supportive care have also been shown to be effective in improving care quality and lowering costs based on many trials evaluating care navigation, emotional support for patients and their families with cancer, electronic patient-reported outcomes integrated into care models that address identified symptoms regularly as well as those formalizing goals of care discussions and advanced care planning [20,21,100–105].

COH, like many large academic centers, has a palliative care department focused on clinical care and research. Projects have been done in pain management, regular biopsychosocial assessments, and care coordination as well as in end of life care. The larger enterprise focus on value-based care has worked to engage the many expert providers and research teams to more formally identify the projects in supportive care that are essential in the value-based framework to provide interval care, care navigation, and care after cancer with the goals of improving care quality as well as clinical outcomes while optimizing the patient and staff experience. The expansion of several ongoing or supportive care pilots was underway before the COVID-19 pandemic expanded our use of telehealth for patient visits across the enterprise. Support services at COH had discussed pilots to expand social work, nutrition counseling, and other projects or pilots were in discussion to expand supportive care pathways to all enterprise sites. With the launch of telehealth technology enterprise-wide after March 2020 that included an iterative EPIC EHR integration and expansion of telehealth coverage by payers, some of these projects are being expedited and optimized through hybrid in-person and telehealth service pathways. Highlights of some of the pathways in our supportive care programs are:

Expanding Goals of Care Discussions as part of shared decision-making can help to avoid ineffective care, especially at the end of life, and help patients to achieve their broad life goals. While there are numerous other services provided by the Department of Supportive Care Medicine at COH, we will highlight three distinct efforts: the integrated care service, Hospice in the Village, and a goals of care pilot in poor prognosis patients.

The Integrated Care Service is a dedicated supportive care team comprised of a palliative care hospitalist, advance practice provider, social worker, chaplain, and pharmacist who provides intensive interdisciplinary inpatient and outpatient care to patients with multiple complex care needs, including high symptom burden, advanced disease, and challenging psychosocial concerns. In the last year, the team has cared for approximately 300 medical oncology and hematology inpatients, with over 50 being cared for in a primary capacity. As processes and results are studied, they will serve as potential models to deploy in partnering with community hospitals where COH faculty practice to build unique faculty outside hospital and clinician partnerships to achieve pathways for similar care improvements for patients hospitalized outside of the academic campus hospital.

City of Hope's Hospice in the Village on the Duarte campus has furnished apartment units with full amenities, specially designed for hospice care. Hospice in the Village allows patients to receive hospice care, from their preferred provider, in a private and comfortable home-like environment, when home is not an option. For the last year and a half, overall, 100 patients were cared for in the village, representing over 550 avoided inpatient days. A pilot in a community site for a similar home-based hospice and higher-level care (below acute level) is being explored.

A collaborative hematology, oncology, and supportive care pilot is underway utilizing various means to identify high-risk patients, including supervised machine learning, which triggers a supportive care

pathway that includes patient screening (including assessing patient perception of prognosis), advance care planning, training on communication, structured goals of care documentation, and engages supportive care disciplines based on patient and family identified needs. The structure of the patient and family meeting pathway is shown in Figure 7, with further information available at the COH website [106].

**Figure 7.** Ten-step supportive care patient and family meeting program pathway with expanded details about step 9 in second section of diagram to improve end of life care at COH.

Implementing Advanced Care Planning Pathways across the enterprise is another key component of our value project's care after cancer pillar. Advance care planning is often poorly incorporated in oncology practices, with reported advance directive (AD) rates generally less than 50% [107]. Concerted efforts were made to improve the overall number of ADs in new patients across the enterprise and specifically for patients undergoing hematopoietic stem cell transplantation (HSCT). The Department of Supportive Care Medicine at COH, in collaboration with medical faculty and administrative support, created a patient-centered ACP pathway program.

The first two years (2013 and 2014) broadly focused on all new COH patients. The last two years (2015 and 2016) included a specific focus on patients undergoing HSCT. The primary goal was a completed AD in the electronic medical record before day 0 of transplant. In addition to provider and transplant team engagement, major time points for supportive care integration to facilitate AD completion were identified, including: (1) registration, (2) new patient orientation, (3) the clinical visit when transplant was decided, (4) pre-transplant education class, (5) clinical social work psychosocial assessment visit, and (6) the pre-transplant hospital days. Between 2012 and 2016 at COH, 1784 transplants were performed. For HSCT patients in 2012, baseline AD capture rate before day 0 of transplant was 28.6%. With the institutional AD program, the AD capture rate before day 0 of transplant was 31.6% for 2014, compared with 2012 (odds ratio, 1.17(95% CI, 0.85–1.60); *p* = 0.33).

With both institutional and hematology-specific programs, the AD capture rate before day 0 was 69.5% for 2016, compared to 2014 (odds ratio, 4.30 (95% CI, 3.14–5.91); *p* < 0.001). While the institutional AD program in 2014 insignificantly impacted HSCT AD completion rates, improving the rate of AD completion from 28.6% to 69.5% in HSCT patients required both institutional AD efforts and a targeted pathway program [108].

After this progress, the team's advance care planning efforts returned to focus on building a scalable advance care planning pathway to support the advanced care planning and AD completions in our 31 network clinics. While advance directives in California can be finalized by two witnesses or a notary, at COH, for our predominantly cancer-focused population, we have opted to not solicit witnesses and elected to offer complimentary notarization services for healthcare documents to minimize a major identified barrier to AD completion. Presently, a pilot pathway is under way, leveraging the Prepare for Your Care advance directive developed by the University of California San Francisco (UCSF) and having these advance directives notarized electronically. Several staff advance directives have been successfully completed and initial patient tests are under way to ensure ease of use, feasibility, and adoption. Building the structured steps and documents into the EHR is on the project roadmap so the AD pathway tools can be available enterprise-wide and outcomes can be tracked and reported for targeted improvements. Additional information regarding City of Hope's advance care planning program is available on the website [109].

Pathways to address, prevent, and treat cancer symptoms through standardizing assessments, triage, and treatments is not only a key component of our value framework's management of patients on therapy pillar, it is a key component of ASCO's Quality Oncology Practice Initiative (QOPI) metrics and Center for Medicare and Medicaid Services' (CMS) Merit-Based Incentive Payment System (MIPS) measures, which we track for compliance [110,111]. Standardized use of patient reported outcome tools has also been shown to improve survival as well as relieve suffering [20,112,113].

Phone Triage Pathways: Our EHR regimens are all built to include evidence-based nausea, vomiting, and hypersensitivity medications that can then be obtained pretreatment and discussed at the chemotherapy teaching visit. The medications are based on the NCCN low, minimal, moderate, and high emetogenic risk of the regimen, which automates a major quality metric to ensure appropriate preand post-medications.

Support Screen Pathways: Our management of care pillar for value-based care also uses pathways for phone triage, biopsychosocial, and per visit symptom assessments and care. We implemented the triage pathway as a pilot in 2018 with the campus breast disease team, which was expanded to campus medical oncology clinics in 2019 and is now being prepared for piloting in the community sites using the EPIC phone triage tool with embedded Schmidt Thompson evidence-based assessment pathways and per disease team or site standard operating protocols for triaging patients with cancer symptoms. We are expanding the nationally recognized work of the supportive care faculty's biopsychosocial Support Screen tool with validated questionnaires and integrated support materials or referrals in a pilot at two community sites. Patients, at specific visits or remotely, can respond to questions on iPads in clinic or through our EPIC MyChart portal. Ongoing work that will be informed by the community pilots will work to address the challenges of addressing complex biopsychosocial needs through hybrid models using community resources supplemented by academic experts. Challenges of identifying resources and activating them based on individual patient needs across an enterprise remains an active project to expand these well validated supportive care pathways for biopsychosocial symptom monitoring and interventions across the enterprise.

Oncology Review of System Pathways: We have also built a multidisciplinary community developed oncology review of symptoms (ONC-ROS) questionnaire into a patient-reported outcome questionnaire available to patients through the EPIC MyChart portal before each visit or in clinic within the EPIC rooming questionnaire that medical assistants can complete with patients prior to visits. The 39 questions cover the most common toxicities for cancer patients receiving medical oncology, surgical oncology, and radiation oncology therapies. Community doctors preferred one questionnaire that medical assistants, who help multiple doctors, could be trained and become familiar with for efficient use. The 39 symptoms are divided into systems which require answering only the positive symptoms; then, with one click, the remaining symptoms can be noted as negative. For the nine symptoms that are monitored for the ASCO-QOPI and CMS-MIPS metrics (n/v, diarrhea, constipation, anxiety, depression, pain, fatigue, falls, and shortness of breath), a positive response opens additional questions to assess patient acceptance of that toxicity or desire for interventions that can be addressed at the visit. The questionnaire responses integrate into the EPIC visit note template as the review of systems for clinician review and editing to ensure symptoms are addressed, which has been shown to lessen hospital and ER visits while speeding up complex documentation to support appropriate billing levels and providing mineable data [20]. The use of the ONC ROS after teachings on capturing a pain score and pain plan of care in EPIC after the December 2017 go live numerically improved the capture of the pain plan of care QOPI metric average or the three reporting periods (Fall 2016, Spring and Fall 2017) before our EPIC transition compared to the four reporting periods (Spring and Fall 2018 and Spring and Fall 2019) after the EPIC transition. This occurred in both the campus (64% pre-EPIC to 93% post) and community sites (39% pre-EPIC to 86% post) as shown in Figure 8. COH's quality team, who collect and report the QOPI scores from the EHR, reported that increased symptom-reporting data from use of the ONC ROS questionnaire in patient visit notes as well as having the NCCN nausea risk level medication orders built into the chemotherapy orders in our EPIC EHR were the main reasons for our improved enterprise QOPI scores for the reporting periods after EPIC compared to those before EPIC implementation. As shown in Figure 9, the score improved from 84% pre-EPIC without the OncROS and nausea regimens built into regimens to 87.6% post-EPIC with the use of the OncROS and nausea medications built into chemotherapy regimens for the enterprise. Ongoing work evaluating which patient-reported outcome tool might work with our triage pathways, multidisciplinary pathways, the EPIC EHR, and our diverse patients is in progress toward a standardized PRO tool for all services. Our criteria for choosing an enterprise-wide tool include patient and provider satisfaction for ease of use, efficacy to improve clinical care, and integration of the PRO tool into care pathways to improve

value-based outcomes and ease documentation to support team-based triage communication and provide mineable data to support quality assessments.

**Figure 8.** Pain plan of care compliance for those reporting moderate to severe pain. Blue represents results from the 3 QOPI reporting quarters (Fall 2016, Spring and Fall 2017) from the previous EHRs (Allscripts on campus and TouchWorks in community sites) prior to the EPIC implementation. Orange represents results from the 4 QOPI reporting quarters (Spring and Fall 2018 and Spring and Fall 2019) after the December 2017 EPIC transition for both the campus and community sites.

**Figure 9.** COH enterprise QOPI final certification scores. Blue represents the 3 reporting quarters (Fall 2016, Spring 2017 and Fall 2017) before the -EPIC implementation. Orange represents the 4 QOPI reporting quarters (Spring and Fall 2018 and Spring and Fall 2019) after the EPIC go live December 2017.

#### **7. Other Enterprise Academic Community Teamwork Supporting Personalized Pathways for Value-Based Care Delivery**

Disease team meetings include academic and community doctors and APPs who get to know and develop respect for each other. They discuss the newest science, clinical trials, and care opportunities that are not yet in pathways, pending disease committee reviews and the 6–12-month processes. Disease teams

maintain internal pathway flow diagrams with some modifications embraced before they are formally available in national pathway tools.

New pre-ClinPath pathway update tool for practice changing therapies in development: COH has identified the need for a new tool to capture and share in a standardized monthly update format new practice changing updates agreed to by disease teams. Tracking analytics are being developed concurrently to track the sharing of new practice changing knowledge with the agreement to add to pathways, the availability in the pathway tool, and the uptake by faculty across the enterprise for long-term quality improvement work. Baseline studies in three tumor types are underway to inform this new process.

Tumor boards: These are all now done for tumor types by teleconferencing so the multidisciplinary teams of campus and community faculty actively participate. Some community doctors report that they are submitting at least one case to each tumor board and regularly submit to the tumor board of 1–2 cancers which they may be particularly focused on. They report that the move to tele-tumor boards for all faculty has increased participation and feel that they are sharing and learning more from the entire multidisciplinary team in very open and collegial discussions. This is true for the weekly molecular tumor board as well.

Joint clinics: Many academic faculty have agreed to cover or see patients from time to time in different community clinics and some community doctors spend regular time in academic campus clinics. The academic faculty appreciates the tight knit and efficient teamwork at community sites who evolved to be very efficient in a highly managed care world. Academic experts get to share one-on-one the latest molecular or other discoveries during real-time patient encounters with their community faculty colleagues and observe the use of ClinPath and other pathways at community sites. Friendships and increased collegiality have blossomed, which encourages ongoing sharing that offers every patient access to cutting-edge knowledge. A system to facilitate second opinions between the campus and community faculty has also helped to increase patient peace of mind while allowing them to receive care at the site closest to their home.

Regional, disease team and campus symposia and lectures are shared with all faculty. The medical oncology department holds quarterly half-day symposia where campus and community faculty present the latest science and care pathways as well as sharing clinical trial, value-based care, and programmatic updates. Disease teams also hold retreats to share the latest disease-specific knowledge and clinical trials, which is another helpful format for networking and building of collegial relationships. The pandemic restrictions on in person meetings led the medical oncology department to hold their sixth symposia in October of 2020, where 118 members participated in a 5-h update on science and clinical programs. The feedback was enthusiastic for the format, where there was active engagement and extensive group and interpersonal sharing of ideas in the chat function while members avoided having to travel from family on a weekend.

#### **8. Other Value-Based Care Framework Projects Integrated with the Pathways Programs**

Development of enterprise disease trees is a newer initiative being done to aid further in the standardization process of developing personalized multidisciplinary care plans throughout the enterprise. Community disease and regional leaders are closely involved with academic-based colleagues in building the initial decision tree for breast cancer. Next up in the queue are GI and GU multispecialty decision tree builds. The campus disease trees include locations for complex surgeries or other cancers as well as timing and genomic testing for both somatic and germline mutations. Within these disease trees is an enterprise-wide initiative to extend germline and somatic genomic testing to all patients as well as the integration of the results into the disease pathways. This will be piloted in one site (South Pasadena) close to the academic center and then expand to all sites by 2021. Pulling in the disease-specific pathways

from medical oncology, hematology, radiation oncology, and surgical oncology, along with supportive care pathways, nutrition, genetic counseling, and testing as well as other rehabilitative services, will help to define when in a patient's course of care is provided and in which sites by which experts.

Increased interoperability of EHRs, especially EPIC, is an ongoing project. An advantage of EPIC, which, in CA, is the primary EHR for major university cancer centers (COH and Stanford) as well as the UC system hospitals (UCLA, UCI, UCSD, UC Merced, UCR, and UCSF), as well as the Kaiser Permanente Health System, which serves 40% of California residents, is the ability to review past records, therapies, and diagnoses, which can better inform current care plan development. Continued regulatory pressure to improve the interoperability of all EHR systems and the laboratory, imaging, and diagnostic companies is important to ensure that accurate patient data are available to treating clinicians to facilitate proper care plan development.

Improved functionality of EHRs to capture data more efficiently and support customized care planning is an ongoing focus for most oncologists to empower value-based care initiatives. The goal that we started with at COH, to develop personalized multidisciplinary CARE PLANS for each patient, continues to advance as disease and discipline-specific pathways are implemented, which can be then be made for each patient with our fast and frugal tree approach (described above). Getting to the point where every patient can be given a comprehensive, multidisciplinary care plan that can be navigated and managed remains a top priority for COH and other organizations focused on actualizing all the elements of VBCC.

Staging initiative for discrete data elements engages faculty in defining the data elements needed for decision support and downstream outcomes for each type of cancer. The AJCC has identified components for staging diseases that now include three categories: traditional staging elements, prognostic features, and clinical care features [114]. While the latest, AJCC 8, which became active in the US for staging in January of 2018, expanded these ideas in breast and some other cancers, most cancers do not have these details defined by the AJCC [114]. These are the growing number of discrete disease and patient-specific data components that track to pathway prompting when the patient goals and other comorbidities make it likely that a standard evidence-based pathway will have the best outcome [115]. Capture of these discrete elements facilitates the interfacing of data entry in the main EHR for an organization, then linking those features into decision support tools to minimize duplicate data entry and improve practice efficiencies, as we have pioneered and described above. Our initial 10-month, enterprise-wide 2020 staging initiative showed that 92% of all new patients seen in medical surgical and radiation oncology between January and September 15 of 2020 had required staging data elements entered in the EHR. The next steps for 2021 will be expanding the required data elements for entry and dashboard reporting to clinicians to include elements that are capturable in the EHR that can map to decision support in the pathway tool while we simultaneously expand the tumor types we have full mapping of EHR data elements into the pathway tool and back to therapy orders for medical and radiation oncology, as discussed above.

Oral hemotherapy drugs are built into the EHR as regimens: In preparation for further integration between our EHR and the ClinPath pathway tool, we identified 58 oral cancer drugs which had not been built into our chemotherapy ordering system (Beacon for EPIC). These are included in 96 protocols which will be prompted from the pathway program, so both academic and community medical oncologists then hematologists were recruited, agreed to build specific protocols including the references, associated NCCN nausea level, dosing, visit, lab, and specific drug guidance using a standardized Excel tool. Examples of current protocols were shared at individual or small group teaching sessions held through video visits. The builds for all but the last 25 protocols have been completed, including validation by an oncology Pharm D since April 2020, which was stimulated by the March 2020 COVID-19 pandemic work from home emphasis. We anticipate that the remaining 25 regimens will be built and validated by December of 2020 and have a goal that all of the 96 regimens will be available in the EPIC Beacon ordering system by Q1 2021.

Once completed, we will launch a best practice alert if a clinician orders an oral chemotherapy agent in the stand-alone prescription system of the HER rather than ordering it in the fully integrated chemotherapy order section in the EPIC Beacon system. The advantage of the build in the specific chemotherapy order system is the tie into the pathway, prompting the ability to standardize chemotherapy teaching visits in each order, support for the authorization processes, and integration into our recently launched specialty pharmacy program. With these integrations, any additional coverage needs for high-cost specialty and other cancer therapies can be identified and addressed early on while also facilitating analytic reporting on specific patient tumor types and disease features treated by specific regimens over measurable time periods for clinical and financial outcomes.

Outreach to community hospitals and ambulatory surgery centers is ongoing to share ERAS and other supportive care pathway work and engage diverse teams in meeting the care steps. Working across systems remains a challenge as noted. There is a clear focus, however, on developing these so that, wherever the patient chooses to get their care, the same high-quality care standards will be incorporated into the care processes and embraced by all staff, whether they work for our enterprise, a partnered entity, or a standalone facility. An advantage of our academic community faculty team model is the easy sharing and customization abilities of our clinicians and administrators so that the additional tools and processes developed at our NCI designated comprehensive cancer center can be shared with community-based hospitals and clinicians outside our enterprise.

The COH Enterprise data warehouse collects data elements from all digital platforms used by the enterprise and supports VBCC as well as research and other initiatives. It is overseen by a governance committee with clinical informatics, precision medicine, and information technology experts, who make data available to empower care and insights. From the data warehouse, Tableau dashboards can be created for feedback to clinicians and administrators along with many other informatic formats to process large datasets for study, some of which are shown as figures in this manuscript. The digital framework to turn information into real-world data is described in Figure 10.

**Figure 10.** City of Hope's overriding digital framework to empower value-based care and real-world insights.

#### **9. Challenges of Keeping Pathways Updated and Minimizing Burnout**

Among the many workstreams we have discussed here, two other challenges should be highlighted: the challenge to keep decision support tools up to date in real time and the importance of engaging physicians in the design and efficient implementation of value-based projects as team members to minimize burnout. Our enterprise has recognized and is addressing methods to keep pathway tools up to date and engage clinicians in the many value-based project processes. These challenges are both minimized by the integrated academic community faculty foundation teamwork, which not only gains from the diverse experiences and wisdom of the faculty but can help to reduce the growing burnout rates and earlier retirement events for cancer clinicians. Three components of burnout described by the AMA Steps Forward highlight a culture of wellness, personal resilience, and practice efficiencies as the major areas contributing to burnout [116]. Internal COH faculty surveys identified practice efficiencies as the most important factor contributing to burnout in our faculty as collected and shared at regular faculty meetings. These surveys have prioritized work to address efficiencies and teamwork as we implement VBCC projects with academic and community team inputs.

Implementing multidisciplinary pathways across an enterprise and expanding to outside organizations faces the ongoing challenge of keeping pathway and decision support tools up to date in real time so that clinicians can deliver value-based care across the enterprise. Oncology is fortunate to have so many continuing new discoveries that can improve outcomes, but they require real-time tools and programming to bring practicing changing diagnostic and therapeutic knowledge to clinicians. Doing this with academic disease team leads working with integrated academic community disease teams helps to share the knowledge and build real-time enterprise tool updates until the larger pathway decision tools are updated. This can also improve workflows that can enhance patient and practitioner

satisfaction [116]. It can also help to reduce burnout from extended work hours from individual clinicians having to individually evaluate the copious amounts of detailed new cancer information for the many kinds of cancer patients that they see.

A second challenge is to engage clinicians and staff teams in pathways and value-based initiative work so that the data entry, programming of digital tools, and workflow processes align to minimize their contribution to burnout while empowering evidence-based care delivery. Faculty who are encouraged and rewarded by leadership to participate in their passion for cutting-edge care and clinical research, while contributing to enterprise and professional initiatives, can lessen the sense of loss of meaning in work and a feeling they are a replaceable robot in an assembly line of patient care. Dashboard feedback that engages doctors and their teams with the results of discrete data entry and clinical trial outcomes can fuel pride in their and their care team's work with patients as well as in their role in the success of the larger enterprise. While many clinicians expressed initial push back on the increased workloads to enter discrete staging and pathway navigation data, as familiarity with the systems increased and the integration of the tools as well as the clinical trials expanded, there was a growing sense, which has not been formally measured, that the time spent entering defined, discrete data once is likely offset by time savings from authorization clerks, searching for patient information to inform clinical trial prioritization, expanded efficiencies in database-related clinical research, improved inter-specialty communication of the patient's disease and course, and, as sequences of disease descriptions build across a patient journey, markedly speeding up the ability of clinicians to gain a rapid understanding of what a patient has had, what they were treated with, and where the next steps in diagnostics, follow-up, or subsequent therapies should be. All of these tradeoffs would benefit from formal study that might also help to inform new pilots about return on investments regarding who and when on the team with what training are optimal for specific data entry if not the doctors or APPs [116].

This work can also be informed by many experts outside of medicine as we continue to iterate and learn from pioneering efforts. Dr. Etienne Wenger, PhD is one such globally recognized thought leader in the field of learning theory and its application to business who is a pioneer of "community of practice" research. Especially relevant to our work is his statement that "Communities of practice are important to the functioning of any organization, but they become crucial to those that recognize knowledge as a key asset" [117].

COH as an enterprise is addressing these two challenges along with the many others discussed. The pathways and care plan development programs benefit with leadership supporting active clinician engagement across our academic and community cancer center sites, which addresses several major components to minimize burnout: improving work place efficiencies, reducing social isolation at work, and the sense of loss of meaning in work that care standardization without engagement can bring. This has helped clinicians and staff to embrace the shift to value-based care through personalized precision medicine with integrated clinical research. The addition of teamwork across the enterprise by clinicians, managers, nurses, pharmacists, administrators, quality and informatics experts with our IT teams also iterates improvements in EHR functionalities, allowing more patient-reported input and team-based data entry, which can free up clinician time to analyze data and engage patients in developing a personalized, multidisciplinary care plan. The many projects outlined with multidisciplinary oncology pathways as a key component towards enterprise delivery of high-quality cancer care can be done with thoughtful attention to methodologies and team engagement so that value-based care, defined as the combination of high-quality care, patient satisfaction, provider satisfaction, and enterprise health at the appropriate costs, can be achieved.

#### **10. Conclusions**

Patients, clinicians, patients' families, employers, and payers want to have the best care plan available, discussed and implemented seamlessly with ongoing patient support to achieve the best cancer care outcome, whether this is a cure or improved quality of life. Diagnostic testing, genomics, imaging, and clinical assessments have increasingly complex discrete data elements that need to be transformed into actionable information and decision support. This knowledge also needs to be communicated effectively in the shared decision-making process with patients and payers. Pathway programs with decision support tools integrated into the clinical workflows can inform high-quality clinical trial, treatment, and supportive care decisions efficiently when overseen and supplemented by disease teams informed by integrated academic community faculty partnerships. This knowledge, integrated into EHRs and digital tools, can empower true VBCC throughout a fully integrated, multisite enterprise.

Thus, while many see oncology pathways as focusing only on medical oncology and hematology regimens, they are just as important for specific patient populations such as geriatrics, clinical cancer genetics, AYA, and pediatric populations as well as for surgical oncology, radiation oncology, and supportive care. Combining pathways personalized to each patient can then build the evidence-based comprehensive care plans to optimize each patient's cancer journey for their best health outcome. Foundational work on multidisciplinary oncology pathways, which started with COH's early engagement and leadership with the NCCN, has progressed to the development and implementation of disease-specific, supportive care, subpopulation-specific (i.e., geriatrics), and multidisciplinary pathways toward our goal of comprehensive, individual patient care plans. This work, coordinated under our Value Realization enterprise project, is overseen by faculty and administrative teams with diverse scientific and practical operational knowledge as well as academic and community expertise. The partnered teams have facilitated this pioneering work as an enterprise to develop, implement, and analyze the methodology and infrastructure to care for each cancer patient with a personalized care plan based on their cancer diagnosis, detailed and evolving molecular diagnostics, clinical trials, and bio-psychosocial needs with tools to support engagement in shared decision-making at initial diagnosis and any subsequent recurrences.

As more community oncology practitioners with practical implementation and local cultural expertise join local, regional, and national networks with experts who maintain the intense, detailed focus in subspecialty areas of oncology care, this combined knowledge and perspective can be shared through standardized decision support oncology pathway tools supplemented with real-time, team-led modifications as new knowledge helps us to achieve better outcomes with less toxicity for patients, regardless of where they live. Our COH enterprise commitment to one standard of high-quality care for every patient led to the pilots and multidisciplinary oncology pathway programs using tools, teams, and processes that can serve to inform others as network affiliations grow. While this work is empowered by the team-based, one faculty, one standard of care philosophy across our enterprise, there are many operational and implementation steps to deliver and measure this care standard. The work has benefitted from the combined strength of the academic faculty's deep scientific knowledge and focus with the community faculty's deep clinical and operational knowledge in partnership with administrators, project leads, informaticists, IT experts, and committees.

Implementing and integrating multidisciplinary oncology pathway programs also benefits from other partnered projects with subspecialty groups, IT interfaces, EHR functionalities, care teams, and analytic reporting as reviewed. While multidisciplinary oncology pathways are key components in the delivery of high-quality cancer care, they are only one of the many important component projects of VBCC. Achieving measurable delivery of all component projects within the three VBCC pillars of evidenced-based care, care management, and care after cancer with high levels of patient satisfaction, clinician satisfaction, and

institutional health with optimized costs remains a work in progress. Early reports of COH's successes and the successes reported by others can serve to inform and inspire ongoing work in VBCC [8,11,26,27,30,34].

**Author Contributions:** Conceptualization: L.D.B., B.Y., S.S., H.C., S.G. (Stacy Gray), F.Z., E.K., S.G. (Scott Glaser), R.S., D.J. Formal Analysis: L.D.B., F.Z., D.D., A.B.L., D.M., S.G.(Scott Glaser), J.C., A.N., D.J. Data Curation: L.D.B., B.Y., S.S., A.P., F.Z., D.D., A.B.L., M.S.S., E.K., S.G.(Scott Glaser), D.M., A.N., D.J. Writing—original draft preparation: L.D.B., B.Y., C.Y., H.C., S.S., A.P., F.Z., A.B.L., M.S.S., E.K., S.G. (Stacy Gray), S.G. (Scott Glaser), T.T., A.N., I.B.P., P.K., D.J. Writing—review and editing: L.D.B., M.C., B.Y., C.Y., H.C., S.S., A.P., F.Z., D.D., A.B.L., M.S.S., E.K., S.G. (Stacy Gray), D.M., S.G. (Scott Glaser), G.B., C.A., T.T., J.C., A.N., I.B.P., L.K., P.R., K.S., A.N., P.K., R.S., J.Y., D.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding. The work was conducted as quality improvement projects under institutional programs.

**Conflicts of Interest:** Linda Bosserman serves as an unpaid member and co-chair of the breast cancer pathway committee for Elsevier's Clinical Pathways and serves as an unpaid member of the adult oncology steering board for the EPIC electronic health record. Chao disclosed he has financial relationships in the last 36 months with Merck, Amgen, Macrogenetics, Ono Pharmaceutical, Foundation Medicine and Daiichi-Sankyo. Elisabeth King is a member of a Pfizer advisory board and on AstraZeneca's speaker's bureau. No conflicts were reported by other coauthors related to this manuscript.

#### **References**








© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### **Frontline Management of Epithelial Ovarian Cancer—Combining Clinical Expertise with Community Practice Collaboration and Cutting-Edge Research**

**Edward Wenge Wang 1,\*, Christina Hsiao Wei 2, Sariah Liu 1, Stephen Jae-Jin Lee 3, Susan Shehayeb 4, Scott Glaser 5, Richard Li 5, Siamak Saadat 1, James Shen 1, Thanh Dellinger 3, Ernest Soyoung Han 3, Daphne Stewart 1, Sharon Wilczynski 2, Mihaela Cristea <sup>1</sup> and Lorna Rodriguez-Rodriguez <sup>3</sup>**


Received: 21 July 2020; Accepted: 28 August 2020; Published: 1 September 2020

**Abstract:** Epithelial ovarian cancer (EOC) is the most common histology of ovarian cancer defined as epithelial cancer derived from the ovaries, fallopian tubes, or primary peritoneum. It is the fifth most common cause of cancer-related death in women in the United States. Because of a lack of effective screening and non-specific symptoms, EOC is typically diagnosed at an advanced stage (FIGO stage III or IV) and approximately one third of patients have malignant ascites at initial presentation. The treatment of ovarian cancer consists of a combination of cytoreductive surgery and systemic chemotherapy. Despite the advances with new cytotoxic and targeted therapies, the five-year survival rate for all-stage EOC in the United States is 48.6%. Delivery of up-to-date guideline care and multidisciplinary team efforts are important drivers of overall survival. In this paper, we review our frontline management of EOC that relies on a multi-disciplinary approach drawing on clinical expertise and collaboration combined with community practice and cutting edge clinical and translational research. By optimizing partnerships through team medicine and clinical research, we combine our cancer center clinical expertise, community practice partnership, and clinical and translational research to understand the biology of this deadly disease, advance therapy and connect our patients with the optimal treatment that offers the best possible outcomes.

**Keywords:** epithelial ovarian cancer; frontline treatment; surgical debulking; adjuvant chemotherapy; maintenance therapy; PARP inhibitor; genetics counseling; clinical research; team medicine

#### **1. Introduction**

Epithelial ovarian cancer (EOC) is the most common histology of ovarian cancer, defined as epithelial cancer derived from the ovaries, fallopian tubes, or primary peritoneum [1]. It is the fifth most common cause of cancer-related death in women in the United States, with an estimated 21,750 new cases and 13,940 deaths in 2020 [2]. Because of a lack of effective screening [3] and non-specific symptoms, EOC is typically diagnosed at an advanced stage (FIGO stage III or IV) and approximately one third of patients have malignant ascites at initial presentation. The treatment of ovarian cancer is primarily limited to cytoreductive surgery and systemic chemotherapy. Despite the advances with new cytotoxic and targeted therapies, the five-year survival rate for all-stage EOC in the United States is 48.6% [4]. The delivery of up-to-date guideline care and multidisciplinary team efforts are important drivers of overall survival [5].

The City of Hope National Medical Center (COH) is an NCI-designated Comprehensive Cancer Center based in Duarte, California. Its service area includes Los Angeles, San Bernadino, Riverside, and Orange Counties. Together, these four counties are home to 46% of California's total population. COH delivers high quality cancer care to this sizable demographic through its large network of community oncology practice clinics in the area. In this paper, we review the frontline management of EOC and how we combine our cancer center clinical expertise, community practice partnership, and clinical and translational research to understand the biology of this deadly disease and advance therapy.

#### **2. Surgical Management**

Cytoreductive surgery (debulking) plays a fundamental role in managing EOC. Studies show that survival is inversely correlated with the volume of residual disease after cytoreductive surgery [6–13]. Thus, the goal of surgery is to remove all visible disease [6,9,12,14–18]. In a 2011 meta-analysis of 11 retrospective studies of primary cytoreduction for advanced EOC, there was improved survival with optimal (residual disease ≤ 1 cm in maximum tumor diameter) versus suboptimal (residual disease > 1 cm in maximum tumor diameter) cytoreduction (hazard ratio (HR) 1.36, 95% CI 1.10–1.68), and further improved survival with no gross residual disease (HR 2.20, 95% CI 1.90–2.54) [19]. In a 2013 meta-analysis of 18 studies (retrospective and prospective) of women with stage IIB or higher EOC who underwent cytoreduction and platinum/taxane chemotherapy, each 10% increase in the proportion of patients undergoing complete cytoreduction was associated with a 2.3 month increase in median survival compared with a 1.8 month increase for optimal cytoreduction [14].

Furthermore, improved outcomes in advanced EOC have been shown in high volume hospitals (≥20 cases/year) and high-volume surgeons (≥10 cases/year) [20]. Given the importance of the extent of cytoreduction and volume of cases on outcome and the potential morbidity with an extensive major abdominal surgery, predicting which patients will be able to have at least an optimal cytoreduction is valuable. This is primarily performed through physical examination and computed tomography (CT) of the chest, abdomen, and pelvis. Diagnostic laparoscopy can also be utilized to help triage patients with primary debulking or neoadjuvant chemotherapy [21–23]. It is of utmost importance that a gynecologic oncologist experienced in extensive cytoreductive surgeries evaluates the patient to determine resectability, as achieving no gross residual disease or optimal cytoreduction largely depends on the judgment, experience, skill, and aggressiveness of the surgeon. Additionally, patient factors, such as age, performance status, medical comorbidities, and preoperative nutritional status, are important considerations, as some patients may not be able to tolerate an extensive cytoreduction. The commonly accepted criteria for unresectability include mesenteric root involvement, diffuse involvement of the stomach and/or large parts of the small or large bowel, extra-abdominal disease, infiltration of the duodenum and/or parts of the pancreas (not limited to the pancreatic tail), or involvement of the large vessels of the hepatoduodenal ligament, celiac trunk, or behind the porta hepatis [24].

Our strong partnership with community practices provides a large number of patients in Los Angeles and the Greater Los Angeles area with access to a high volume, high complexity cancer center. In addition to hysterectomy, bilateral salpingo-oophorectomy, and omentectomy, additional procedures can include small bowel resection, large bowel resection, stoma formation, diaphragm peritonectomy plus/minus segmental full-thickness diaphragm resection, splenectomy plus/minus distal pancreatectomy, segmental liver resection, cholecystectomy, partial stomach resection, and partial bladder/ureteral resection. We advocate against routine lymphadenectomy (pelvic, para-aortic) in patients undergoing cytoreduction for stage III or IV disease as it has not been shown to improve overall survival and results in increased postoperative morbidity [25]. However, we do resect suspicious or enlarged lymph nodes to achieve a complete or optimal cytoreduction. An intraperitoneal (IP) catheter for IP delivery of adjuvant chemotherapy may be placed in select patients who have obtained optimal primary cytoreduction, as combination treatment with intravenous (IV) and IP chemotherapy has been shown to prolong overall survival [26–28]; although newer trials have advocated for IV delivery of chemotherapeutics that may have similar outcomes but less morbidity than IP chemotherapy [29].

Patients referred to COH from our community clinics for the surgical management of EOC are assessed by our gynecologic surgical oncologist team and we perform primary cytoreduction for EOC in selected patients (those medically fit to undergo an extensive surgery and in whom it is deemed a resection to no gross residual disease or at least in whom an optimal debulking can be achieved) followed by adjuvant chemotherapy. Other patients deemed unresectable may undergo neoadjuvant chemotherapy and then re-evaluation for possible interval cytoreduction. We perform heated intraperitoneal chemotherapy (HIPEC) in a clinical trial setting for translational purpose toward personalized medicine. We collect biospecimens including peritoneal samples with and without tumor cells, blood samples before and after HIPEC. Paired tumor/normal whole exome sequencing (WES) and whole transcriptome sequencing (RNAseq) is performed for analyses of germline and somatic genomic landscapes, as well as gene expression phenotypes before and after treatment, including the assessment of driver mutations, mutation signatures, tumor mutation burden, and immune signatures. Hyperthermia increases the penetration of chemotherapy and increases the chemosensitivity of the cancer by impairing DNA repair. Additionally, hyperthermia induces apoptosis and activates heat-shock proteins that serve as receptors for natural killer cells, inhibits angiogenesis, and has a direct cytotoxic effect by promoting the denaturation of proteins. In a 2018 randomized trial, van Driel et al. reported a nearly 12-month survival benefit in those receiving HIPEC versus no HIPEC after undergoing at least an optimal interval cytoreduction with a similar rate of grade 3 or 4 adverse events between the two groups [30]. It is unclear if the IP administration, the heat, or the additional dose of chemotherapy is responsible for the benefit as all three interventions were utilized. These results are encouraging; however, further studies are needed before there is widespread adoption of this technique, which requires additional technical expertise [31,32].

Pressurized intraperitoneal aerosol chemotherapy (PIPAC) is another approach we are evaluating in the clinical trial setting. PIPAC is a novel minimally-invasive drug delivery system in which normothermic chemotherapy is administered into the abdominal cavity as an aerosol under pressure [33,34]. This approach uses the advantage of the physical properties of gas and pressure by generating an artificial pressure gradient and enhancing tissue uptake of the aerosolized chemotherapy. Due to high local bioavailability during PIPAC, lower concentrations of chemotherapy can be utilized, thus minimizing side effects and toxicity.

#### **3. Gynecologic Pathology: Diagnostic Evaluation**

Accurate pathologic diagnosis is the cornerstone of our treatment approach. When patients come to COH with a diagnosis of EOC made in the community, their surgical pathology is reviewed by our gynecologic pathology team. There are four major histologic types of ovarian epithelial tumors—serous, mucinous, endometrioid, and clear cell. High grade serous carcinoma (HGSC) is the most common, and lethal histologic subtypes of all ovarian epithelial malignancies are diagnosed, often presenting at an advanced stage. A subset of these patients carry germline mutations in double-strand DNA repair genes, such as BRCA1, BRCA2, RAD51c, and PALB2. Therefore, diagnosis of HGSC carries specific prognostic, therapeutic, and genetic implications. The ovarian cancer TCGA study showed that HGSC is characterized by a near universal p53 mutation [35]. Most of the p53 mutations lead to the overexpression or deletion of the protein, and these can be detected using immunohistochemistry. In morphologically ambiguous cases, performing a p53 mutation analysis may be helpful, and p53 mutation status can be used to temporally track patients' tumors over time. Knowledge about the clinical and functional consequences of various p53 mutations is emerging. We perform whole-exome and RNA sequencing using the next generation sequencing platform for HGSC tumors. This allows us to define the p53 mutation profile in tumors and helps us to better understand clinical and treatment significance.

HGSC also displays genomic instability with high copy-number variations across the genome [36]. This unstable genomic landscape is a collective reflection of high tumor replication rate and the tumor cells' underlying defective DNA repair mechanisms, specifically homologous recombination repair (HRR) [37]. In HGSC, which displays homologous recombination deficiency (HRD), tumors rely on alternative but error-prone pathways, including non-homologous end-joining and single-strand annealing repair pathways [38]. Women with germline BRCA1/2 mutations are enriched for the HRD phenotype [39]. The underlying HRD phenotype explains why some HGSC patients are sensitive to platinum-based chemotherapy (carboplatin, paclitaxel, or docetaxel) or poly-(ADP-ribose)-polymerase 1 (PARP1) inhibitors (such as olaparib and niraparib). Platinum-based chemotherapy induces synthetic lethality by covalent binding with DNA, forming DNA-platinum adducts that eventually trigger double strand break. PARP1 inhibitors impede the PARP1-mediated repair of DNA single strand breaks, a component of the HRR pathway. In HGSC with underlying HRD, double strand breaks cannot be repaired efficiently and their accumulation in the genome result in cell death [38].

HGSC is diagnosed using the MD Anderson histologic 2-tier system [40,41]. Corroborating with the molecular event of p53 mutation, the diagnosis of HGSC can be further supported by performing immunohistochemical staining for p53. HGSC is staged using the current American Joint Committee on Cancer/College of American Pathologists Cancer Staging Form and the FIGO Staging System. The molecular diagnosis of ovarian cancer subtypes that correlate with prognosis may also be adopted as standard procedure in the future. Verhaak et al. analyzed the TCGA database and revealed four ovarian tumor subtypes, each associated with a different prognosis [42].

#### **4. Molecular Studies Available for Diagnostic or Therapeutic Decision Support and Emerging Translational Research**

We perform extensive molecular testing, including whole exome sequencing, transcriptomic sequencing, copy number information, mismatch repair (MMR) deficiency, microsatellite instability (MSI) status, tumor mutation burden (TMB), HRD, and PD-L1 protein expression levels, using paired formalin-fixed paraffin-embedded tumor tissue and patient saliva or peripheral blood. This comprehensive approach allows us to detect somatic and germline mutations, clinically actionable mutations, potential therapeutic targets, and markers to help guide checkpoint inhibitor therapy. The genomic analysis makes tailored therapy possible and informs clinical trial options that best match with patient tumor genotype.

Germline and somatic BRCA1 and BRCA2 mutations are assessed in specific clinical contexts to inform genetic counseling and therapy selection. Younger age at presentation and family history of tubo-ovarian and breast cancer malignancies are risk factors suggestive of the presence of germline cancer predisposition syndrome. Referral to a genetic counselor and establishing germline mutation information is crucial for informing patients about BRCA-related cancer risks for themselves and their family members. Most importantly, this allows patients the opportunity to access BRCA-related cancer risk reduction surgeries (e.g., risk-reducing salpingo-ophorectomy, mastectomy), where the timing of surgery can be crucial to successful risk reduction.

Germline and somatic BRCA1/2 mutation information is also important for informing PARP inhibitor eligibility in Stage II, III, and IV HGSC patients post primary treatment. The NCCN guidelines recommend screening for BRCA mutations early in the treatment course to avoid the possibility of delay in instituting PARP inhibitor therapy [43].

HRD positivity is determined by BRCA mutation status (deleterious or suspected deleterious) or HRD/genomic instability score (mathematically derived from genomic assessment of loss of heterozygosity, telomeric allelic imbalance, and large-scale state transitions). Due to the inherent biocomputational complexity with HRD score derivation and inter-laboratory analytic variability, most large medical centers perform HRD testing on a research basis and not for routine clinical diagnostic use.

Circulating miRNAs in blood and urine are being explored as potential early detection markers. However, the evidence on this approach is currently limited, and no consistent miRNA signatures have emerged [44–46]. The lack of reproducibility may be attributable, in part, to technical issues, such as different statistical modeling and approaches, the utilization of different miRNA detection platforms, and patient and tumor heterogeneity [46]. Besides early detection, liquid biopsy-based circulating tumor cells have been leveraged in a recent small pilot preclinical study to provide chemosensitivity information and therapy response prediction in patients presenting with recurrent ovarian cancer [47]. The quest for providing precision oncology to patients using minimally invasive liquid biopsies is expanding, and hopefully it will become a reality in the not so distant future.

With numerous genomic alterations present in HGSC, an integrative analytical approach is necessary to characterize the dominant biologic drivers of carcinogenesis, cancer progression, and prognosis. The TCGA (Cancer Genome Atlas Research Network) and CPTAC (Clinical Proteomic Tumor Analysis Consortium) investigators have paved the way for combining multiple omics in ovarian HGSC—including genomics, proteomics, and phosphoproteomics. Using transcriptomic data, TCGA has built a HGSC molecular taxonomy comprised of four subtypes: differentiated, mesenchymal, immunoreactive, and proliferative [35]. This framework was recapitulated using the proteomic data [36]. However, this molecular taxonomy does not correlate with patient survival [36]. Instead, proteomic signatures (cytoskeleton involved in invasion and migration, apoptosis, and epithelial junction/adhesion) showed more robust correlation with survival [36]. However, this proteomic signature is currently research-based only, awaiting further validation in larger independent cohorts, and is not currently used in clinical setting.

#### **5. Adjuvant Chemotherapy**

With the exception of patients with early-stage disease and low-grade cancers with a high cure rate, such as stage 1A and 1B grade 1 endometrioid ovarian cancer, mucinous carcinoma, and low grade serous carcinoma [48–50], patients with EOC who have undergone surgical debulking usually require adjuvant platinum- and taxane-based chemotherapy to reduce the risk of recurrence or prolong disease-free survival. Optimal time from surgery to initiate adjuvant chemotherapy has been shown to be 4–6 weeks [49,51]. Table 1 summarizes the main clinical studies of frontline treatment and maintenance of EOC. The standard adjuvant chemotherapy regimen includes: IV paclitaxel 175 mg/m<sup>2</sup> and carboplatin AUC 5–6 every 3 weeks. Alternatively, dose dense weekly paclitaxel 80 mg/m2 and carboplatin AUC 5–6 every 3 weeks may be applied [52–55]—although this regimen has shown differing outcomes in different studies—the JGOG3016 study [52,56] showed a favorable outcome over every 3-week standard regimen, while the ICON-8 [55], and GOG-262 studies [53] failed to showed a significant improvement. The MITO-7 study used weekly paclitaxel 60 mg/m2 and carboplatin AUC 2 for up to 18 weeks—this regimen has a high tolerance and is effective for elderly patients or those with poor performance status [54]. Single agent carboplatin is also acceptable if patients cannot tolerate the combination treatment. Docetaxel is an acceptable taxane alternative to paclitaxel with equivalent efficacy [57]. Carboplatin plus liposomal doxorubicin is also an acceptable combination for adjuvant chemotherapy when patients cannot tolerate taxanes [58,59]. Recently, bevacizumab was incorporated into the adjuvant chemotherapeutic regimen, showing improved progression-free survival and also overall survival in the high risk of progression subgroup, including those with stage IV disease and inoperable or sub-optimally debulked stage III disease (ICON-7, GOG-218) [60,61], especially in patients with ascites [60,62,63].

In patients with EOC, the peritoneal cavity is usually the primary site of recurrence. Thus, the administration of adjuvant IV/IP chemotherapy to treat residual cancer cells with highly concentrated chemotherapeutics is an attractive approach. The GOG-172 study showed that IV paclitaxel 135 mg/m<sup>2</sup> on day 1 plus IP cisplatin 75–100 mg/m2 on day 2 and IP paclitaxel 60 mg/m2 on day 8, every 3 weeks for up to six cycles, improved survival by 16 months in patients with optimally debulked stage III EOC compared with IV delivery of paclitaxel and cisplatin [27]; IP carboplatin is a suitable substitute for IP cisplatin in the GOG-252 study, as the median progression-free survival and overall survival were similar in the IP carboplatin and IP cisplatin arms [28]. However, the IV/IP chemotherapy regimen resulted in more side effects [64], including abdominal pain, catheter-related infection and blockage, and myelosuppression, all of which may delay treatment and compromise efficacy. We routinely use IV/IP adjuvant chemotherapy based on the favorable survival outcomes [27,65]. A recent publication showed that, when bevacizumab was added to IV/IV carboplatin and paclitaxel, IV/IP carboplatin and paclitaxel, or IV/IP cisplatin and paclitaxel, there was no significant difference in progression-free survival in all of these groups of patients [28]. Therefore, there is debate as to whether or not IP chemotherapy is still an acceptable option in primary adjuvant chemotherapy for patients with advanced EOC, given its higher toxicity, inconvenience, catheter complications, and uncertain long-term benefits [29]. At City of Hope, we have been treating patients with the IV/IP protocol. Due to recent advances in maintenance therapy, we are reconsidering if it is still necessary to perform the IP delivery of chemotherapeutics.

#### **6. Maintenance Therapy**

EOC patients who undergo surgical debulking and adjuvant chemotherapy still experience a high rate of disease recurrence. Thus, there is a need for effective maintenance therapy after adjuvant chemotherapy for patients with EOC to help prevent recurrence or prolong disease-free survival. In the past, patients who completed adjuvant chemotherapy usually underwent active surveillance with regular follow-up, labs, and imaging as needed. However, this practice was changed after the ICON-7 and GOG-218 studies showed clinical benefit by adding bevacizumab to the adjuvant chemotherapy regimen [59–62]. The ICON-7 study added bevacizumab (7.5 mg/kg) to IV paclitaxel and carboplatin on day 1, repeated every 3 weeks for 5–6 cycles, continuing bevacizumab for up to 12 additional cycles and showed a modest prolongation of progression-free survival by 2.4 months. Overall, survival was also increased in patients with a poor prognosis [61,66]. The GOG-218 study added bevacizumab to IV paclitaxel and carboplatin on day 1 of cycle 2 (15 mg/kg), every 3 weeks for up to 22 cycles. This regimen showed a significant benefit to progression-free survival (14.1 months vs. 10.3 months, *p* < 0.001). Patients with ascites who received bevacizumab in addition to paclitaxel and carboplatin had significantly improved progression-free survival and overall survival compared to those who received paclitaxel and carboplatin alone [63]. However, maintenance with PARP inhibitors may be favored over bevacizumab due to improved survival.

Following success in treating recurrent EOC, PARP inhibitors have also recently become an attractive choice for maintenance after adjuvant chemotherapy in newly diagnosed EOC patients. Olaparib was FDA-approved (2018) for the maintenance treatment of adult patients with deleterious or suspected deleterious germline or somatic BRCA-mutated advanced EOC who are experiencing a complete or partial response to first-line platinum-based chemotherapy. This is based on the SOLO-1 study [67], a randomized, double-blind, placebo-controlled, multi-center trial that compared the efficacy of olaparib with placebo in patients with BRCA-mutated advanced ovarian, fallopian tube, or primary peritoneal cancer following first-line platinum-based chemotherapy. After a median follow-up of 41 months, the risk of disease progression or death was 70% lower with olaparib than with placebo. In May 2020, the FDA expanded the indication of olaparib to include its combination with bevacizumab for first-line maintenance treatment of adult patients with advanced EOC who have complete or partial response to first-line platinum-based chemotherapy and whose cancers are HRD-positive, defined by either a deleterious or suspected deleterious BRCA mutation and/or genomic instability score. This

recommendation was based on the study by Ray-Coquard et al. [68], which showed that, in patients with advanced EOC receiving first-line standard therapy bevacizumab, the addition of maintenance olaparib provided a significant progression-free survival benefit, which was substantial in patients with HRD-positive tumors (37.2 vs. 17.7 months). Patients with HRD-positivity but without a BRCA mutation also had significantly improved progression-free survival (28.1 vs. 16.6 months).

Niraparib, another PARP inhibitor, was granted approval by the FDA in April 2020 as a first-line maintenance treatment of adult patients with advanced EOC who experienced a complete or partial response to first-line platinum-based chemotherapy, regardless of biomarker status. This recommendation is based on the PRIMA study [69] (Table 1) which showed that patients with newly diagnosed advanced EOC who had a response to platinum-based chemotherapy and received niraparib had significantly longer progression-free survival than those who received placebo (13.8 vs. 8.2 months), regardless of the presence or absence of HRD. We use niraparib for patients without BRCA mutation or HRD, or patients with unknown BCRA/HRD status.

Additional maintenance options are being studied in clinical trials, including new PARP inhibitors, anti-angiogenesis agents, immune checkpoint inhibitors, agents targeting other signal transduction pathways, and new rational combinations. We expect to have improved maintenance options in the future to further reduce recurrence and prolong disease-free survival. Choosing the right maintenance therapy for each patient is highly complex and benefits from multi-disciplinary discussion. At COH, the community oncologists have access to the COH Gynecologic Cancer Tumor Board (discussed further below) to present their challenging cases for in-depth discussion.


### **1.** Major clinical trials on frontline treatment of epithelial ovarian cancer.

**Table**

#### *J. Clin. Med.* **2020**, *9*, 2830

#### **7. Genetic Counseling**

HGSC is a single case indicator for germline genetic testing [70]. Germline genetic testing should be considered both due to the relatively high percentage of hereditary ovarian cancer with some studies estimating that more than 20% is hereditary in etiology [71–73], and due to the potential for treatment implications [74]. Generally, it is preferable for an individual to undergo germline testing as soon as diagnosis occurs [75,76]. This allows ample time to obtain and disclose results, especially in the setting of a patient who may have a guarded prognosis. Urgent testing of BRCA1/2 and other breast cancer genes with high or moderate penetrance by multi-gene panel can currently be performed. While this strategy is often used for women with breast cancer undergoing surgical decision-making, it can also be employed in the gynecologic oncology setting to provide results that may affect eligibility for PARP inhibitors or other therapies in a timely manner.

Germline testing in an affected individual is the most informative strategy and can help clarify risk for relatives. Close female relatives may have increased empiric risk to develop EOC, although older studies may include some families with risk alleles that would be identified by current technology [77,78]. The ascertainment of a multi-generational pedigree allows both for appropriate test selection as well as for proper assessment of family structure and identification of at-risk relatives [79]. Pedigree assessment in the setting of genetic counseling can also facilitate understanding of social relationships between relatives to help develop appropriate strategies to encourage familial communication about risk.

Germline testing for women with EOC at our center typically includes evaluation via a multi-gene panel to include EOC risk genes beyond BRCA1/2, such as the mismatch repair (Lynch syndrome) genes, BRIP1, RAD51C, and RAD51D [71,80]. Beyond informing therapeutic strategy, germline testing in the setting of appropriate counseling can have significant implications for patients and family members. Germline testing can help stratify the risk of developing other cancers and guide the development of appropriate management strategies, especially as the prognosis for EOC improves with better treatment options. For example, patients with Lynch syndrome are at significantly elevated risk to develop colorectal cancer [81] and patients with pathogenic alterations in the BRCA genes are at significantly elevated risk to develop breast cancer [82]. Understanding a patient's risk may help prevent a second primary cancer, especially in the setting of well-controlled ovarian disease or in the setting where the development of a new cancer may interfere with the patient's current treatment.

Germline testing may be even more impactful in terms of implications for relatives. Identifying an ovarian cancer risk allele can allow relatives with the same allele to undergo preventative measures, such as risk-reducing salpingo-oopherectomy, which is especially relevant when screening is not effective. Moreover, in some cases, over-treatment may be avoided in relatives who do not carry the risk allele but who may have otherwise chosen to move forward with preventative measures or screening due to concerns over risk, based on family history. Many genes implicated in EOC in the setting of a monoallelic pathogenic variant also have implications for typically childhood-onset syndromes in the setting of biallelic pathogenic variants. For example, biallelic variants in BRCA1/2, BRIP1, and RAD51C [83–86] are associated with Fanconi anemia and biallelic variants in the mismatch repair genes are associated with Constitutional Mismatch Repair Deficiency syndrome [81]. Thus, individuals contemplating childbearing may also wish to learn their germline status to inform reproductive decisions.

Importantly, negative somatic testing does not obviate the need for germline testing. Reasons for this can include the loss of a germline mutation in the tumor, limited analysis of the tumor genome, and differences in variant calling between somatic and germline laboratories. Conversely, somatic testing may identify variants that are germline in origin [87,88]. Therefore, patients should be counseled about this possibility, and if somatic results are available, they should be reviewed to help inform germline test selection. Other genes may also be included based on clinical suspicion and the evaluation of additional personal and family history. Reevaluation should be considered over time as changes to the family history, as well as advances in the field of cancer genetics, occur [79].

#### **8. Team Medicine: Optimizing Partnerships and Clinical Research**

We have a number of initiatives to ensure the inclusion of our community partners in research, education, and the integration of research-based advances into novel therapeutics by clinical trials. We aim to personalize therapy for patients so our community physicians can recommend improved therapy considerations, including clinical trials beyond the standard of care. One way we achieve this is via comprehensive molecular testing. All EOC patients at COH undergo GEM ExTra® testing (facilitated by TGen, a COH affiliate). This test reports clinically actionable mutations, copy number alterations, transcript variants, and fusions, detected in any gene in patient DNA or RNA. The goal is to uncover true tumor-specific (somatic) alterations by comparing the sequence of the tumor against the paired normal DNA from each patient. The test also includes whole-transcriptome RNA profiling, interrogating the patient's tumor transcriptome for fusions and transcriptional variants known to be relevant to cancer (e.g., EGFR vIII). Each tumor's cancer-specific alterations are then queried against a proprietary knowledge base algorithm to identify potential therapeutic associations. The final report provides the physician with a list of FDA-approved agents that are associated with tumor-specific DNA alterations, as well as biomarker summaries on the variants found and tumor-specific evidence for drug matches, including matches with investigational agents, as available on clinicaltrials.gov. The results are reviewed by our multidisciplinary gynecologic cancer research team to aid in treatment decision-making, highlight on-going studies and identify study candidates.

Our current clinical research portfolio in the frontline management of EOC focuses on developing superior treatment options for patients that reduce recurrence and prolong disease-free survival. We are exploring the use of HIPEC and PIPAC in the clinical trial setting as well as novel drug combinations that help to tailor and personalize treatment for superior results. Our HIPEC trial includes studying the molecular changes triggered by HIPEC to identify molecular signatures of response. Our PIPAC trial is the first in the United States to study aerosolized, pressurized chemotherapy for patients with peritoneal carcinomatosis, including ovarian cancer. Our community oncologists play an important role in these studies by referring patients, thereby allowing us to complete accrual expeditiously.

#### **9. Summary**

Management of EOC requires a multi-disciplinary approach, drawing on clinical expertise and collaboration combined with community practice and cutting edge clinical and translational research. Our goal is to understand the biology of this disease, advance therapy and connect our patients with the optimal treatment that offers the best possible outcomes.

**Author Contributions:** Conceptualization, E.W.W., C.H.W., S.L., S.J.-J.L., S.S. (Susan Shehayeb), S.G., R.L., S.S. (Siamak Saadat), J.S.; validation, E.W.W., M.C. and L.R.-R.; writing—original draft preparation, E.W.W., C.H.W., S.L., S.J.-J.L., S.S. (Susan Shehayeb), S.G., R.L., S.S. (Siamak Saadat), J.S.; writing—review and editing, E.W.W., C.H.W., S.L., S.J.-J.L., S.S. (Susan Shehayeb), T.D., E.S.H., D.S., S.W., M.C. and L.R.-R.; supervision, E.W.W. and L.R.-R.; project administration, E.W.W.; funding acquisition, E.W.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors thank Nicola Welch, CMPP for assistance with writing and editing the manuscript. This work was supported by the National Cancer Institute of the National Institutes of Health under award number K12CA001727. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Small Cell Lung Cancer from Traditional to Innovative Therapeutics: Building a Comprehensive Network to Optimize Clinical and Translational Research**

**Shanmuga Subbiah, Arin Nam, Natasha Garg, Amita Behal, Prakash Kulkarni and Ravi Salgia \***

Department of Medical Oncology and Therapeutics Research, City of Hope National Medical Center, Duarte, CA 91010, USA; ssubbiah@coh.org (S.S.); anam@coh.org (A.N.); ngarg@coh.org (N.G.); abehal@coh.org (A.B.); pkulkarni@coh.org (P.K.)

**\*** Correspondence: rsalgia@coh.org

Received: 21 June 2020; Accepted: 28 July 2020; Published: 30 July 2020

**Abstract:** Small cell lung cancer (SCLC) is an aggressive, complex disease with a distinct biology that contributes to its poor prognosis. Management of SCLC is still widely limited to chemotherapy and radiation therapy, and research recruitment still poses a considerable challenge. Here, we review the current standard of care for SCLC and advances made in utilizing immunotherapy. We also highlight research in the development of targeted therapies and emphasize the importance of a team-based approach to make clinical advances. Building an integrative network between an academic site and community practice sites optimizes biomarker and drug target discovery for managing and treating a difficult disease like SCLC.

**Keywords:** small cell lung cancer; translational research; immunotherapy; clinical trials; team medicine; community practice

#### **1. Introduction**

Lung cancer is the leading cause of cancer deaths in both men and women in the United States [1]. Small cell lung cancer (SCLC) is a subtype of lung cancer that has an incidence of 13% and is strongly associated with smoking [2,3]. A distinct biology, aggressive clinical course with distant metastasis, and poor survival outcomes characterize SCLC. The disease is classified into extensive stage and limited stage. While limited stage SCLC (LS-SCLC) is disease confined to one hemithorax that can be enclosed within a radiation field, extensive stage SCLC (ES-SCLC) is more prevalent (66%) and includes malignant pleural or pericardial effusions along with distant metastasis [4]. Despite the bleak prognosis, standard chemotherapies for patients with SCLC have not changed significantly in the last 30 years. However, recently, immunotherapy with checkpoint inhibitors have shown promising efficacy in advanced disease [5,6]. The lack of advances in SCLC therapies are partly due to disease complexity, research recruitment, and resource utilization. Thus, collaborative efforts between academic and community practices that combine knowledge, skills, experiences, and expertise of academicians, clinicians, and researchers, can accelerate advances in treatment and patient care. Academic centers are critical in the advancement of cancer treatment, but community hospital care plays an equally important and complementary role. Indeed, academic–community collaboration, or 'team medicine', has become an emerging culture to advance and shape clinical care.

In this article, we first highlight the current standard treatments in SCLC as well as recent advances in immunotherapies. We also review potential targeted therapies and underscore the importance of a team-based approach toward SCLC based on our experience at the City of Hope (COH).

#### **2. Current Standard Therapies**

#### *2.1. Radiation Therapy*

For LS-SCLC the standard of care is chemotherapy with concurrent radiation therapy [7]. Two meta-analyses established that concurrent cisplatin and etoposide treatment combined with radiation therapy improves survival compared to chemotherapy alone [8,9], although the dosage (once daily vs. twice daily) of radiation with chemotherapy remains equivocal. One study showed a significant survival advantage of patients who received 1.5 Gy in 30 fractions twice daily compared to 1.8 Gy in 25 fractions after a median follow-up of 8 years [10]. However, a more recent trial that randomized patients to receiving 1.5 Gy twice daily fractions (45 Gy dose) or 2 Gy once daily fractions (66 Gy dose) concurrently with platinum based chemotherapy showed that survival outcomes did not differ between the regimens, although the trial was not powered for equivalence [11]. The ongoing trial of CALGB 30,610 comparing 45 Gy twice daily to 70 Gy once daily (NCT00632853) is likely to shed more light on this issue.

#### *2.2. Chemotherapy*

Regardless of the stage, platinum with etoposide (EP) is the standard of care for patients with SCLC in the United States. Outside the United States, some patients are given platinum with irinotecan as an alternative treatment [12–14]. The overall response rates (ORR) range from 40–70% with up to 10% of the patients achieving complete radiographic response, and the median overall survival (OS) spans 7–12 months, with a two-year survival rate of less than 5% [15]. Eventually, however, most SCLC tumors become resistant to chemotherapy resulting in disease progression. For patients with disease relapse, topotecan as a single agent is the only approved second-line drug that has demonstrated increased survival compared to supportive care [16,17]. Nonetheless, the ORR of patients treated with topotecan is only about 5% [18] and even worse, in SCLC patients who develop disease recurrence within 3 months of the first-line platinum doublet chemotherapy, in which case topotecan is ineffective. Other chemotherapeutic agents such as gemcitabine, docetaxel, paclitaxel, temozolomide, irinotecan, and vinorelbine, may be used in certain cases based on limited clinical evidence [19–24]. Amrubicin is an anthracycline agent that has been developed more recently and approved only in Japan for second-line therapy [25]. Unfortunately, beyond second-line therapy, currently there are no standard guidelines of care although, newer immunotherapies appear promising in some patients.

#### *2.3. Surgery*

Compared to non-small cell lung cancer (NSCLC), SCLC is rarely treated surgically. However, over the years the fraction of SCLC patients treated surgically has increased considerably from 14.9% in 2004 to 28.5% in 2013. This is at least in part due to availability of better diagnostic tools in the form of positron emission tomography (PET) scans and increasing usage of low-dose computed tomography (CT) screening. Randomized trials reported in the late 1960s and early 1970s showed no survival advantage for surgery alone or in combination with radiation therapy compared with radiation therapy alone [26,27]. Subsequently, it was reported that chemotherapy given sequentially with radiation and then randomized to surgery vs. non-surgical group did not show any benefit to surgery [28]. However, recently there have been increasing numbers of retrospective studies showing survival benefit of surgery compared to non-surgical therapy [29–32]. For example, a meta-analysis published by Liu et al. [33], which included two randomized control trials described above and thirteen retrospective studies for a total of 41,483 patients, concluded that surgical resection significantly improved overall survival when compared to non-surgical treatment (hazard ratio (HR) = 0.56, *p* < 0.001) for retrospective studies, and in the two randomized trials there was no survival advantage to the surgical arm. This meta-analysis also showed that lobectomy was associated with superior OS compared with sub lobar resection (HR = 0.64, *p* < 0.001). Based on these data, the National Comprehensive Cancer Network (NCCN) also recommends surgery for T1-2N0M0 SCLC provided preoperative evaluation of mediastinal lymph

nodes are done. Unfortunately, there are no ongoing randomized trials evaluating surgery in SCLC, since less than 5% of patients present with stage I SCLC. However, a collaborative engagement with community clinic sites where majority of cancer patients are seen and academic institutes similar to COH should help accrue enough patients to conduct a prospective trial.

#### **3. Novel Therapies**

Immunotherapy for SCLC was considered viable due to frequent somatic mutations as a result of smoking and the presence of paraneoplastic disorders [34–36]. Furthermore, in light of the remarkable success seen in NSCLC, parallel studies undertaken in SCLC have also shown considerable promise for immunotherapies that include antibodies against programmed cell death protein 1 (PD-1), programmed death-ligand 1 (PD-L1), and cytotoxic T-lymphocyte antigen 4 (CTLA4; Figure 1) [37,38] discussed below.

**Figure 1.** Current investigational immunotherapies and targeted therapies for small cell lung cancer SCLC.

#### *3.1. Atezolizumab*

In treatment-naïve ES-SCLC patients, a recently published a phase III trial involving 403 patients, IMpower-133, combining atezolizumab with carboplatin and etoposide (EP) demonstrated an improved progression-free survival (PFS) as well as overall survival (OS) [39]. More specifically, the patients who did not progress after 4 cycles of induction therapy, received atezolizumab or placebo as maintenance every 3 weeks until disease progression or intolerable toxicity. Median OS for those treated with atezolizumab was 12.3 months compared to 10.3 months for the placebo group, with a hazard ratio for death of 0.70. Median PFS was also improved in the atezolizumab group, which was 5.2 months vs. 4.3 months, with a hazard ratio for disease progression at 0.77, resulting in the approval of atezolizumab with EP for ES-SCLC in the first-line setting. However, blood-based tumor mutational burden (TMB) was not associated with clinical benefit in this study.

#### *3.2. Durvalumab*

Another phase III trial, the CASPIAN trial, which used durvalumab as the immunotherapy in combination with platinum with etoposide to treat treatment-naïve ES-SCLC patients, also showed improvement in OS compared to platinum-etoposide alone (13 months vs. 10.3 months, with a hazard ratio of 0.73) [40]. Based on these results, the Food and Drug Administration (FDA) also approved durvalumab for ES-SCLC.

#### *3.3. Ipilimumab and Nivolumab*

In contrast to atezolizumab or durvalumab, ipilimumab (an anti-cytotoxic T-lymphocyte-associated protein 4 (CTLA4) antibody) in combination with chemotherapy prolongs PFS, but does not improve OS in treatment-naïve ES-SCLC [41]. However, maintenance therapy in such patients with nivolumab/ ipilimumab combination or nivolumab alone did not show improvement in OS, according to results from the phase III CheckMate 451 study presented at the recent European Lung Cancer Congress 2019 [42]. Another trial CheckMate 032 assessed nivolumab as a single agent or in combination with ipilimumab in previously treated SCLC and found that ORR with single agent nivolumab was 11% compared to 22% in the cohort with combination of nivolumab with ipilimumab. The median OS for nivolumab alone was 4.1 months, and for nivolumab with ipilimumab, it was 6 months to 7.8 months based on the doses received [43]. Because the long-term survival benefits with nivolumab alone demonstrated better outcomes compared to previous agents used in the third-line setting, nivolumab received FDA approval for third-line treatment of SCLC.

#### *3.4. Pembrolizumab*

Pembrolizumab was studied in relapsed SCLC patients in the KEYNOTE-028 and KEYNOTE-158 trials. In KEYNOTE-028, the study included only patients with PD-L1 combined positive score (CPS) ≥1%. Among 24 patients with relapsed SCLC, 12.5% were treated with pembrolizumab in the second-line setting and 50% in the third-line. ORR was 33%, median PFS was 1.9 months, one-year PFS was 23.8%, median OS was 9.7 months, and the one-year OS was 37.7% [44]. In the KEYNOTE-158 trial, 79% of 107 patients with relapsed SCLC were treated with pembrolizumab in the second-line or third-line setting. A total of 47% of patients were PD-L1-negative, with an ORR of 18.7% (35.7% in the PD-L1-positive subgroup and 6.0% PD-L1-negative subgroup). The median PFS was 2 months, and median OS was 9.1 months [45]. This led to the approval of pembrolizumab in metastatic SCLC patients whose disease progresses on or after platinum-based chemotherapy and at least one other line of treatment. Considered together, although immunotherapy appears promising for SCLC patients, its benefits are modest, and there is significant room for further improvement.

#### **4. Targeted Therapy**

Unlike in the case of NSCLC, there are currently no targeted therapies available for SCLC. The lack of knowledge of the key genetic mutations and molecular targets that drive SCLC initiation and progress to a more aggressive disease has been a major impediment in developing targeted therapies. However, recent genome-wide studies have identified the universal loss of tumor suppressor genes such as tumor protein 53 (TP53) in 75–90% of patients and retinoblastoma 1 (RB1) and by frequent 3P deletion [46–51]. Consistent with these observations, studies using genetically engineered mouse models have confirmed that the introduction of these two events in pulmonary cells can give rise to high frequency of SCLC development [52]. Nonetheless, more than 120 clinical trials are ongoing that are evaluating new drugs in SCLC targeting various/multiple pathways. Below, we review a few key studies and are depicted in Figure 1.

Aberrant signaling driven by epidermal growth factor receptor (EGFR), stem cell factor receptor tyrosine kinase (c-KIT), PI3K/AKT/mTOR, insulin-like growth factor receptor (IGFR1), and hedgehog signaling pathways have been identified in SCLC. However, inhibitors targeting these pathways have shown minimal efficacy in first-line, maintenance and relapsed SCLC [34,53–70]. Additionally, overexpression and amplification of MET and fibroblast growth factor receptor (FGFR) that are associated with regulating cell proliferation, survival, motility, ability of invasion, and chemoresistance were also observed in SCLC. However, therapies targeting these pathways have not fared well and next generation inhibitors need to be evaluated in combination with either chemotherapy or immunotherapy [71–76].

The apoptotic pathway and the cell cycle checkpoint are some of the other pathways that have been targeted in SCLC. In the former case, B-cell lymphoma 2 (BCL-2) is the favorite therapeutic target. However, BCL-2 antisense oligonucleotide oblimersen and other agents, including obatoclax and navitoclax, have not shown significant activity against SCLC in both phase I and phase II trials [77,78]. Another BCL-2-specific inhibitor venetoclax, demonstrated efficacy in SCLC cell lines expressing high levels of BCL-2 [79] and a phase I trial of venetoclax together with ABBV-075, a bromodomain and extra-terminal domain (BET) inhibitor, is currently under way (NCT02391480). As far as the cell cycle checkpoint is concerned, ataxia telangiectasia and Rad3-related protein (ATR), checkpoint kinase-1 (CHK1), WEE1, and aurora kinase (AURKA), have been the preferred targets among others [80].

In addition to the pathways discussed above, transcription and DNA repair pathways have also been investigated in SCLC. Of these, the MYC pathway stands out since MYC is amplified in a significant (~20%) fraction of SCLC patients and appears to have higher sensitivity to certain newer targeted therapies, such as AURKA and BET inhibitors [81–84]. Other agents that are currently being evaluated are: chiauranib, an aurora B kinase inhibitor for relapsed SCLC (NCT03216343), GSK525762, a BET inhibitor as monotherapy for patients harboring MYC amplification (NCT01587703), and in combination with trametinib for patients carrying RAS mutations (NCT03266159).

Wee-like protein kinase 1 (WEE1) is a key tyrosine kinase involved in halting the G2-to-M phase transition of the cell cycle upon DNA damage [85] that is overexpressed in SCLC [86]. In a preclinical study, the combination of Poly(ADP-Ribose) Polymerase (PARP) inhibitors and WEE1 inhibitors demonstrated a synergistic effect [87]. A phase II, multi-arm trial (BALTIC) is currently evaluating the efficacy of novel therapies in patients with ES-SCLC refractory to platinum-based agents. These novel therapies include PD-L1 inhibitor durvalumab, PARP inhibitor olaparib, and WEE1 kinase inhibitor AZD1775.

Poly (ADP-ribose) polymerase 1 (PARP1) another key player in DNA repair is overexpressed in SCLC [88]. PARP inhibitors prevent cancer cells from repairing DNA damage caused by cytotoxic drugs. Several PARP inhibitors have demonstrated antitumor efficacy in preclinical SCLC models and are currently being studied in several clinical trials. A phase II study investigating veliparib in combination with cisplatin and etoposide in untreated ES-SCLC patients showed improvement in the primary endpoint PFS (6.1 months vs. 5.5 months) although no significant differences in OS were observed [89]. However, Schlafen-11 (SLFN11), which is involved in regulating response to DNA damage and is overexpressed in about 48% of SCLC, has been identified as a potential biomarker for veliparib benefit [90]. Recently, another PARP inhibitor talazoparib also caused higher sensitization to radiotherapy in SCLC cell lines and patient-derived xenografts. Thus, PARP inhibitors have great potential to emerge as a promising therapy for SCLC [91].

Activation of the Notch pathway is oncogenic in some cancer types, but in SCLC, the inhibition of Notch pathway is involved in tumorigenic signaling, progression, and chemoresistance [92]. Consistently, the inhibitory Notch ligand Delta-like protein 3 (DLL3) is upregulated in 85% of SCLCs compared to normal lung [93]. Rovalpituzumab tesirine (Rova-T), a first-in-class DLL3 antibody-drug conjugate, initially exhibited promising results of 18% ORR in heavily pretreated SCLC [94]. Unfortunately, high toxicity rates in the phase II TRINITY trial (NCT02674568) precluded the study from meeting its primary endpoint. In addition, the phase III MERU trial (NCT03033511) evaluating Rova-T in the maintenance setting following first-line chemotherapy, was also terminated early as a result of lack of survival benefit at an interim analysis. Likewise, another phase III (TAHOE) study that assessed Rova-T as a second-line therapy for advanced SCLC compared to topotecan,

stopped enrollment due to shorter OS in the Rova-T group compared with the topotecan control group [95,96]. A phase I/II study evaluating the safety of Rova-T administered in combination with nivolumab or nivolumab and ipilimumab for adults with ES-SCLC has been recently completed and could decide the future of Rova-T (NCT03026166).

#### **5. Protein Phosphatase 2A (PP2A)**

Protein phosphatase 2A (PP2A), a serine/threonine phosphatase that functions as a tumor suppressor in many cancers [97], is also involved in various cellular processes, such as protein synthesis, cellular signaling, cell cycle, apoptosis, metabolism, and stress responses [98]. Several small-molecule activators of PP2A (SMAPs) have emerged as first-in-class agents for this target [99–102]. Further, a recent study has shown that PP2A suppression leads to resistance to kinase inhibitors in KRAS-driven lung cancer cell lines. In KRAS-driven xenograft mouse models, combination treatment of SMAP and selumetinib (a MEK inhibitor used in clinical trials) led to significant tumor regression compared to either agent alone [103,104]. Although PP2A is generally held to have tumor suppressor function, several lines of evidence suggest that it could also function as an oncogene. Thus, small molecule inhibitors of PP2A such as LB-100 [105], are emerging as novel strategies for SCLC. Furthermore, since PP2A is also associated with immune response by downregulating cytotoxic T-lymphocyte function [106,107], PP2A inhibition combined with immunotherapy appears to be effective in mediating an antitumor response. Indeed, in colon cancer and melanoma cells, a combination of LB-100 and an immune checkpoint inhibitor led to greater T-cell-dependent anti-tumor response, with more effector T-cell and reduced regulatory T-cell infiltration [106]. Carbonic anhydrase IX (CAIX), an enzyme involved in hypoxia inducible factor 1 alpha (HIF-1α) hypoxic signaling, is a promising target for chimeric antigen receptor-T (CAR-T) cells in an intracranial mouse model for glioblastoma [108]. In this mouse model, LB-100 was shown to augment the cytotoxic response of anti-carbonic anhydrase (CAIX) CAR-T cells, underscoring the therapeutic potential of synergistic LB-100 and CAR-T cell therapy in SCLC and other solid tumors [109].

#### **6. Mitochondria**

Mitochondria play an essential role in cell survival, apoptosis, adenosine triphosphate (ATP) production, as well as tumorigenesis [110]. Multiple therapeutic strategies have been developed to target mitochondrial functions, such as oxidative phosphorylation, glycolysis, tricarboxylic acid (TCA) cycle, apoptosis, reactive oxygen species (ROS) regulation, permeability transition pore complex, mitochondrial DNA, and dihydroorotate de-hydrogenase (DHODH)-linked pyrimidine synthesis [111]. Drugs that target mitochondrial metabolism through inhibition of pyruvate dehydrogenase (CPI-613, dichloroacetate), isocitrate dehydrogenase (AG-22, AG-120, AG-881) and targeting apoptotic pathways (birinapant, Minnelide, ME-334, Debio 1143, ONC201, LCL161) have been studied in early phase clinical trials [112–117]. These drugs have modest clinical activity as single agents in various tumors including SCLC and are currently being explored as combination therapies with chemotherapy, immunotherapy, and other targeted therapies [118].

#### **7. Stem Cell Therapy**

Cancer stem cells (CSCs) are defined as a small population of cells within a heterogeneous tumor that exhibit similar traits of normal stem cells. CSCs can originate from either somatic stem cells or differentiated progenitor cells. They prompt tumorigenic activity by undergoing self-renewal and differentiation, leading to tumor relapse, resistance to therapy, and metastasis [119–121]. Hence, targeting CSCs has become a novel therapeutic strategy for cancer treatment. CSCs often have upregulated signaling involved in development and tissue homeostasis, such as Notch, Hedgehog and wingless type 1 (WNT) pathways, all of which can be found in SCLC [122,123]. Unfortunately, Notch signaling inhibition with Rova-T and tarextumab have failed, but combination therapies studies are ongoing that could potentially yield positive results. Lysine demethylase 1 (LSD1) is implicated

in maintaining stemness properties and hence, has emerged as a potential target for inhibiting lung CSCs [124]. A phase II trial (CLEPSIDRA) investigating the LSD1 inhibitor iadademstat in combination with standard-of-care in relapsing SCLC patients showed remarkable response rates (up to 75%). Preclinical studies with another LSD1 inhibitor GSK2879552 using 150 cancer cell lines, showed that SCLC and acute myeloid leukemia (AML) cell lines were sensitive to growth inhibition by the LSD1 inhibitor [125–127]. Furthermore, dual inhibition of LSD1 and PD-1 appear to be more effective than either therapy alone [128]. Another route to target stem cells is through CD47 inhibition by RRx-001, which targets tumor-associated macrophages and CSCs via downregulation of the antiphagocytic CD47/SIRPα checkpoint axis. A phase II trial (QUADRUPLE THREAT) involving 26 previously platinum-treated third-line SCLC patients showed that the OS and PFS for patients treated with RRx-001 and a reintroduced platinum doublet were 8.6 months and 7.5 months, respectively, which is much higher for a third-line treatment reported in literature [129]. Interestingly, biopsies taken from patients have correlated response to CD47 inhibition with a high density of infiltrated tumor-associated macrophages that are abundant in SCLC. Based on these observations, a phase III (REPLATINUM) randomized study of RRx-001 with platinum doublet vs. only platinum doublet in third-line SCLC is currently ongoing (NCT03699956).

#### **8. Improving Outcomes**

#### *8.1. Biomarker-Based Therapy*

Although immunotherapy has improved median survival for treatment-naïve ES-SCLC patients, the benefits have been limited with improvement in both PFS and OS approximately 1 and 2 months, respectively, with only 12.6% of patients remaining progression-free after one-year [39]. To improve the outcomes, better selection of patients based on predictive biomarkers, given the high mutational load and rapid resistance in SCLC, and therapies targeting multiple pathways with combination strategies need to be developed.

PD-L1 remains the most common immune-based biomarker for several malignancies. PD-L1 staining in SCLC is less intense and infrequent compared to NSCLC [130]. In KEYNOTE-158, a combined score >1 for PD-L1 expression by the Dako 22C3 assay appeared to predict increased response to pembrolizumab and improved survival when compared with patients negative for PD-L1 [44]; however, the PD-L1 positivity based on Dako 28-8 assay as in the CheckMate 032 did not replicate those results [131]. The assays differ in that KEYNOTE-158 used a PD-L1 score based on staining of tumor cells, lymphocytes and macrophages, whereas CheckMate 032 used staining of only tumor cells to determine positivity. The ongoing phase II REACTION (NCT02580994) and the phase III KEYNOTE-604 (NCT03066778) trials will require measurement of PD-L1 at baseline to provide insight into the predictive role of PD-L1 expression. Higher tumor mutation burden (TMB) has been recognized as a likely predictor of response to immunotherapy across disease types [132]. In an exome-sequencing analysis of CheckMate 032, patients with high TMB appeared to have a greater improvement in OS when treated with nivolumab. Patients with high-, medium-, and low-TMB had a median OS of 5.4 months, 3.9 months, and 3.1 months, respectively; the one-year OS rates were 35.2%, 26.0%, and 22.1%, respectively [133]. However, in the IMpower-133, a blood-based TMB failed to predict benefit for atezolizumab, thus requiring further prospective randomized validation TMB studies. Circulating tumor cells (CTCs) can be detected in 85% of SCLC patients and can potentially serve as a biomarker [134]. CTCs have been explored in multiple studies as a biomarker to predict response and resistance to therapy; however, additional studies looking into the genomic, epigenetic, and transcriptomic heterogeneity of CTCs at diagnosis and during relapse need to be done before it can be applied in clinics [135,136]. Cell-free DNA (cfDNA) widely used in NSCLC has also been examined in SCLC. A study with 27 patients showed cfDNA was able to mirror treatment response and even identified disease recurrence before radiological progression [137]. Future work based on tumor or blood-based biomarkers will help a long way in understanding treatment resistance in SCLC.

#### *8.2. Combination Therapy*

To overcome treatment resistance, novel combination approaches targeting multiple pathways are being explored in combination with chemotherapy and immunotherapies. Lurbinectedin, a novel cytotoxic drug, is a transcriptional inhibitor that inhibits RNA polymerase II. In a phase II trial for both sensitive and resistant disease, lurbinectedin was active as a single agent in second-line SCLC with an ORR of 35.2% [138]. Based on this study, the FDA has approved lurbinectedin for the treatment of ES-SCLC patients with disease progression after platinum-based chemotherapy. A phase III study (ATLANTIS trial, NCT02566993) of lurbinectedin in combination with doxorubicin vs. chemotherapy has completed recruitment and results are pending. Current trials investigating targeted therapies with or without chemotherapy include WEE1 inhibitor AZD1775 in combination with carboplatin (NCT02937818) and olaparib (NCT02511795), checkpoint kinase 1 (CHK1) inhibitor SRA737 in combination with cisplatin/gemcitabine (NCT027979770), ataxia–telangiectasia and Rad3 related (ATR) inhibitor AZD6738 in combination with olaparib (NCT02937818), another ATR inhibitor VX-970 in combination with topotecan (NCT02487095), Bcl-2 inhibitor navitoclax and mTOR inhibitor vistusertib (NCT03366103), Bcl-2 inhibitor venetoclax and ABBV-075 (NCT02391480), and Aurora B kinase inhibitor (NCT02579226) are all ongoing, which should shed some light on the future of targeted therapy in SCLC. There are numerous early phase trials investigating the combination of immunotherapy and targeted drugs as well. Durvalumab with olaparib (NCT02734004, MENDIOLA), avelumab with utomilumab, which is a humanized monoclonal antibody (mAb) that stimulates signaling through CD137 (NCT02554812), nivolumab plus ipilimumab with dendritic cell-based p53 vaccine 9 (NCT03406715) in relapsed SCLC, atezolizumab with chemotherapy and a CDK 4/6 inhibitor trilaciclib (NCT03041311) in first-line ES-SCLC are some of the novel combinations of immunotherapy with targeted agents.

#### *8.3. Other Modalities*

Other modalities of therapies targeting cell surface antigens expressed on tumor cells by monoclonal antibodies or surface-targeting immunotherapies, such as CAR-T cells and bispecific T-cell engagers (BiTEs), arein early stages of development. CD56is expressedin almost all SCLC tumors, and thus, presents to be an attractive target for treating SCLC [139]. In a preclinical study, promiximab-duocarmycin (DUBA), a CD56 antibody conjugated to a potent DNA alkylating agent with a novel linker, showed promising results. This antibody drug conjugate (ADC) demonstrated high efficacy in SCLC xenograft models [140], suggesting that further clinical evaluation of this compound may be beneficial. Another ADC sacituzumab govitecan is comprised of a humanized mAb targeting Trop-2 (trophoblastic antigen-2), which is highly expressed in several epithelial cancers [141], fused to SN-38 (the active metabolite of irinotecan), which induces double- and single-strand DNA breaks by inhibiting topoisomerase I [142]. Sacituzumab govitecan is currently being investigated in several trials, including a phase I/II trial where it is being evaluated as a single agent in patients with previously treated, advanced SCLC (NCT01631552). CAR-T cells targeting DLL3 have entered a phase I clinical trial (NCT03392064). AMG 757, a BiTE, is also being evaluated in a phase I trial that includes ES-SCLC patients requiring first-line maintenance therapy and those with SCLC recurrence (NCT03319940). In patients with metastatic solid tumors including relapsed SCLC, targeting other immune checkpoints, such as PD-1 and CTLA-4, with immunotherapies, including TIM3 and LAG3, are being evaluated in clinical trials in combination with anti-PD-1 or anti-PD-L1 antibodies (NCT03708328, NCT03365791). Finally, radiation therapy is assumed to modulate immune response, as it can increase tumor antigen production and presentation and also enhance cytotoxic T-lymphocyte activity [143]. Potential synergy of radiotherapy in combination with immunotherapy in patients with ES-SCLC is being evaluated in innovative ongoing trials, and results are expected in the near future. Altogether, advancement in biomarkers, targeting multiple critical pathways, and enhancing immunotherapy efficacy in SCLC will hopefully improve the survival outcomes for SCLC patients, which has been elusive for many years.

#### **9. Community Network-City of Hope Experience**

City of Hope is a National Cancer Institute (NCI)-designated Comprehensive Cancer Center and a member of the National Comprehensive Cancer Network (NCCN). In addition, all clinical sites accept the Via Oncology Pathways (modified by City of Hope) for evaluation, antitumor treatment and surveillance after treatments have concluded. City of Hope is composed of a central academic site in Duarte and several satellite sites within Southern California. At the academic center, preclinical work is performed, and clinical trials on that translational research can be rapidly deployed across the entire enterprise, making bench-to-bedside research feasible and fascicle. The collaboration between basic research done on the main campus and clinical research done at both the main and satellite campuses, furthers the discovery of disease biomarkers and novel drug targets. This results in more rational drug design, improved therapeutic efficacy, and quicker optimization of high priority drugs for clinical use (Figure 2).

**Figure 2.** Collaboration between main academic site and community sites for clinical research.

Clinical trials are initiated at the academic campus in Duarte and at one or more of our 27 affiliated network community cancer center offices that are staffed with 43 medical oncologists, 40 radiation oncologists, seven advanced practice providers (APPs) and a clinical trials coordinator. As directed by the Recalcitrant Cancer Research Congressional Act of 2012 (H.R.733), the National Cancer Institute (NCI) allocates resources for research and treatment of recalcitrant cancers having five-year relative survival rates of <20% and estimated to cause at least 30,000 deaths in the US per year. SCLC is considered a recalcitrant cancer having a dismal five-year survival rate of less than 7%. One of the major limitations to ongoing research in SCLC is tumor tissue availability, as the disease is rarely treated surgically. Another issue is many clinical trials in SCLC cannot be completed due to lack of accrual. Using the academic collaboration model with the academic center along with 27 community sites, enrollment becomes more feasible. The rapid progression of disease in SCLC relapse also places research on an urgent timeline to test new agents with a small window to observe treatment efficacy. Given single Institutional Review Board (IRB) approval in our institution, clinical trials can be opened at multiple sites in a rapid fashion. At the academic site, preclinical investigations are done, and clinical trials based on translational research data can be rapidly designed and disseminated across the entire

enterprise, facilitating bench-to-bedside SCLC research in a more feasible manner. The collective knowledge gained from the interaction between the academic and community sites will provide insight into how to overcome challenges that continuously hinder therapeutic advancements in SCLC.

#### **10. Future Directions**

Traditionally, SCLC has been regarded as a homogenous disease, which led to most SCLC patients being treated with essentially one standard regimen. Recent studies from molecular analysis of patient tissues and genetically defined models indicate that there is notable heterogeneity among SCLCs in terms of histology, growth characteristics, expression of neuroendocrine cell differentiation markers, MYC activation, Notch pathway inactivation, and role of neuronal lineage-specific transcription factors in this disease (ASCL1, achaete-scute homologue 1; NeuroD1, neurogenic differentiation factor 1; POU2F3, POU class 2 homeobox 3; YAP1, yes-associated protein 1) [144–148]. Currently, SCLC is classified into four subtypes based on increased expression of different markers: ASCL1 high (SCLC-A), NEUROD1 high (SCLC-N), POU2F3 high (SCLC-P), and YAP1 high (SCLC-Y) [79]. SCLC-A and SCLC-N are neuroendocrine subtypes, whereas SCLC-P and SCLC-Y are non-neuroendocrine subtypes. These subtypes can be associated with specific biomarkers that are either drug-specific targets or predictors of drug response (e.g., DLL3 in SCLC-A, AURKA in SCLC-N, CDK4/6 in SCLC-Y and IGF1R in SCLC-P). These distinct gene expression profiles will guide us in designing new clinical trials. Recent advances in using patient-derived xenograft (PDX) models based on biopsy/resected tumors, CTCs, genetically engineered mouse models (GEMM), as well as omics profiling will drastically enhance our capacity to identify and test novel drugs and discover biomarkers for treatment and prognostication [93,149].

In conclusion, we recommend: (i) setting up a centralized biobank and repository leading to creation of a database incorporating full genomic, proteomic, and microRNA information; (ii) enrolling a higher proportion of SCLC patients into clinical trials with obligatory biomarker analysis; (iii) creating a master protocol which will help reduce duplicative effort and thus ease the eligibility requirements for clinical trials; (iv) create and incentivize academic and community research partnership centers of excellence, since most SCLC patients are treated in community sites; (v) collaborate with bioengineers, cancer biologists, and biophysicists to gather the genetic aberrations discovered and harness the power of computational modeling of genetic information, which will be a powerful tool in understanding SCLC and developing future therapies. Given the academic and community partnership we have established at City of Hope, this should be achievable and pave way for success in treating this challenging disease.

**Author Contributions:** Conceptualization, S.S., A.N., and R.S.; methodology, S.S., A.N., and N.G.; writing—original draft preparation, S.S. and A.N.; writing—review and editing, S.S., A.N., N.G., A.B., P.K., and R.S.; supervision, R.S.; project administration, P.K.; funding acquisition, R.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
