**The Community Oncology and Academic Medical Center Alliance in the Age of Precision Medicine: Cancer Genetics and Genomics Considerations**

### **Marilena Melas 1, Shanmuga Subbiah 2, Siamak Saadat 3, Swapnil Rajurkar <sup>4</sup> and Kevin J. McDonnell 5,6,\***


Received: 13 June 2020; Accepted: 2 July 2020; Published: 6 July 2020

**Abstract:** Recent public policy, governmental regulatory and economic trends have motivated the establishment and deepening of community health and academic medical center alliances. Accordingly, community oncology practices now deliver a significant portion of their oncology care in association with academic cancer centers. In the age of precision medicine, this alliance has acquired critical importance; novel advances in nucleic acid sequencing, the generation and analysis of immense data sets, the changing clinical landscape of hereditary cancer predisposition and ongoing discovery of novel, targeted therapies challenge community-based oncologists to deliver molecularly-informed health care. The active engagement of community oncology practices with academic partners helps with meeting these challenges; community/academic alliances result in improved cancer patient care and provider efficacy. Here, we review the community oncology and academic medical center alliance. We examine how practitioners may leverage academic center precision medicine-based cancer genetics and genomics programs to advance their patients' needs. We highlight a number of project initiatives at the City of Hope Comprehensive Cancer Center that seek to optimize community oncology and academic cancer center precision medicine interactions.

**Keywords:** community oncology; academic cancer center; precision medicine; cancer genetics; cancer genomics

#### **1. Introduction**

Historically, the practice and delivery of healthcare in the community contrasted significantly with medical care provided at the academic medical center [1,2]. These differences manifested across specialty practices, including oncology [3,4]. Rapid advances in molecular diagnostics, the advent of targeted therapies and the introduction of precision medicine amplified differences between community and academic oncology practices [5,6]. Reversing this historical divide, however, new financial realities, public policy initiatives and legislative mandateshave forced community oncologists and academic cancer centers to more closely align their healthcare efforts [7]. This forced alliance has lessened the

separation between community and academic oncology practices and permitted broader access and utilization of precision medicine-based cancer genetics services and tumor genomic analyses. The alliance between community and academic oncology expands the capabilities and effectiveness of the community practitioner, reinforces the mission of the academic cancer center and, ultimately, secures better oncologic care for the cancer patient.

#### **2. The Emergence and Evolution of the Community Health Care and Academic Medical Center Alliance**

A number of key distinctions differentiate the medical care provided at community health centers (CHCs) versus academic health centers (AHCs); these differences result in complementary advantages. The overwhelming majority of patients receive their healthcare through CHCs; the CHC patient population typically exhibits great diversity across economic, racial, ethnic and social spectra [8]. CHCs offer their patients increased accessibility and enhanced client engagement [9]. In contrast, AHCs, characteristically, have focused on specialty medical care, biomedical research, the education and training of health care professionals and the stopgap provision of health care to uninsured and destitute populations [10]. These activities underlie the strengths of AHCs. These strengths include the presence of medical expertise, scientific innovation and clinical trial availability; additionally, AHCs possess unique physical resources such as libraries, computerized database management and informatics infrastructure, research laboratories and emergency room facilities [11,12]. Leveraging these strengths, AHCs have established their reputations and acquired leadership roles in shaping medical care and policy [10].

Until two decades ago, CHCs and AHCs functioned largely in parallel, without administrative or operational intersection. A variety of recent economic, social and regulatory circumstances, however, diminished the independence of AHCs. With the rise of community-based health care markets, particularly managed care plans, many of the operations traditionally carried out at AHCs shifted to CHCs; this shift often undercut the previously reliable revenue streams of AHCs. This situation forced reconsideration of the AHC financial model and provided impetus for the implementation of more efficient, cost-effective health care delivery strategies [13–18]. At the same time, governmental funding agencies, to ensure faithful representation of population diseases, placed a premium on the inclusion of community patients into research protocols. These agencies also issued directives to AHCs to provide comprehensive population care and mandated the formal reporting of AHC involvement with community patient populations [19–25]. Overall, these influences forced AHCs to redefine their core mission with a new emphasis on the integration of the CHC and their patient populations [26–28]. Given their previous work in shaping medical policy, their stewardship of medical education, and their diverse and extensive resources, AHCs readily assumed a leadership role in the restructuring of the CHC/AHC relationship and the creation of integrated partnerships [2,29–35].

The alliance between CHCs and AHCs provides advantages to both partners. CHCs and AHCs enjoy better positioning within the healthcare marketplace. The improved marketplace positioning results primarily from economy of scale pricing that accompanies the integration and expansion of patient services, procedures and therapeutics; the alliance secures for both partners more stable financial footings [36]. The alliance makes possible specific benefits for the CHC. This alliance permits the CHC more direct access to AHC-generated experimental therapeutics, clinical trials, translational research, medical devices and protocols [37]. Further, evidence suggests that affiliation with an AHC often enhances the prestige and attractiveness of the CHC, increases patient and clinical staff retention, fosters more opportunity for continuing professional development, frequently results in greater professional satisfaction and has the potential to enhance the quality and efficacy of the CHC [38–40]. For the AHC, partnerships with a CHC allow for enhanced opportunities to interact more tangibly with the community patient population and expand and diversify patient pools for translational research and clinical trial enrollment; partnerships also increase the ability of AHCs to mitigate outcomes and

patient access disparities [41]. Multiple examples of successful CHC/AHC partnerships exist; they serve as models for the feasibility and potential future CHC/AHC partnerships [42–44].

#### **3. Community Oncology and Academic Cancer Center Alliance**

The integration of CHCs with AHCs most tangibly manifests as practice changes within specific departments, including, prominently, medical oncology [45–51]. During recent years cancer care has transitioned from primarily private, CHC-based oncology practices to AHC-affiliated and -integrated network cancer centers [51–55]. This transition has advantaged the community cancer patient as the services associated with the academic cancer center provide added value.

At the City of Hope Comprehensive Cancer Center (COHCCC), patients identify a number of key value elements associated with the academic cancer center including access to cancer disease specialists, the availability of clinical, translational and basic science researchers, potential for clinical trial participation and enhanced comprehensive care coordinated through multidisciplinary clinical teams [56].

Across a broad range of cancers, patients experience improved survival when receiving treatment at an academic cancer center or at a community hospital associated with an AHC [57–64]. Academic cancer centers provide additional value to community practices through the discovery and provision of novel drugs, experimental medical devices, treatment protocols and technological advancements [65–71]. Reciprocally, academic cancer centers benefit from their alliance with community oncology practices by expanding clinical trial portfolios [72–74], increasing patient diversity in cancer translational and basic research initiatives [75–79], enhancing cancer center core mission accomplishment through community cancer patient engagement [80] and reducing cancer care costs resulting from increased patient volumes [81].

The introduction of new technologies and scientific techniques underscores the importance and potential of the alliance between community oncology practices and academic cancer centers. Specifically, recent advances in genetics and tumor genomics have provided a foundation for the emergence of precision oncology; the community/academic oncology alliance promises to accelerate significantly the clinical utility of precision oncology for the cancer care of community patients [82–85].

#### **4. The Age of Precision Oncology**

Cancers exhibit highly complex genomic and epigenomic alterations; these alterations dictate their overall phenotypic behavior that includes growth characteristics, metastatic potential, interplay between cells and microenvironmental interactions and responses. Over the past several decades, scientific strategies to prevent, diagnose and treat cancer have radically shifted from histology-based to genomically- and immunologically-informed approaches [86].

Since completion of the Human Genome Project in 2003 [87], a series of convergent technological advances resulted from academic-based initiatives. These advances include the introduction and adoption of next generation nucleic acid sequencing (NGS), exponential improvements in computer hardware capabilities, optimization of data processing approaches, evolution of increasingly sophisticated computational biological methods and the discovery and utilization of targeted cancer therapies. Together these advances made possible precision medicine and, more exactly, precision oncology [82,88–90].

NGS arose from innovative DNA sequencing methodologies, most notably massively parallel signature sequencing [91,92]. NGS permits tractable high throughput sequencing of immensely large and complex DNA samples such as whole human exomes and genomes [93,94]. Geneticists first employed NGS to sequence accurately and rapidly the human germline genome [95], allowing insights into the cause of inherited disease [96,97]; investigators then extended the technology to sequence somatic cancer genomes [98]. Scientists further refined the applications of NGS technology. New applications permitted assessment of not only single nucleotide variation and nucleotide insertions and deletions, but also the transcriptome to assess gene expression [99–101], copy number variation [102], complex genomic structural variation [103], protein-DNA interactions [104], targetable epigenetic alterations [105] and epigenetic mechanisms regulating 3D genome structure [106].

In addition to examining tumor genomics, there arose an interest in understanding the immune profiles of the tumor and its microenvironment using NGS; in part, this interest developed from the recognition that tumor genomic changes frequently result in the production of unique, highly immunogenic neoantigens that render the tumor vulnerable to immune surveillance and destruction [107,108]. With the appreciation that the immune system plays an important role in cancer initiation and progression, there has also occurred new interest in targeted therapies aimed at activation of the immune axis [109,110].

NGS generates enormous caches of data; use of these immense data sets for precision oncology requires ever increasing levels of computer hardware performance. Employment of Dennard scaling [111] and multicore architectures [112] have sustained exponential increases in computer chip performance [113–115]. Data processing innovations have included parallel algorithm implementation [116] and parallel data computing [117]; such innovations have force multiplied the efficiency and speed of computation. These approaches allow data analysts to keep pace with the ever increasing information workloads of precision oncology [118].

The realization of precision oncology required adoption of computational biological approaches. The creation of computational biology as an independent academic discipline resulted from the complexity and size of biological data sets. In the case of NGS, the sheer number of nucleotides reads, the task of aligning these reads to reference sequences, predicting functional consequences of genomic variation and the translation of these findings into clinically actionable information necessitated computational biological expertise [119–122]. Computational biological analysis now constitutes an integral element of the data workflow in precision oncology [123–126]; effective clinical translation depends inextricably upon the availability of these computational resources [127–130].

In the early 1970 s, Drs. Janet Rowley, Peter Nowell and Alfred Knudson, studying leukemia cell chromosomes under the microscope, suggested that a specific chromosomal translocation that resulted in the formation of the BCR-ABL fusion oncogene caused chronic myelogenous leukemia (CML); this observation established a foundation for clinical cancer genomics [131]. Oncogenic proteins consequently became a focus of therapeutic drug design; targeted therapies aimed to suppress the aberrant functions of these proteins in order to inhibit tumor progression [132,133].

The successful harnessing of precision therapeutics in oncology ultimately relies upon the availability and efficacy of targeted agents. The discovery that imatinib effectively treats CML harboring the BCR-ABL fusion protein [134] led to the drug's FDA approval in 2001 [135], demonstrated the utility of targeted cancer therapy [136,137], kindled enthusiasm for the identification of other genetically vulnerable cancers and their treatments [90,138] and underscored the clinical value and potential of precision oncology [98,139,140]. Since the success of imatinib, the FDA has approved a multitude of additional therapies to target molecularly-altered cancers [141,142].

The clinical provision of precision oncology requires multidisciplinary support [143]; the complexity of this support will become more intense as precision oncology continues to undergo accelerating change [144–146]. AHCs possess the resources and organization to create this support structure; their alliance with CHC oncology practitioners will make precision oncology available to the larger CHC cancer population.

#### **5. The Community Oncology**/**Academic Cancer Center Alliance in Germline Cancer Genetics**

NGS and precision oncology have had a profound effect upon the practice of cancer genetics, including the evaluation and care of community patients with hereditary predisposition to cancer [147–151]. Until recently, genetic testing involved clinical assessment followed by sequential, single gene Sanger sequencing of suspect genes [152–156]. The advent of NGS brought high throughput germline multigene panel [157–161], whole exome [162–167] and whole genome assessment [168–172] to clinical cancer genetics. These platforms provide tremendous benefit to cancer genetics patients both

in community oncology practices and at academic cancer centers; these advantages include increased diagnostic yield, increased speed of testing, optimized testing workflows, decreased expense and the discovery of new cancer-causing genes [173–177]. However, together with advantages, challenges and limitations arise; AHCs have the specialized resources to address these issues.

In accordance with the American College of Medical Genetics and the Association for Molecular Pathology guidelines, variants from clinical genetic testing fall along a spectrum ranging from pathogenic/likely pathogenic to benign/likely benign [178]; variants of uncertain significance (VUS) occur when there exists insufficient information for variant assignment to either the pathogenic or benign categories [179,180]. For pathogenic/likely pathogenic and benign/likely benign variants, genetic providers typically have the ability to communicate clear interpretation of results and to provide consensus health recommendations. As their pathogenicity remains uncertain, VUS challenge health care specialists to formulate and relay unambiguous health care instructions [181–185]; furthermore, VUS frequently cause confusion and anxiety for the patient [186–190]. VUS impose a significant clinical burden. More than one third of NGS-based cancer gene panel tests result in identification of a VUS [191]; whole exome and genome testing generate even greater numbers of VUS [192–195]. Moreover, if a patient belongs to a minority group, for whom genome annotations remain less well confirmed, VUS additionally increase [196].

Geneticists classify genes according to their penetrance, that is, how likely will a pathogenic variant of a gene cause disease [197]. For pathogenic variants of high penetrance genes, clinicians more often have firmly established guidelines that inform recommendations for patient screening and surveillance. However, for pathological variants of low penetrance genes, less definitive clinical guidelines exist. NGS-based testing results in increasing detection of pathogenic variants of low penetrance genes; this increased detection adds complexity and uncertainty to patient management [198,199].

Clinicians face another challenge when selecting NGS gene panels for genetic evaluation: they must select the composition of the gene panel that they will employ. This selection requires specialized education and training [200,201]. The cancer genetics expertise required to address this challenge remains scarce [202–205]; the wider use of NGS platforms in clinical oncology and continued technological advances has made this expertise even more scarce [206–208].

AHCs possess the clinical expertise, facilities, support personnel, and administrative structures to meet the burgeoning demands of cancer genetics and to overcome the obstacles associated with the use of NGS in the clinic. Allied community oncology practices and their patients have access to these resources and services through their partnerships with AHCs. Four access models enable community oncology patient engagement with the AHC: (1) patient consultation visits to the academic cancer center, (2) cancer genetics specialist visits to community oncology sites, (3) telemedicine- and web-based remote visits and (4) AHC-sponsored genetic education initiatives that train community oncology practitioners to assess and manage cancer genetic risk and disease (Figure 1).

Conventionally, community oncology patients have received their cancer genetics care by consulting, in person, with a specialist at an AHC [155,209–211]. This model disadvantages community patients who live substantial distances from an AHC as it involves significant travel time and cost commitments [210,212,213]. Alternative cancer genetics delivery models have the potential to mitigate these problems.

In the community satellite clinic model, AHC cancer genetic specialists travel to the CHC clinic on an interval basis to meet the cancer genetic needs of community patients. This approach has proven successful in a variety of circumstances where logistical or economic challenges create barriers to effective cancer genetics care [214–217].

**Figure 1.** Community health center (CHC) patients requiring genetics care interface with specialists at academic health centers (AHC) through four modes of interaction. (1) The CHC patient may travel to the AHC for assessment. (2) The AHC genetics specialist may travel to a satellite CHC genetics clinic to evaluate the CHC patient. (3) CHC patients and AHC genetic specialists may interact via telemedicine consultation. (4) In order to provide genetics care to their patients, CHC physicians may undergo genetics specialty training sponsored by AHCs.

In our digital era, innovative cancer genetics delivery models have emerged; telemedicine platforms that involve both telephony and video communication platforms represent one such model [218–222]. The Division of Clinical Cancer Genetics (CCG) at COHCCC has assumed a national leadership position in the adoption of digital age technologies to provide academic center cancer genetic services to community oncology practices and their patients.

The CCG formed the Cancer Screening and Program Network (CSPPN), building a bridge to community oncology practices; the CSPPN utilizes innovative videoconferencing, telemedicine and wed-based applications to provide cancer genetics services [223]. Innovation continues at the CCG with the ongoing construction of new software and web-based platforms to permit effective communication between academic cancer genetics providers and community-based patients and practitioners [224].Alongside the use of these digital platforms, the CCG has administered a landmark educational program to provide community oncology healthcare providers with the necessary training that allows them to function as competent cancer risk assessment specialists in their own communities [225]. This program, funded by the National Cancer Institute, has expanded the workforce of qualified germline genetics providers and has helped to alleviate the shortage of cancer genetics expertise in CHC practices.

Educational programs, such as that sponsored by the CCG, have acquired additional practical importance as many healthcare systems now require, prior to genetic testing, assessment by a healthcare provider trained in genetics. These requirements may hinder effective cancer genetics care, particularly in underserved communities [226]; the availability of training will help eliminate this hindrance.

#### **6. The Community Oncology**/**Academic Cancer Center Alliance in Somatic Tumor Genomics**

The use of clinical NGS in oncology has risen exponentially [227]. Hundreds of commercial and academic laboratories now offer NGS-based clinical sequencing of cancer specimens [119,228]. The NGS sequencing formats for somatic tumor sequencing include, among others, whole exome,

whole genome, targeted panel, transcriptome and liquid biopsy assessments [229–235]. Various factors have driven the increased clinical application of NGS for somatic tumor assessment. The number of targetable genomic alterations increases substantially each year. Currently, there exist well over one hundred FDA-approved targeted therapies available for the treatment of both solid and hematological cancers [98]; over the past year alone, the FDA granted approval to nearly 20 new drugs or new indications for previously approved drugs [96]. With inclusion of therapy based upon molecular pathway considerations or off-label usage based on tissue-agnostic variant matching, the set of molecular targets and usage indications expands geometrically [236–244]. Purposing NGS-based somatic testing to determine clinical trial eligibility further increases the utility of NGS [245–248]; moreover, the demonstrated efficacy of testing to achieve improved outcomes has also motivated demand [248–252]. The decision by the Centers for Medicare and Medicaid Services to provide insurance coverage for NGS-based sequencing tests removed a financial barrier against the use of NGS, and contributed to the expanded use of this technology [253–255]. All told, currently over three quarters of oncologists use now NGS-base clinical testing to guide treatment decisions [256].

Significant challenges, however, temper enthusiasm for the clinical institution of somatic tumor NGS. A majority of oncologists report difficulty interpreting NGS somatic tumor testing, lack understanding of the clinical indications for testing and have inadequate opportunities to acquire the necessary training to properly use testing. One quarter of oncologists refer patients to other specialists to assist with NGS testing, and approximately 1 in 5 oncologists did not feel they had the proper knowledge to use properly NGS testing [256–258]. Additionally, oncologists report challenges with managing the large data volumes generated from NGS somatic testing. Oncologists also feel that they do not have the ability to distill from these reports actionable information; further, they lack the skill to manage germline variants detected as incidental findings in somatic NGS tumor testing [259–262]. These obstacles may be amplified for the CHC-based oncologist who lacks access to the necessary computational resources, logistical support and expertise in targeted therapeutics [263–266].

The CHC/AHC alliance provides solutions to alleviate these obstacles. Innovative AHC-based web applications make available to community oncologists an analytic framework and the computational tools to aid in the interpretation and clinical implementation of NGS sequencing results (Table 1). CIViC, an open access web resource, serves as a public central repository of NGS data "supporting clinical interpretations related to cancer" [267]. OncoKB, a precision oncology database, aids therapeutic decision-making based upon cancer gene variant status [268]; similarly, the web applications Personalized Cancer Therapy and My Cancer Genome assist both community and academic oncologists in selecting therapeutic options resulting from the somatic NGS of tumor specimens [269,270]. The SMART Cancer Navigator aggregates variant and clinical data from multiple data bases to assist community-based oncologists with the processing of NGS reports and the identification of effective targeted therapies [271]. At the COHCCC, investigators have configured an interactive web interface, HOPE-Genomics, that community patients and oncologists may use to better understand genomic sequencing results and treatment recommendations [224]. The COHCCC also provides to its community practice partners in-house NGS panel testing as part of its HOPESEQ molecular testing panel [272]; HOPESEQ includes genomic test reports designed to assist clinicians with interpreting the genetic testing results and clinical decision making. Furthermore, COHCCC physicians and community partners have access to Via Oncology; this tool provides a web-based clinical pathway system to help match patients with clinical trials and insurance reimbursement for NGS driven treatments [273].

Precision oncology tumor boards (POTBs) represent another solution to the problem of implementing NGS data in the CHC oncology clinic. POTBs arose from the need to assess, process and generate clinical treatment plans from the highly dense and complex data sets that arise from somatic NGS of tumor specimens [274]. POTBs serve two primary functions: targeted therapy drug matching and molecularly-informed clinical trial enrollment [275–280] (Figure 2).


**Table 1.** Web-based genomics resources available to community oncologists.

**Figure 2.** Multidisciplinary precision oncology tumor boards (POTBs) provide expert targeted drug matching and molecularly informed clinical trial enrollment for community oncology patients. Tumor specimens from community patients undergo nucleic acid sequencing with computational analysis to identify molecular alterations; this information provides a basis to discover candidate targeted therapies and determine clinical trial eligibility. An academic health center (AHC) POTB comprising, among others, clinical oncologists, pathologists, surgeons, geneticists and computational biologists, in consultation with community health center (CHC) oncologists, reviews patients' clinical cases and their sequencing results to select appropriate targeted therapies and clinical trials.

POTBs originated within AHCs as these centers possess the multidisciplinary expertise including, among others, clinical oncologists, pathologists, genomics specialists, computational biologists, pharmacologists and clinical geneticists to efficiently identify targeted therapies and clinical trials [281–290]. Targeted therapy drug matching requires comprehensive molecular mutational profiling and downstream pathway analyses of the tumor, combined with the identification of safe and effective therapeutic agents that redress these molecular alterations [291–293]; CHCs typically do not possess the analytic or pharmacologic capabilities to adequately perform these activities. Most clinical trials fail [294–296]; these failures result from a number of factors including deficient clinical trial design, poor proof of concept planning and insufficient administrative support and compliance [297–300]. Such failures have adverse consequences for both the clinical trial sponsors as well as the patients; failure has significant economic cost and results in lost therapeutic opportunity, in addition to potentially exposing the patient to harm from the investigational protocol and drugs [301–303]. These clinical trial-related matters may be more acute at CHCs given their more limited resources and the absence of experienced clinical trialists [304–307]. The POTB provides appropriate, molecularly-informed clinical trial assignment for patients, maximizing both the utility of clinical trial participation and potential patient benefit [308–320].

Given the resource limitations of the CHC oncology clinic, community POTB operation requires innovation and dedicated planning [83,321,322]. One innovation available to community oncologists, the web-based ASCO Multidisciplinary Molecular Tumor Boards, assists oncologists with understanding precision medicine-based tumor testing and the therapy recommendations resulting from these tests [323] (Table 1). Helio Learn Genomics, another web platform, offers a number of educational modules, including POTB cases, to help providers understand the molecular bases of carcinogenesis and precision therapeutics [324]. The Pancreatic Cancer Action Network administers a Know Your Tumor program, a turn-key precision medicine initiative, that allows community oncology practitioners to submit their patients' pancreatic cancer specimens for NGS molecular testing and to receive back a precision medicine-based treatment plan [325].

Another version of the POTB, the virtual POTB, permits the distance participation of community oncologists in an academic POTB. In this model an AHC hosts the POTB and reviews the clinical history and precision oncology testing results of the community oncology patient; subsequently, the POTB discusses with the community oncologist, using a live interactive video teleconferencing link, targeted treatment and clinical trial recommendations [263,311,326–330]. The Translational Genomics Research Institute (TGEN), an academic affiliate of the COHCCC, has successfully built a comprehensive, integrated, high-throughput sequencing and reporting framework that, when combined with remote teleconferencing, has proven tremendously successful in establishing efficient collaborative POTBs [331–335]. Together, these various models of providing clinical somatic NGS demonstrate the feasibility of leveraging precision oncology for the community-based cancer oncologist and their patients.

#### **7. Conclusions**

We have entered the age of precision oncology. Precision oncology offers the potential of molecularly informed medicine for the assessment of inherited cancer predisposition, as well as for the diagnosis and treatment of cancers. Realization of this potential depends upon access to specialized expertise and significant analytic and technological resources. While frequently available at AHCs, these resources have previously been limited for CHC oncology practices and their patients. In this paper, we have examined the CHC/AHC alliance and discussed examples illustrating how this alliance provides a structure that allows community cancer patients to benefit from germline and somatic precision oncology advances. Looking forward, multidisciplinary efforts, improved technology and continuing innovation promise to strengthen and facilitate the CHC/AHC alliance in oncology; this alliance offers community oncologists and their patients the prospect of unambiguous interpretation of genetic and genomic test results and optimized precision oncology care.

**Author Contributions:** Writing-Original Draft Preparation, M.M., S.S. (Shanmuga Subbiah), K.J.M.; Writing-Review and Editing, S.S. (Swapnil Rajurkar). 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/).

### *Review* **Non-Small Cell Lung Cancer from Genomics to Therapeutics: A Framework for Community Practice Integration to Arrive at Personalized Therapy Strategies**

### **Swapnil Rajurkar** †**, Isa Mambetsariev** †**, Rebecca Pharaon, Benjamin Leach, TingTing Tan, Prakash Kulkarni and Ravi Salgia \***

Department of Medical Oncology and Therapeutics Research, City of Hope, Duarte, CA 91010, USA; srajurkar@coh.org (S.R.); Imambetsariev@coh.org (I.M.); rpharaon@coh.org (R.P.); bleach@coh.org (B.L.); titan@coh.org (T.T.); pkulkarni@coh.org (P.K.)


Received: 26 May 2020; Accepted: 12 June 2020; Published: 15 June 2020

**Abstract:** Non-small cell lung cancer (NSCLC) is a heterogeneous disease, and therapeutic management has advanced with the identification of various key oncogenic mutations that promote lung cancer tumorigenesis. Subsequent studies have developed targeted therapies against these oncogenes in the hope of personalizing therapy based on the molecular genomics of the tumor. This review presents approved treatments against actionable mutations in NSCLC as well as promising targets and therapies. We also discuss the current status of molecular testing practices in community oncology sites that would help to direct oncologists in lung cancer decision-making. We propose a collaborative framework between community practice and academic sites that can help improve the utilization of personalized strategies in the community, through incorporation of increased testing rates, virtual molecular tumor boards, vendor-based oncology clinical pathways, and an academic-type singular electronic health record system.

**Keywords:** non-small cell lung cancer; driver mutations; testing rates; receptor tyrosine kinases; team medicine

#### **1. Introduction**

Lung cancer remains the leading cause of cancer deaths in the United States and, in 2020, it will be responsible for an estimated 230,000 cases and 135,000 deaths in the US alone [1]. Non-small cell lung cancer (NSCLC) is the major histological subtype that accounts for approximately 85% of all lung cancer cases and encompasses several subtypes, including adenocarcinoma, squamous cell carcinoma, and large cell carcinoma [2]. Despite advances in screening and diagnosis, most patients still present with metastatic disease, at which point surgical intervention is no longer an option [3]. The advent of targeted therapy and immunotherapy has altered the course of treatment for the majority of patients—with molecular testing now a standard recommendation for late-stage lung adenocarcinoma patients. Tyrosine kinase inhibitors (TKIs) that target abnormalities in several genes, such as *ALK* and *EGFR*, have shown better progression-free survival (PFS) as compared with standard chemotherapy in a number of NSCLC trials [4–6]. More recently, other molecular markers, including ROS1, RET, NTRK, BRAF, and MET, have delivered similar clinical benefits to patients with late-stage NSCLC [7–12]. Furthermore, mature outcome data from second-generation TKIs is showing durable overall survival benefit for patients [13,14], a factor that was previously disputed with earlier TKIs [15].

Several molecular targets that were previously considered "unactionable", such as KRAS, now have several targeted therapies under consideration with promising early results [16,17]. Nevertheless, for patients without an actionable target or progression of disease, immune checkpoint inhibitors (ICIs) have resulted in durable outcomes and clinical benefit across several NSCLC trials in various lines of therapy [18–24]. Protein expression testing of programmed death-ligand 1 (PD-L1) has been identified as a potential, though not definitive, biomarker of predicting response to immunotherapy [21,25–27]. Beyond tumor response, recent results from KEYNOTE-001 showed that pembrolizumab monotherapy was associated with a 23.2% 5-year overall survival as compared to 15.5% for previously treated patients [28]. However, therapeutic advancements and outcome improvements have not been uniformly applied in practice, with the majority of trials and novel therapies being more prevalent in academic sites as compared to community practice. We previously showed in a retrospective study that in a cohort of 253 patients from nine community practice centers, the molecular testing rate for first-line treatment decisions was 81.75%, with testing for PD-L1 at only 56% [29]. This suggests that while community sites are on pace to improving their testing rates, the current results are inadequate and require more education and understanding of novel upcoming personalized therapies. The purpose of the current review is to shed light on the available and upcoming therapies in lung cancer, to report the gaps in community practice testing rates, and to identify the available tools that can assist in complex lung cancer management and decision-making.

#### **2. Advances in Genomic Testing and Personalized Therapy**

In the last 20 years, therapeutic management of lung cancer has progressed from cytotoxic chemotherapies to personalized targeted therapies that act upon specific genomic alterations. Prior to this, while cytotoxic therapies showed a benefit for early-stage disease [30,31], there was no reported outcome benefit in patients with late-stage lung cancer [32]. Following the completion of the multi-billion dollar endeavor of the Human Genome Project in 2003 [33], the development of next-generation sequencing with high-throughput has enabled large-scale parallel sequencing of the lung cancer genome revealing a plethora of genomic targets including EGFR (10–50%), KRAS (25%), ALK (2–7%), ROS1 (1–2%), RET (1%), BRAF (4%), and others [34,35]. Initially, EGFR tyrosine kinase inhibitors were evaluated in unselected populations with mixed responses due to inadequate selection of patients with EGFR alterations [36,37]. However, the results from randomized Phase III trials for EGFR and ALK tyrosine kinase inhibitors [5] led to the acceptance of genomic testing for ALK and EGFR alterations in routine clinical practice, and in turn, led to the development of faster and more efficient next-generation sequencing platforms that were Clinical Laboratory Improvement Amendments (CLIA)-certified and became widely accepted commercially and at academic sites [38]. While first-generation EGFR TKIs, including gefitinib and erlotinib, showed improved progression-free survival, retrospective studies and outcomes data failed to show improvements in overall survival outcomes [13,39–42]. In contrast to these results, the FLAURA trial for second-generation TKI, osimertinib, showed significant progression-free survival benefit (median PFS 18.9 vs. 10.2 months) and a considerable overall survival benefit of 35.8 months as compared to 27.0 months in the control [43]. The durable survival benefit of targeted therapies had previously been disputed, but recent results from the long-term survival of advanced ALK-rearranged patients treated with crizotinib showed an undisputable benefit of median OS of 6.8 years and a 5-year OS rate of 36% as compared to the historical 2% [44]. Moreover, advances in immunotherapy have yielded similar improvements and KEYNOTE-189 showed that patients who received immunotherapy resulted in a 20% improvement in the overall survival [45].

The promise of precision medicine and the arrival of personalized therapy has transformed lung cancer care with a number of genetic alterations that have come to fruition or are quickly rising with promising trial results, including EGFR, ALK, ROS1, MET, RET, NTRK, BRAF, KRAS, and immunotherapies (Table 1). However, the rapid and dynamic nature of emerging trial results has made lung cancer management difficult and while academic sites are familiar with trial results

and the latest available therapies, a community oncologist, who may see a variety of solid tumors, may have difficulty grasping the complexity of these genomic alterations. In our experience at the academic site, actionable alterations were identified in 53.5% of patients with lung cancer, and the use of genomic-informed therapy was associated with improved survival benefit as compared to patients with no actionable alterations [46]. The use of genomic-informed therapy and selective immunotherapy must be standardized within community practice to ensure improved outcomes.


**Table 1.** Actionable targets in lung cancer and available therapeutics.

#### *2.1. EGFR*

The epidermal growth factor receptor is a transmembrane cell-surface receptor that is activated in 10–50% of patients with NSCLC, which varies based on populations and is more common in Asians and nonsmokers [34,48]. The receptors in the EGFR family exist as inactive monomers, but the binding of extracellular growth factors, such as epidermal growth factor (EGF), has been shown to cause receptor dimerization and induced autophosphorylation of the tyrosine kinase domain, with downstream and intercellular signaling cascades that in turn affect cell motility, invasion, proliferation, and angiogenesis [49]. Initial mutations in EGFR were first described in 2004 and activating mutations in EGFR occurring in exons 18–21 of the kinase domain were associated with sensitivity and response to gefitinib and erlotinib [50–52]. This led to the selection of patients with adenocarcinoma histology and EGFR alterations and, in 2009, a landmark Phase III Iressa Pan-Asia Study (IPASS) identified clinical responsiveness and increased progression-free survival in EGFR mutant patients who received gefitinib as compared to standard chemotherapy [50]. The landmark Phase III trial, EURTAC, evaluating erlotinib, an EGFR TKI, as a first-line therapy for patients with EGFR mutations, showed an increased

median PFS of 9.7 months as compared to 5.2 months with standard chemotherapy [53]. Two other Phase III trials, the OPTIMAL and ENSURE trials, showed a similar improvement with erlotinib and the US Food and Drug Administration (FDA) approved erlotinib as a first-line cancer therapy for EGFR mutation-positive patients [4,53,54]. Similarly, afatinib, a second-generation TKI, received FDA approval in 2013 following two Phase III trials, Lux-Lung 3 and Lux-Lung 6, that both showed improved PFS of 11.1 months and 11 months respectively, as compared to standard chemotherapy in the first-line setting [55,56].

In 2015, efficacy results for patients with exon 19 deletions or exon 21 (L858R) mutations treated with gefitinib showed a 50% objective response rate (ORR) and led to the FDA approval of gefitinib as a first-line therapy for EGFR mutation-positive patients [57]. However, at that time erlotinib became the standard choice of therapy for many EGFR mutated patients, and mechanisms of primary and secondary resistance to TKI therapy began to emerge. The most commonly identified acquired resistance to early-generation TKIs was the T790M substitution, a secondary EGFR mutation in exon 20, that accounted for approximately 60% of cases [53,55,58,59]. The development of mutant selective pyrimidine-based third-generation TKIs that could block the T790M substitution led to the AURA3 trial evaluating osimertinib, a third-generation TKI, as second-line therapy following T790M EGFR TKI resistance [6]. In 2017, the results of the AURA3 trial showed a significantly improved PFS of 10.1 months and a response rate of 71% as compared to standard chemotherapy [6], and this led to the issuance of FDA approval for osimertinib in the second-line setting for EGFR T790M mutation-positive patients treated with first-line EGFR TKI. Compounding results also exhibited higher CNS response rates with osimertinib (40% vs. 17%) and a longer CNS PFS of 11.7 months vs. 5.6 months [60]. Brain metastases occur in approximately 20–40% of EGFR patients at presentation [61,62] and CNS activity of osimertinib hinted at its potential as a first-line therapy. Unsurprisingly, in 2018, the results of the FLAURA trial showed osimertinib as superior in the first-line setting as compared to first-generation TKIs, with a median PFS of 18.9 months (vs. 10.2 months), ORR of 77% (vs. 69%), and a median duration of response (DOR) at 17.6 months (vs. 9.6 months) [13]. This led to the issuance of FDA approval for osimertinib as the first-line therapy option for EGFR mutant lung cancer. Furthermore, mature data from the FLAURA trial also showed a medial overall survival benefit of 38.6 months over 31.8 months in the control and there was a significant improvement in quality of life, a clinical factor that was never previously achieved in first-generation TKIs [43].

However, despite advances in therapy, acquired resistance inevitably occurs, including EGFR-dependent resistance (6–10%), MET and HER2 amplifications (8–17%), small cell lung cancer (SCLC), and squamous cell carcinoma (SCC) transformation (15%), and others [63]. EGFR-dependent resistance includes S768I, L861Q, G719X, and other alterations that are resistant to most first-generation TKIs except for afatinib that was approved for first-line therapy for patients with rare EGFR alterations [64]. Additional TKIs such as poziotinib are currently under consideration for such alterations and Phase II preliminary data showed a response rate of 43% and a median PFS of 5.5 months in previously treated EGFR-mutant patients [65]. Additionally, other TKIs including TAK-788 (NCT03807778), TAS6417 (NCT04036682), and tarloxotinib (TH-4000) (NCT03805841) are currently under investigation in this setting. There are other trials available for less-frequent mutations of EGFR, such as exon 18 or exon 20 EGFR insertions. The availability of numerous EGFR TKIs in the first and refractory setting is strictly contingent upon appropriate assignment to therapy following reflex molecular testing. The improvements in survival are dependent on early identification of molecular markers and appropriate sequence of TKI therapy. In one retrospective study of rates of molecular testing in a community-based academic center, EGFR testing following the approval of reflex testing was only 62% [66]. In another larger cohort of 814 community practice patients, testing rates were similarly low, with only 69% of patients who were tested for EGFR mutations, and approximately 70% of patients who tested positive received appropriate targeted therapy [67]. In a retrospective evaluation of 1,203 advanced NSCLC patients from five community oncology practices, the testing rates of EGFR were at 54% [68]. A comprehensive retrospective cohort of 191 community oncology practices with 5688 patients performed by Flatiron Health, selected patients who were tested for EGFR alterations with either broad genomic sequencing or routine-testing and identified 154 EGFR-mutated patients in the broad-based sequencing group, but reported that only 25% of these patients received appropriate EGFR-targeted therapy [69]. The findings of the study concluded that there was no survival difference between broad-based and routine genomic sequencing, but this misrepresented the utility of broad-based genomic sequencing in the community, as better outcomes cannot be achieved without appropriate assignment to targeted therapy. Meanwhile, in our own community practice experience of 253 patients, we reported testing rates of 94% for EGFR and 96.2% of patients with an EGFR sensitizing mutation received a TKI therapy [29]. The translation of outcomes reported in clinical trials to real-world outcomes requires cooperation and acceptance of molecular testing within community practice and the integration of targeted therapies in community decision-making.

#### *2.2. ALK*

ALK, a receptor tyrosine kinase, was originally identified in lung cancer in 2007 with the detection of an echinoderm microtubule-associated protein-like 4 (EML4) gene and anaplastic lymphoma kinase (ALK) gene fusion from a surgically resected lung adenocarcinoma patient [70]. This gene rearrangement is largely independent of EGFR alterations and has been described as an actionable oncogene with incidence in 1–7% of lung cancer patients [71]. ALK-rearranged patients tend to be younger and—similar to EGFR—have a limited history of smoking. Crizotinib, while originally developed as a MET therapeutic, showed a preclinical efficacy for ALK [72]. The Phase I trial lead to the FDA approval of crizotinib in ALK-positive NSCLC [5]. In 2013, the results of the Phase III trial evaluating crizotinib compared to standard chemotherapy showed PFS of 7.7 months (vs. 3.0 months) and ORR of 65% (vs. 20%) [5], resulting in FDA approval of crizotinib for first-line therapy as a standard of care. As with other TKIs, while patients initially respond to ALK inhibitors, resistance invariably develops and one of the most common resistance mechanisms is an acquired ALK mutation (1151Tins, L1152R, C1156Y, F1174V/L, G1269A, and others) [73]. Other resistance mechanisms include EGFR activation, KIT activation, KRAS mutation, and IGF1R activation [74–79]. It was estimated that 25% of ALK-mutated patients do not respond to crizotinib in the first-line setting and, in response to these resistance mechanisms [77], other ALK TKIs have been developed. In 2014, the results from the Phase I trial evaluating ceritinib as a potential therapy in ALK-rearranged NSCLC patients with disease progression on crizotinib showed a median progression-free survival of 7.0 months and a response rate of 56% [80]. Based on only the Phase I trial results, the FDA approved ceritinib in patients who have progressed on crizotinib, and in 2017, it expanded its approval for first-line use. Alectinib received similar approval in 2015 in the refractory setting that was later expanded to first-line in 2017 [81–83]. In the first-line, alectinib showed a median PFS of 34.8 months with an OS rate of 62.5% as compared to crizotinib with 11 months and 52% [81–83]. Brigatinib, a second-generation ALK TKI, was initially identified to have preclinical efficacy and grater potency against all 17 ALK mutants as compared with crizotinib [84,85]. Initial results for brigatinib from a Phase II trial in the refractory setting showed promising responses and yielded FDA approval in 2017 [86]. While alectinib has been shown to be effective against L1196M, C1156Y, and F1174L ALK gatekeeper mutations [87], brigatinib has shown efficacy against ROS1, FLT3, and IGF-1 secondary mutations [88]. The results of the Phase III trial for brigatinib vs. crizotinib in the first-line showed an estimated PFS of 12 months as compared to 11 months with crizotinib, and two-year follow-up data showed brigatinib reduced the risk of progression or death by 76% [14,89]. Several other new generation ALK TKIs including lorlatinib and ensartinib demonstrated 73% and 72% ORR, respectively, following crizotinib and we are awaiting first-line results [90,91].

The availability of a number of ALK inhibitors has complicated management of ALK patients, but in a long-term assessment of 110 patients with an ALK inhibitor, a remarkable OS for advanced ALK NSCLC patients of 6.8 years was reported with 78.4% of patients receiving another ALK inhibitor after first-line progression [44]. Therefore, many studies are reporting that the success of ALK inhibition

therapy may lie in the sequence of administrating ALK inhibitors based on metastatic progression and resistance profiles [92,93]. In a retrospective analysis of 31,483 patients with advanced NSCLC at community practices, ALK overall testing rates were 53.1% and rose to 62.1% in 2016, with 21.5% of patients who were initiated into non-targeted therapy before receiving test results [94]. Gierman et al. in 2019 evaluated 1,203 advanced NSCLC patients from five community practices and results showed that only 51% of patients were tested for ALK rearrangement, with approximately 45% of actionable patients receiving targeted therapy [68]. A concurrent study of 814 community practice patients showed that only 65% were tested for ALK alterations [67]. A retrospective study of advanced NSCLC across over 70 community sites in the US showed that only ~50% of patients were tested for ALK alterations during their cancer care [95], suggesting that advancements in liquid biopsies and testing are not translating to real-world practice. The use of liquid biopsies in a large cfDNA study showed that genomic results were concordant with tissue and utilizing cfDNA liquid biopsies increased detection and rates of testing by 48% [96]. The integration of liquid biopsy testing and further controls on tissue biopsy testing may improve the rates of ALK testing and translate the 6.8-year median survival benefit from academic site-wide studies into real-world efficacy.

#### *2.3. ROS1*

*ROS1* has been identified as an oncogene in lung cancer and rearrangements have been reported in 1 to 2% of patients with NSCLC [34]. The fusion mutations lead to the dysregulation of the tyrosine-kinase dependent multi-use intracellular signaling pathway, which in turn accelerates growth, proliferation, and progression [97]. Similar to EGFR and ALK alterations, *ROS1* fusions and rearrangements are mutually exclusive and independent of other oncogenes such as KRAS or MET [98]. Following the discovery of *ROS1* fusions in 2007 and in part due to the high degree of homology between *ALK* and *ROS1,* the tyrosine kinase inhibitor crizotinib was explored as a therapeutic option [99,100]. Crizotinib was approved by the FDA in 2016 contingent upon clinical benefit from a PROFILE 1001 Phase I study, where patients had a median PFS of 19.2 months and an ORR of 72% [101]. A Phase II study of ceritinib with 32 patients showed an ORR response rate of 62% and a PFS of 19.3 months for crizotinib-naïve patients, but FDA approval is pending and ceritinib was ineffective against resistance mutations but had activity against CNS disease, as intracranial ORR was 25% and intracranial DCR was 63% [102]. Unlike ceritinib, entrectinib has been shown to be effective against some resistance mutations and had similar CNS activity with a median PFS of 13.6 months and ORR of 55% for patients with CNS disease [103]. This led to the FDA's approval of entrectinib in the management of ROS1-positive NSCLC. However, lorlatinib is currently the only inhibitor under consideration for *ROS1* that is effective against most resistance mutations and in a Phase II trial it induced an ORR of 26.5% with a PFS of 8.5, with considerable CNS activity inducing an ORR of 52.6% [104]. Other agents such as DS6051b (NCT02279433) and repotrectinib (NCT03093116) are also currently under investigation with results awaiting. A 2018 study by Friends of Cancer Research and Deerfield Institute announced the response of a survey of 157 oncologists and showed that ROS1 testing in the community centers was 32% [105]. However, a comprehensive study of 14,461 patients treated in the community showed testing rates for ROS1 were incrementally lower at 5.7% with 35.5% and 32.9% for *EGFR* and *ALK* respectively [106]. Of the three major approved alterations, ROS1 has the lowest testing rates in several studies [67,105,106]. While tissue biopsies remain the gold standard in detecting *ROS1* fusions and rearrangements, advances in liquid biopsy have shown that it is a viable option for *ROS1* and implementation of this practice may increase the testing rates within the community practice [29,107].

#### *2.4. MET*

*MET* oncogenic mutations and amplification has been noted in various solid tumor malignancies, including NSCLC, breast cancer, and head and neck cancer [108–112]. MET alterations or its ligand activation (hepatocyte growth factor) causes the activation of the tyrosine kinase which subsequently activates downstream signaling pathways related to cell growth, apoptosis, motility, and invasiveness [113]. Initially discovered in familial and sporadic papillary renal carcinomas [114], subsequent studies revealed the incidence of *MET* alterations in SCLC and NSCLC, especially MET exon 14 skipping as identified initially by our laboratory [115,116]. *MET* alterations have an incidence rate of 6% in lung adenocarcinoma and 3% of lung squamous cell carcinoma [117,118]. The most frequent alteration is the *MET* exon 14 skipping mutation, which has been identified in 4% of lung cancers. A 2015 study was the first to demonstrate clinical efficacy of crizotinib or cabozantinib in NSCLC patients with *MET* exon 14 skipping mutations [119]. A recent study enrolled 69 NSCLC patients harboring *MET* exon 14 alterations that were treated with crizotinib and reported an ORR of 32% and a median PFS of 7.3 months, suggesting antitumor activity with crizotinib treatment [120]. Several clinical trials, such as the GEOMETRY mono-1 trial and the VISION trial, are evaluating other TKIs like capmatinib and tepotinib in MET exon 14-mutated NSCLC and have shown promising results [12,121]. Interim results of the Phase II GEOMETRY mono-1 trial with 97 enrolled patients reported good ORR and a median PFS of 9.13 months in the treatment-naïve cohort [12]. Recently, capmantinib was granted accelerated FDA approval in metastatic NSCLC patients with *MET* exon 14 skipping mutation, the first TKI approved for MET NSCLC patients. MET amplification, which accounts for 1–4% of NSCLC patients who have not been treated with EGFR TKIs, is associated with a poor prognosis [122,123]. A Phase I trial investigated telisotuzumab vedotin, an antibody-drug conjugate, in NSCLC patients with MET overexpression and demonstrated safety and tolerability of the drug with promising antitumor efficacy [124]. In a study of NGS testing rates of genomic biomarkers in NSCLC patients treated at community sites, only 15% of the 814 patients underwent NGS testing for MET, a sharp decline compared to EGFR (69%) or ALK (65%) testing rates [67]. This testing rate was recapitulated in another community analysis [69], however, MET testing rates were reported as low as 6% in an analysis of NGS screening rates between private clinics, academic centers, and community sites [105].

#### *2.5. RET*

Activation of RET results in downstream pathway signaling including MAPK, JAK/STAT, and PI3K/AKT, leading to cell proliferation and migration. Alterations in *RET* are most frequently found in medullary thyroid carcinoma and NSCLC. In NSCLC, RET rearrangements are found in approximately 1–2% of cases [117]. These patients tend to be non- or former light smokers with adenocarcinoma histology and present with advanced disease [125]. Since its discovery, several targeted therapies have been investigated including multikinase inhibitors and selective RET inhibitors. A Phase II trial of RET fusion-positive NSCLC patients were treated with cabozantinib, a TKI targeting RET, VEGFR, and MET. The results demonstrated good clinical efficacy with an ORR of 28% and a median PFS of 5.5 months [126]. The most promising selective RET inhibitors currently under investigation are BLU-677 and selpercatinib (LOXO-292). Interim results from a Phase I clinical trial of 79 RET fusion-positive NSCLC patients treated with BLU-677 demonstrated an ORR of 56% among the 57 evaluable patients and encouraging central nervous system (CNS) activity against brain metastases [127]. The Phase I/II LIBRETTO-001 trial evaluating selpercatinib in a cohort of previously treated NSCLC patients with RET rearrangements (N = 105) also demonstrated marked antitumor efficacy with an ORR of 68%, a remarkable CNS response of 91%, and a median PFS of 18.4 months [8]. In the treatment-naïve cohort (N = 34) of the trial, the ORR was 85%, resulting in the FDA approval of selpercatinib for patients with RET-positive NSCLC. Like MET testing rates, RET demonstrated a 14–15% testing rate in community NSCLC patients [67,69]. Also similar to MET, RET testing rates were reported as low as 8% [105]. This is a staggeringly low rate considering the recent FDA approval and great antitumor activity of selective RET inhibitors.

#### *2.6. NTRK*

*NTRK* genes (*NTRK1*, *NTRK2*, and *NTRK3)* encode three TRK proteins (TRKA, TRKB, and TRKC), which play an important role in the cell growth, differentiation, and apoptosis of peripheral and CNS

neurons [128]. *NTRK1* and *NTRK2* rearrangements account for 3–4% of NSCLC cases [129]. Several clinical trials have shown the efficacy of TRK inhibitor treatment in *TRK*-positive tumors. Larotrectinib (LOXO-101), a highly selective pan-TRK inhibitor, was first evaluated in a study of 55 pediatric and adult patients with various *TRK* fusion-positive malignancies, four of whom had lung cancer, and reported an ORR of 75% [10]. Remarkably, responses were shown to be durable with a response rate of 71% while 51% of patients stayed progression-free at one year. A multicenter analysis of three major Phase I/II clinical trials—STARTRK-1, STARTRK-2, and ALKA-372-001—investigating entrectinib in 54 patients diagnosed with advanced or metastatic *NTRK*-positive tumors demonstrated an ORR of 57%, a median PFS of 11.2 months, and a median OS of 20.9 months [130]. Larotrectinib and entrectinib are currently FDA-approved for the treatment of advanced *NTRK* fusion-positive NSCLC. Although these clinical trials have shown strong and durable responses to first-generation TRK TKIs, acquired resistance mutations have been identified in colorectal and mammary analogue secretory carcinomas, requiring the development of second-generation TKIs [131,132]. LOXO-195, a second-generation TRK-selective inhibitor, has shown preclinical efficacy and clinical activity in a Phase I trial of *NTRK* fusion-positive cancers previously treated with larotrectinib, demonstrating an ORR of 45% [133,134]. Despite the great clinical response elicited by NTRK-targeted therapies, NTRK testing rates were shown to range from 0–15% in several community site analyses [69,105].

#### *2.7. BRAF*

*BRAF* mutations represent 7% of NSCLC cases and are more commonly found in current or former smokers and female patients [117]. The most frequent *BRAF* activating mutation, V600E, carries a poorer prognosis and a shorter disease-free survival [135]. A Phase II trial investigated combination treatments of dabrafenib and trametinib in chemotherapy-pretreated patients diagnosed with *BRAF* V600E-mutated NSCLC and reported an ORR of 63% and a median PFS of 9.7 months in 52 evaluable patients [11]. In a Phase II trial of treatment-naïve patients with *BRAF* V600E-mutated NSCLC, treatment with dabrafenib and trametinib resulted in an ORR of 64% and a median PFS of 10.9 months, although 69% of patients experienced at least one grade 3/4 adverse event [136]. Currently, the combination of dabrafenib and trametinib is FDA approved for the treatment of advanced NSCLC harboring the *BRAF* V600E mutation regardless of the previous therapy. In an analysis by Gutierrez et al., BRAF NGS testing rates in 814 community site patients were reported to be 18%, similar to MET and RET NGS testing rates [67]. Other analyses demonstrated consistent rates of 12–29% [68,69,105]. Interestingly, rates of BRAF testing were shown to be as low as 0.1% in a larger analysis of 14,461 NSCLC patients treated in the community [106].

#### *2.8. KRAS*

Alterations in *KRAS*, one of the most frequent oncogenes in solid tumor malignancies, represent up to 32% of lung adenocarcinoma cases [117]. They are generally found in smokers [137] and are associated with a poor prognosis [138], although recent data have reported that it has a minimal effect on overall survival in early-stage NSCLC [139]. Therapeutic targeting of *KRAS* has been notoriously difficult, thus dubbing the molecular marker as an "undruggable" target. However, research into KRAS small molecule inhibitors targeting mutational variants of *KRAS* has shown preclinical and clinical efficacy. AMG-510, an inhibitor targeting KRAS G12C, which accounts for 13% of *KRAS* mutant NSCLC [140], is currently under investigation in a Phase I/II clinical trial of advanced *KRAS* mutant solid tumors. Interim results were recently presented and showed that out of the 29 patients, 10 were diagnosed with NSCLC, of which 90% (*N* = 9) of patients exhibited either a partial response or stable disease [16]. Although there are currently no FDA-approved drugs targeting *KRAS*, small molecule inhibitors like AMG-510 and JNJ-74699157 continue to demonstrate good clinical activity. Another drug, MRTX849, has also shown potent efficacy in vitro and in vivo for G12C positive lung cancer, with pronounced tumor regression in 17 of 26 (65%) KRAS G12C positive cell lines [141]. Preliminary data from the Phase I trial also showed a ~30% decrease in target lesions in heavily pre-treated lung

cancer patients [141]. NGS testing of KRAS, although still important now, will become necessary once targeted therapies become approved. In several studies of molecular testing rates in community sites, KRAS testing has widely varied, ranging from 0–43% [66,67,69,105]. As more and more targets such as KRAS become clinically actionable, the landscape of lung cancer therapeutic management will continue to change. However, a number of actionable alterations are currently FDA approved and have distinct therapeutic strategies currently available (Figure 1).

The testing rates reported in the community have been rising over the years, and the main driver of this transformation has been education and dissemination of novel therapeutics available for the different oncogenes. However, more effort is required as the primary challenge remains that many newly approved targets face an astronomical hurdle in being implemented in daily community practice (Table 2). The most distinct example of this is the testing rates of BRAF reported in community practice at 0.1% in 14,445 patients—the lack of testing also poses a threat towards clinical trial enrollment and delivery of novel therapeutics to patients [106].


**Table 2.** Reported testing rates of clinically actionable and clinically relevant oncogenes in community practice.

**Figure 1.** Genomic-informed and immunotherapy-focused management of NSCLC based on approved therapies. The role of immunotherapy is not clear in all of the actionable targets but is currently under investigation.

*J. Clin. Med.* **2020** , *9*, 1870

#### *2.9. Immunotherapy*

The availability and discovery of more and more targeted therapies makes it a priority that all advanced NSCLC patients are tested at presentation. However, when an actionable alteration is not available, treatment decisions may depend on PD-L1 expression, histology, or the onset of progressive disease. In these situations, immune checkpoint inhibitors have induced response through interaction with cytotoxic T cells, helper T cells, NK cells, macrophages, and other immune mechanisms. In 2015, the first results of monoclonal antibodies against programmed death ligand-1 (PD-1) in the refractory setting showed efficacy of nivolumab, PD-1 inhibitor, with OS (12.2 months) as compared to second-line chemotherapy (9.5 months) [18–20]. This led to the FDA approval of nivolumab in advanced NSCLC. Similar approval of pembrolizumab, a PD-1 inhibitor, was contingent upon results from KEYNOTE-001 that showed ORR of 19.4 in refractory NSCLC patients [21]. Soon after, two PD-L1 inhibitors, atezolizumab for stage IV metastatic disease and durvalumab for stage III disease, were also approved based on positive ORRs and OS [22,144]. However, the preliminary analysis reported that PD-L1 expression may be a potential biomarker of response and resistance with only 6.6% of patients whose tumors were negative to PD-L1 responding to durvalumab [22]. In the front-line setting, pembrolizumab was the first immune checkpoint inhibitor (ICI) to demonstrate median PFS of 10.3 months (vs. 6 months) and a response rate of 44.8% (vs. 27.8%) based on the results of KEYNOTE-024 as compared to chemotherapy [145], and it can be utilized as a monotherapy or in combination with chemotherapy depending on PD-L1 expression and the performance status of the patient at presentation [146]. The addition of chemotherapy to pembrolizumab resulted in an increased OS at 12 months of 69.2% (vs. 49.4%) and a median PFS of 8.8 months (vs. 4.9 months), with a comparable adverse event rate of 67.2% vs. 65.8% [147]. These results were surprisingly not recreated when nivolumab was evaluated as a monotherapy, showing a median PFS of 4.2 months with nivolumab vs. 5.9 months, and a similar OS benefit of 14.4 months vs. 13.2 months in the chemotherapy control group [23]. However, it did have success in combination with ipilimumab, showing an improvement in overall survival of 17.1 months vs. 13.9 months with chemotherapy, and a nominal duration of response of 23.3 months (vs. 6.2 months) for the front line setting [148].

Nivolumab plus ipilimumab remains a controversial choice due to grade 3 and 4 adverse events in 32.8% of patients [148]. Atezolizumab monotherapy achieved similar approval with incremental improvements in OS [24], but durvalumab in combination and alone failed to improve survival [149]. While the availability of therapies is beneficial to patients, pembrolizumab is slowly becoming the first-choice option for front-line immunotherapy, partially due to its favorable toxicity profile and versatility as a monotherapy and in combination therapy [150]. However, the availability of therapies has not translated into practice and a retrospective observational study of 55,969 NSCLC patients from the community showed that only 1,344 patients received nivolumab or pembrolizumab in the metastatic setting [142]. More surprisingly, only 8% of these patients were tested for PD-L1 expression [142]. More so, an outcomes study of 423 patients with high PD-L1 who received first-line pembrolizumab monotherapy in the community showed that community clinical outcomes were comparable to clinical trial results with a median PFS of 6.8 months vs. 6.1 months and a median OS of 19.1 months vs. 20 months [151]. A larger study of 10,689 patients in the community showed that utilization of immunotherapy in the first-line is not yet implemented, with <1% of patients treated with immunotherapy in the first-line, but rates were improved in the second and third-line setting [143]. PD-L1 expression was equally underperformed and was tested in <1% of patients [143]. Furthermore, in a quality improvement study of 100 patients who received immunotherapy in the community, only 61% fully completed immunotherapy as planned and 81% had immune-related adverse events [152]. While it is concerning that the reported use of immunotherapy in the community practice is limited, based on experience from melanoma and immunotherapy, the rates are anticipated to slowly increase over time with more education and acceptance of various immunotherapy options [153].

While PD-L1 remains an imperfect biomarker, several subgroup analyses in the trials mentioned above show an increased benefit in patients with PD-L1 ≥1% or ≥50%. Therefore, PD-L1 testing should be considered in everyday decision-making, and currently four PD-L1 testing types are available: 22C3, 28-8, SP263, and SP142 [154]. The 22C3 IHC assays were developed alongside pembrolizumab in the Phase I trial as a biomarker for patients who may benefit from treatment [155]. Meanwhile, IHC 28-8 test was developed to be used in conjunction with nivolumab, and SP142 was developed for trial use with atezolizumab [18,19,156,157]. SP263 is the most recent assay that was developed for use with durvalumab, especially in the Stage III setting in NSCLC [156]. All four assays are FDA approved in their individual setting and while testing is not required to initiate treatment, it may support clinical decision-making [156]. Meta-analysis reports show that there is high concordance between 22C3, 28-8, and SP263 assays, but SP142 detected significantly lower PD-L1 expression [154,156]. At the same time, evidence shows that non-commercial laboratory-developed tests (LDTs) used by academic centers detect similar overall percentages of PD-L1 (≥1%) at 63% (vs. 22C3 61%), but PD-L1 ≥50% were much lower at 23% (vs. 22C3 33%) suggesting LDTs are less sensitive than commercial tests [158]. LDTs are becoming more and more utilized in practice and offer a potential solution to the complexity of commercial PD-L1 tests. However, the lack of PD-L1 testing and the difficulty of immune-related toxicities is a challenge that is more difficult to address, and we believe that the integration of community practice with the academic site model is one solution to this grave issue.

#### **3. Integration of Personalized Therapy and Molecular Testing in the Community through an Academic Site to Community Practice Network**

Advances in targeted therapy and immunotherapy have lowered the costs of molecular testing, making it a viable practice in the academic sites and the community [159]. While academic sites have benefited from a close knowledge of clinical trials and novel therapies, the drive of personalized medicine has not been uniform, with the majority of patients in the community lacking appropriate testing and assignment to therapy [66–69,94,95,106,142,143,152]. This is especially concerning as the majority of patients or approximately 85% with cancer are treated in the community setting and 50% of collaborative group trial accruals occur in the community [160]. Several models have been proposed to integrate community oncologists into the academic paradigm of personalized medicine, with the most promising being the establishment of interpersonal relationships between community oncologists and academic site physicians through molecular tumor board (MTB) teams [161–165]. The establishment of an MTB team would allow for the proper evaluation of imaging, histopathology, and genomic information that is required to make the appropriate therapeutic decision [166]. One reported study involving 1725 patients who were evaluated through a cloud-based virtual molecular tumor board (VMTB) showed that oncologists chose the VMTB-derived therapies over others, resulting in an increase of matched therapies [165]. Such a model also allows for the dissemination of information regarding available CLIA-certified vendors and platforms for both tissue and liquid biopsy testing that are imperative to improving testing rates and outcomes [167]. The MTB model can be scaled into the community through virtual or physical collaboration, and would further improve collaboration between community sites and academic sites through the interactions between pathologists, oncologists, primary care physicians, radiologists, and pulmonologists in the decision-making process (Figure 2). This team-based approach can be utilized in all cancers, especially during crises such as the recent pandemic of novel coronavirus [168]. The improvement in the relationships with various experts and free-flow of information from the academic site to the community will invariably yield improvements in patient outcomes.

Another available tool in building the community and academic network is the incorporation of guidelines and pathways into everyday practice. As the majority of oncologists in the community see a number of patients with varying histologies, it is often difficult to keep track of various therapies available, especially for lung cancer. While guidelines such as the National Comprehensive Cancer Network (NCCN) and the American Society of Clinical Oncology provide guidelines regarding the use of immunotherapy and targeted therapy, as well as genomic testing for FDA approved alterations [169], the results in our review show that the gaps in testing rates still remain prevalent and these guidelines are often difficult to interpret during a busy community practice. One proposed solution to this challenge is the implementation of vendor-based oncology clinical pathways (OCPs) that guide physicians in their decision-making based on query questions regarding the patient case [170]. A number of studies have shown that the use of OCPs not only maintains or improves outcomes, but they lower overhead costs for community practice [171–174]. While guidelines offer multiple recommendations that are difficult to interpret, clinical pathways create a local structure and framework from guidelines or evidence, with the goal of providing the single best therapeutic decision that provides value to the patient (Figure 3) [175]. The advantage of OCPs is not only the availability of decision-making support but the collection of analytics data that can be analyzed for research purposes and continuous quality improvement [176]. An OCP implemented in the community not only evaluates the performance of the community practice, but gives the tools to the community to drive improvements in testing rates and personalized therapy. The wide majority of community practice patients do not consider enrollment in clinical trials, as they are unaware of the option [177]. The pathways incorporate the clinical trials open within the entire enterprise, where trial decisions are placed ahead of other recommendations and always count as on-pathway, which encourages trial enrollment and integrates clinical trials into community practice. Our community practice utilizes the ClinicaPath (formerly ViaOncology) pathway systems in the decision-making process, but there are several vendors available [170].

**Figure 2.** The multidisciplinary care model for community and academic practice integration for lung cancer decision-making.

One recent development in our enterprise is the implementation of a standardized electronic health record (EHR) system in the community that mirrors the academic site medical records in a single system and allows for optimization of testing results and physician referrals for clinical trials. The standardization of molecular testing results and reporting in a fast and reliable manner through the medical record is an important barrier for community oncology practice towards improving testing rates [178]. The cohesiveness of a singular EHR not only results in clinical decision support, but allows the community oncologists to participate in the clinical and translational research process

through the evaluation of retrospective patient cohorts in a collaborative model that encompasses a multi-disciplinary team of pathologists, radiologists, and other specialties. The seamless amalgamation of high level genomic and treatment data from the community can be quickly extrapolated from the EHR and utilized in translational studies including evaluation of testing rates and therapy outcomes. This also helps in identifying patients that would be eligible for enrollment in clinical trials available at partnering academic sites, as evidenced by the top accrual rates of the adjuvant EVEREST study in renal cell carcinoma at City of Hope [179]. This is an especially significant strategy to implement in order to enroll and treat older cancer patients who are primarily seen at community sites [180]. Furthermore, the establishment of integrated clinical research has been shown to translate to wider awareness and acceptance of research results, and in 2013, the NCI formed the NCI Community Oncology Research Program (NCORP) [181]. First-cycle results showed that NCORP improved cancer care delivery and access in the community, but challenges remain in growing the program to more organizations across the nation [182]. The evolution of cancer care has to be met with advancements in cancer care and genomic testing access and delivery in community practice. However, the ultimate development of a successful community-based research program requires funding to empower local physicians, infrastructure to support implementation, collaboration between academic and community investigators, and flexibility in operations and organizations.

**Figure 3.** Advantage of guidelines and pathways in clinical scenarios. Patient outcomes are reliant on adherence to evidence-based medicine, which can be facilitated by guidelines and enhanced by pathways.

#### **4. Conclusions**

The advancements in lung cancer therapy and genomic testing have transformed the lung cancer decision-making process in the last decade. Next-generation sequencing has expanded from a few genes tested with routine testing to broad-based sequencing that has identified a plethora of oncogenes that are involved in driving the progression of NSCLC [183–185]. While targeted therapy was initially implemented in the first-line setting, the availability of a number of second- and third-generation TKIs has transitioned from a model of systemic therapy in the refractory setting to a framework of a number of TKIs administered in sequence based on resistance mechanisms and clinical progression of the individual patient [186]. The promise of personalized medicine continues to be realized through the development of ground-breaking immune checkpoint inhibitors and upcoming trials show promise

for chimeric antigen receptor (CAR) T-cell therapy [187]. To further realize this mission of precision medicine and to deliver improved outcomes, rigorous clinical data science, and translational research of the care delivery model and access have to be expanded beyond academic sites and into community practice. As we have brought to attention in this review, the community practice, while currently lagging behind academic sites in delivery oncology care, can be systematically and procedurally integrated with academic centers in a unified model for lung cancer decision-making and clinical collaboration. Our identified tools and collaborative concepts, including pathways and MTBs, can be realized in any community setting to enhance communication and trial enrollment.

**Author Contributions:** Conceptualization, R.S., S.R., P.K, I.M.; Writing—original draft preparation, S.R., I.M., R.P., R.S., P.K.; Writing—review and editing, S.R., I.M., R.P., R.S., P.K., B.L. and T.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** The work was supported by the National Cancer Institute of Health under award numbers P30CA033572, U54CA209978, R01CA247471 and R01CA218545

**Acknowledgments:** We would like to express our deepest gratitude for philanthropic funding by the Tenenblatt Family.

**Conflicts of Interest:** S.R: Speaker for Boehringer Ingelheim Pharmaceuticals Inc., Puma Biotechnology Inc.; I.M., R.P., B.L., T.T., P.K. and R.S. 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* **Integrating Academic and Community Practices in the Management of Colorectal Cancer: The City of Hope Model**

### **Misagh Karimi 1, Chongkai Wang 1, Bahareh Bahadini 2, George Hajjar <sup>2</sup> and Marwan Fakih 1,\***


Received: 27 March 2020; Accepted: 26 May 2020; Published: 2 June 2020

**Abstract:** Colorectal cancer (CRC) management continues to evolve. In metastatic CRC, several clinical and molecular biomarkers are now recommended to guide treatment decisions. Primary tumor location (right versus left) has been shown to predict benefit from anti-epidermal growth factor receptors (EGFRs) in rat sarcoma viral oncogene homologue (*RAS*) and v-raf murine sarcoma viral oncogene homolog B1 (*BRAF*) wild-type patients. Anti-EGFR therapy has not resulted in any benefit in *RAS*-mutated tumors, irrespective of the primary tumor location. *BRAF-V600E* mutations have been associated with poor prognosis and treatment resistance but may benefit from a combination of anti-EGFR therapy and BRAF inhibitors. Human epidermal growth factor receptor 2 (*HER-2*) amplification was recently shown to predict relative resistance to anti-EGFR therapy but a response to dual HER-2 targeting within the *RAS* wild-type population. Finally, the mismatch repair (MMR)-deficient subgroup benefits significantly from immunotherapeutic strategies. In addition to the increasingly complex biomarker landscape in CRC, metastatic CRC remains one of the few malignancies that benefits from metastasectomies, ablative therapies, and regional hepatic treatments. This treatment complexity requires a multi-disciplinary approach to treatment and close collaborations between various stakeholders in large cancer center networks. Here, we describe the City of Hope experience and strategy to enhance colorectal cancer care across its network.

**Keywords:** colorectal cancer; precisian medicine; academic and community oncology

#### **1. Introduction**

Colorectal cancer is the second most common cause of cancer-related death in the United States, with an annual incidence of 145,000 cases in 2019 and 51,000 deaths according to the American Cancer Society [1]. The life-time risk of developing colorectal cancer in men and women is 4.6% and 4.2%, respectively, according to 2014–2016 data [2]. The five-year survival rate for metastatic disease and regional disease are 14.2% and 71.3%, respectively [3,4]. Therefore, significant progress is still needed, especially in metastatic colorectal cancer settings.

In this report, we will review the latest approaches in the treatment of metastatic disease. We will also explore the City of Hope approach to delivering optimized care with partnership between academic researchers and community clinicians in cancer care.

#### **2. Our Path to Precision Medicine in the Treatment of Metastatic CRC**

The management of metastatic colorectal cancer must take into consideration sidedness, as well as molecular biomarkers including *RAS*, *BRAF*, *HER-2*, and MMR status. We have previously extensively reviewed this topic [5,6]. In short, it is now well-established that the benefits of anti-EGFR therapy appear to be limited to *RAS* and *BRAF* wild-type tumors that originate from the left colon [7–9]. For this group of patients, the addition of anti-EGFR therapy to combination chemotherapy in the first-line setting is better than bevacizumab according to subgroup analyses from two large randomized trials [10,11]. On the other hand, right-sided tumors as well as *RAS* mutated tumors benefit from the addition of bevacizumab to combination first-line chemotherapy. Second-line treatments are typically shaped by first-line treatment decisions and are addressed in our prior reviews [6].

*BRAF-V600E*-mutated colorectal cancers constitute approximately 8% of metastatic colorectal cancers. These tumors are associated with a poor prognosis and relative chemotherapeutic resistance [12–14]. Given their aggressive histology, and based on subgroup analyses from the TRIBE clinical trial, we advocate a combination of 5-FU, irinotecan, and oxaliplatin (FOLFOXIRI) with bevacizumab in the first-line treatment for those deemed to be fit enough to tolerate this regimen [15]. Otherwise, the first-line treatment of *BRAF-V600E*-mutated colorectal cancer is typically managed in a similar fashion as that of *RAS*-mutated metastatic colorectal cancer. A ray of hope has finally emerged in the targeted therapy of these *BRAF*-mutated patients. The BEACON trial has recently demonstrated that in the second-line and third-line settings, a combination of a BRAF inhibitor (encorafenib) and EGFR inhibitor (cetuximab) is better than a combination of chemotherapy (irinotecan with or without 5-FU) and cetuximab [16,17]. Encorafenib plus cetuximab is now to be considered a standard second-line therapy in these patients.

HER-2 amplification occurs in 2% of colorectal cancers and is enriched in left-sided and *RAS* and *BRAF* wild-type tumors [18]. These tumors exhibit a relative resistance to anti-EGFR therapy and respond well to lapatinib and trastuzumab or trastuzumab and pertuzumab, based on the HERACLES and MYPATHWAY trials, respectively [19,20]. This has prompted the National Comprehensive Cancer Network (NCCN) to recommend these treatments in the later lines of treatment for this molecular subgroup. The value of anti-EGFR therapy in these patients is still under investigation. We hope that the ongoing SWOG trial (S1613) will finally shed some light on this issue [21].

Finally, MMR deficiency has emerged as a predictive biomarker of response to PD-1 inhibitors, with or without CTLA-4 inhibitors, in colorectal cancer [22,23]. Nivolumab and pembrolizumab have shown remarkably durable responses in the second- and third-line treatment of these patients [24,25]. The addition of ipilimumab to nivolumab appears to enhance the responses and disease control rates, with promising first-line and beyond outcomes being reported from the CHECKMATE 142 trial [26,27]. Therefore, the NCCN has recommended the integration of these agents (monotherapy or combination) in the second-line (and beyond) treatment of MMR-deficient patients. In addition, the NCCN has recommended the consideration of immunotherapy in MMR-deficient frail metastatic colorectal cancer in first-line settings.

In addition to systemic therapies, one must acknowledge an important role for metastasectomies and ablative therapies in colorectal cancer. These should be important considerations for patients with oligometastatic disease—where surgical intervention can result in a curative outcome and/or improved survival [28,29]. Ablative therapies include microwave or radiofrequency ablation as well as stereotactic body radiation therapy. These are typically used in conjunction with or in lieu of surgery in an individualized fashion. Additional regional therapies in patients with liver-only or liver-predominant metastatic disease include radioembolization and hepatic arterial infusion. The discussion around surgery, ablative therapy, and regional therapy is beyond the scope of this manuscript. However, the above highlights the multi-disciplinary needs in the management of metastatic colorectal cancer.

#### **3. Integration of Academic and Community Oncology**

Achieving the best outcome in patient care has been the long-standing desire at City of Hope. While academicians design and conduct clinical trials, community physicians provide care to most patients and are critical to clinical trial enrollment and the application of standard of care therapy. The optimal partnership requires significant planning and efforts on both sides. A multimodality approach is becoming exceedingly important in the care of colorectal cancer and should be integrated seamlessly across academic and community practices. Here, we describe our efforts to enhance partnerships between our City of Hope Cancer Center and our associated Community Practice Satellites.

#### *3.1. Integrating Community Practices in Tumor Board Discussions*

Increasingly, the management of colorectal cancer requires the involvement of a multidisciplinary team including medical oncologists, radiation oncologists, pathologists, gastroenterologists, cancer geneticists, colorectal cancer surgeons, thoracic surgeons, and surgical oncologists. At City of Hope, we have conducted Gastrointestinal Oncology Tumor Boards on a twice-weekly basis (Mondays and Thursdays) with representative members from each of the disciplines above. **Tumor boards are disease specific and are run on a weekly basis, with email invitations generated to all interested community physicians. In general, community practices participate when they have an interesting case that requires input in a multidisciplinary setting.**

During these meetings, complex colorectal cases are discussed to determine the best treatment options. The recommendations made span from chemotherapy/immunotherapy/targeted-therapy refinement to decisions regarding metastasectomy, adjuvant therapy, radiation therapy, stereotactic body radiation therapy (SBRT), radioembolization, and hepatic artery infusion therapy. Community practice physicians present their cases remotely to these conferences, providing them with the opportunity to benefit from a multidisciplinary review of their cases and the receipt of a multispecialty input regarding a comprehensive approach towards colorectal cancer. **Such participation has altered treatment management in select cases (such as recommendations regarding adjuvant or neoadjuvant therapy or complex surgery), in allowing to link them to certain cases with appropriate clinical trials in our Duarte campus.**

#### *3.2. Integrating Clinical Trials in Community Practices*

Clinical trial access has become increasingly important for our colorectal cancer patients. However, the proximity of patients to a main center that provides these treatments has been and remains a main problem that has hindered patient enrollment. City of Hope is supporting a strong initiative to activate clinical trials in our various community centers as part of our mission to enhance research and improve patient access to novel therapeutics. We have partnered with our Community Practices to activate studies of interest to the community, with a strong focus on Phase II and III studies, early-phase investigator-initiated studies, and cooperative group trials. **The end result has been an increase in the accrual rate in community practices, where 70–100 patients have been enrolled in therapeutic clinical trials on a yearly basis.** This has particularly applied to colorectal cancer, where we have activated therapeutic trials that span first-line, second-line, and third-line studies (examples are listed in Table 1).


**Table 1.** Selection of colorectal cancer trials activated in City of Hope Community Practices.

Potential studies are vetted by community physicians for the feasibility of the associated research procedures and the availability and potential interest of an eligible colorectal cancer population (Figure 1). Only once a study is identified to be fit for a specific community practice is it endorsed for activation in that site. To exemplify, CanStem303C (NCT02753127) was activated through our enterprise to address the value of the cancer stem cell inhibitor BBI-608 in the second-line treatment of metastatic CRC. Out of 25 patients enrolled in this study across four research sites, 13 were enrolled in our main Cancer Center campus, while 12 patients were enrolled through three additional community centers.

**Figure 1.** Process for clinical trial activation in a City of Hope Community Practice.

### **4. Referring Routine and Complex Cases**

The management of colorectal cancer, whether metastatic or localized, is well-characterized and can be performed without significant barriers in community practices. The availability of a large network of providers across a large geographical area allows for easy access to the medical provider, less commuting time, and improved patient satisfaction. Most of the cases seen on our campus can therefore be managed in a more convenient location, which suits patients traveling a long distance from our Cancer Center. These options are discussed with our patients seen on our main Duarte campus, with an appropriate referral made to a more convenient City of Hope Community Practice for continuity of care (Figure 2).

**Figure 2.** Cross-referral patterns between the City of Hope main cancer center and community practices.

On another note, certain colorectal cancer patients that are treated in our community practices may benefit from complex specialized services that are only feasible in our main campus. These patients are referred for treatment and continuity of care from our satellite practices to our Duarte campus. For example, the use of hepatic arterial infusion pumps for regional chemotherapy in an adjuvant setting after hepatic metastasis resection requires special surgical expertise as well as a specialized supporting team (interventional radiology, nurses trained for pump access and troubleshooting, and oncologists experienced in regional hepatic chemotherapy). We have taken the conscious decision to centralize the management of these patients at our main campus.

#### **5. Standardization of Treatment Pathways in Colorectal Cancer**

The creation and standardization of treatment pathways are key to the administration of quality care across our network. City of Hope is a National Network Cancer Center Network (NCCN) member. We contribute to various committees in the NCCN and support its published guidelines [30]. However, the guidelines are broad and are not easy to navigate across our sprawling network of community practices. Several years ago, we joined the Via Oncology network (currently ClinicalPath), which provides an easy-to-navigate clinical pathway for medical oncologists across all disease sites. These guidelines strive to reduce variability, focus on cost-effectiveness, and seek patient-friendly (less toxic and easier to administer) regimens. Since these pathways are meant for academic and community practices alike, we have included representative members from our Cancer Center and Community Practices on the Gastrointestinal Cancer Via Oncology Committee. The committee meets on a quarterly basis and discusses recently published or presented clinical data that can impact the recommended Via Oncology treatment guidelines. Every cancer patient treated in our institute is navigated through these pathways electronically, and the data are reviewed to assess treatment adherence and guideline compliance.

#### **6. Educational E**ff**orts**

Educational programs (certified medical education) across a variety of cancers are hosted on a weekly basis on our Duarte Campus (Cancer Center). All faculty members across our satellite offices have the capability of attending in person or remotely. In addition, our medical oncology department hosts bi-annual symposia for medical oncology. During these meetings, each disease site, including colorectal cancer, is co-hosted by a Cancer Center Academician and a Community Practice physician. Community practice physicians leading such efforts have an existing interest and expertise in the assigned respective area and lead the discussions on the latest standards of care. Such programs increase the interaction between our research faculty and clinical faculty and enhance learning collaborations across our network.

#### **7. Conclusions**

The close interplay and collaboration between our Cancer Center and our Community Practices is essential to optimize clinical care for colorectal cancers across our community. These collaborations include educational activities, the standardization of treatment pathways, clinical trials, and the cross-referral of patients to address patient convenience and treatment complexity. The constant cross-talk between our Academic and Clinical Faculty across the network insures that the best standards in colorectal cancer are applied across our capture area.

**Author Contributions:** M.K., C.W., B.B., G.H., and M.F., contributed literature search and review, writing, graphical design, and editing; M.F. contributed conception and design and editing. All authors have read and agreed to the published version of the manuscript.

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

**Conflicts of Interest:** Fakih reports received Honoraria from Amgen and research funding from Astra Zeneca, Amgen and Novartis. Fakih reports serving as advisory for Amgen, Array, Bayer and Pfizer and as speaker bureau for Amgen and Guardant 360. All other authors declared no conflict of interests.

#### **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* **Managing Bladder Cancer Care during the COVID-19 Pandemic Using a Team-Based Approach**

### **Tina Wang, Sariah Liu, Thomas Joseph and Yung Lyou \***

Department of Medical Oncology & Experimental Therapeutics, City of Hope Comprehensive Cancer Center, Duarte, CA 91010, USA; tinawang@coh.org (T.W.); sarliu@coh.org (S.L.); thojoseph@coh.org (T.J.)

**\*** Correspondence: ylyou@coh.org; Tel.: +1-626-218-9200; Fax: +1-626-218-8233

Received: 14 April 2020; Accepted: 19 May 2020; Published: 22 May 2020

**Abstract:** The recent novel coronavirus, named coronavirus disease 2019 (COVID-19), has developed into an international pandemic affecting millions of individuals with hundreds of thousands of deaths worldwide. The highly infectious nature and widespread prevalence of this disease create a new set of obstacles for the bladder cancer community in both delivering and receiving care. In this manuscript, we address the unique issues regarding treatment prioritization for the patient with bladder cancer and how we at City of Hope have adjusted our clinical practices using a team-based approach that utilizes shared decision making with all stakeholders (physicians, patients, caregivers) to optimize outcomes during this difficult time. In addition to taking standard precautions for minimizing COVID-19 risk of exposure for those entering a healthcare facility (screening all personnel upon entry and donning facemasks at all times), we suggest the following three measures: (1) delay post-treatment surveillance visits until there is a decrease in local COVID-19 cases, (2) continue curative intent treatments for localized bladder cancer with COVID-19 precautions (i.e., choosing gemcitabine/cisplatin (GC) over dose-dense methotrexate, vinblastine, doxorubicin, cisplatin (ddMVAC) neoadjuvant chemotherapy), and (3) increase the off-treatment period between cycles of palliative systemic therapy in metastatic urothelial carcinoma patients.

**Keywords:** bladder cancer; urothelial carcinoma; COVID-19; team-based medicine

#### **1. Introduction**

Recently, a novel coronavirus, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has developed into an international pandemic affecting millions of individuals in more than 150 countries with hundreds of thousands of deaths worldwide [1,2]. This disease has been named coronavirus disease 2019 (COVID-19) by the World Health Organization (WHO) [1]. Patients with this disease are at high risk for developing septic shock and hypoxemia, which can frequently progress to acute respiratory distress syndrome (ARDS) and death [3]. This disease creates a new set of obstacles for the bladder cancer community in both delivering and receiving care. In this manuscript, we address the unique issues regarding treatment prioritization for the patient with bladder cancer and how we at City of Hope have adjusted our clinical practices using a team-based, shared decision approach with all stakeholders (patients, caregivers, and physicians) to optimize outcomes during this difficult time.

### **2. Balancing the Need for Bladder Cancer Treatments and Risk of Exposure to COVID-19**

*2.1. Patients with Bladder Cancer Undergoing Treatments Are at a Higher Risk for COVID-19 Infections and Worse Outcomes Compared to the General Population without Cancer*

For the patient with bladder cancer undergoing treatment, there are several safety issues that place them at higher risk of infection for COVID-19 compared to the general population without cancer. First, patients must physically leave the safety of their residences to go to the clinic, infusion center,

or imaging facility where they could potentially be exposed to COVID-19. Second, the platinum-based chemotherapy regimens commonly used in bladder cancer treatments are immunosuppressive and place them at a higher risk for infection. Third, many bladder cancer patients tend to be of older age and also have multiple medical comorbidities, which has been shown to place them in a group with worse outcomes for COVID-19 [2,4]. A retrospective study that examined the outcomes of approximately 72,000 patients with COVID-19 found that those with older age and presence of medical comorbidities were associated with adverse outcomes [2,4]. In another retrospective study by Liang and colleagues, it was suggested that patients with a history of cancer itself may be associated with worse outcomes from COVID-19 [5,6]. However, it should be noted that this particular retrospective study was limited in that only 18 of the 1590 patients who were studied had a history of cancer, making it difficult to form a general conclusion from such a small sample size [5,6]. Regardless, based on the other reasons discussed above, it is clear that patients with bladder cancer undergoing active therapy or post-treatment surveillance are at a higher risk for COVID-19 exposure and could potentially suffer worse outcomes compared to the general population.

#### *2.2. Prioritizing Treatments Appropriately and Applying Social Distancing*

Ensuring patient safety is the key principle when it comes to delivering medical care among all healthcare professions. In the setting of the COVID-19 pandemic, the central question we have asked ourselves as providers while managing each patient's care has been: Will delaying the patient's bladder cancer treatment in accordance with current COVID-19 social distancing measures lead to a worse long-term outcome? Current models suggest that this pandemic may proceed until herd immunity or a vaccine is developed, with repeated waves of infections, which some experts estimate could continue for another 18 months. Since it is not feasible to delay bladder cancer treatments for another 18 months, we at City of Hope have developed a consensus framework to help balance these competing risks (Figure 1). By utilizing this framework, we have been able to guide our clinicians within the network on how to make a shared decision with the patient that can prioritize bladder cancer treatments appropriately while minimizing the risk for COVID-19 exposure (Figure 1).

**Figure 1. Conceptual framework for prioritizing bladder cancer treatments during the COVID-19 pandemic.** This framework provides guidance on key treatments that should still be offered in order to ensure optimal bladder cancer outcomes if possible. We recommend that these listed priorities can be modified based on available local resources and the patient's overall medical status.

#### *2.3. Applying COVID-19 Risk Mitigation Measures for Bladder Cancer Treatment*

In the state of California, there is a "shelter in place" order that was initiated on 19 March 2020 along with other social distancing measures due to the concern that individuals may be at high risk of becoming infected and could also infect others, further propagating this pandemic. Current epidemiology modeling suggests that the peak incidence of COVID-19 will have occurred sometime in mid-to-late April in the state of California. This framework assumes that the number of new cases will start to decrease in the months of May and June 2020. In the case that there is indeed a second wave of infections later during the fall and winter months of 2020, one could reapply this framework based on the expected peaks. As a result, we suggest the following framework to assist the practicing oncologist in determining optimal treatment strategies for the patient with bladder cancer.

#### 2.3.1. Delay Post-Treatment Surveillance Visits until There Is a Decrease in COVID-19 Cases

For patients undergoing surveillance imaging after completion of cystectomy or other definitive therapies, the National Comprehensive Cancer Network (NCCN) guidelines currently recommend imaging every 6–12 months [7]. Keeping these guidelines in mind, we have rescheduled the patient's clinic and imaging visit to avoid the expected COVID-19 peak period (April–May) so that it will take place during the next 2–3 months in June or July as a way to minimize risk of exposure.

#### 2.3.2. Continue Curative Intent Treatments for Localized Bladder Cancer with COVID-19 Precautions

Even in these difficult times, urothelial bladder cancer is an aggressive disease with poor prognosis when it progresses to metastatic disease. Therefore, we have been vigilant in continuing to deliver curative intent treatments when patients have localized urothelial carcinoma, if possible in a timely manner. A meta-analysis of 13 studies suggested that a delay of more than 12 weeks from time of diagnosis to execution of radical cystectomy, only in muscle invasive urothelial cancer, was associated with worse outcomes [8]. Another study showed that initiating neoadjuvant chemotherapy (NAC) with a delay of more than 8 weeks from time of diagnosis led to worse outcomes [9]. Therefore, we have continued to offer cisplatin-eligible patients NAC within 8 weeks and cisplatin-ineligible patients radical cystectomy within 12 weeks from time of diagnosis, while the infusion center and operating room resources are available for those patients with localized disease since there is a limited window of curative treatment opportunity.

The first set of measures we have instituted to minimize potential risk for COVID-19 exposure for all on-site people (visitors and healthcare workers) is to create a single, separate point of entry to the active clinical areas and institute a strict policy limiting visitors to patients only. Prior to entering the clinical area, all personnel (including patients and healthcare workers) are screened for COVID-19 symptoms (i.e., cough, dyspnea, and fever) and have their temperatures measured. People determined to be asymptomatic and afebrile are then required to don a face mask and are issued an entrance band indicating they have passed screening measures for that day. If someone is found to be symptomatic, we then refer this individual to an on-site "fever clinic" staffed by designated clinical personnel who have been trained and equipped with the appropriate personal protective equipment (PPE) to perform a nasopharyngeal swab for in-house COVID-19 testing. We have also repurposed one of our hospital wards with negative pressure rooms to serve as the COVID-19 unit with its own set of designated staff to decrease exposure within the facility. In both the inpatient and outpatient areas, all people (patients and healthcare workers) are required to don a face mask at all times, which has been suggested as a way to prevent sustained exposure to COVID-19 and reduce risk for infection [10].

The second set of COVID-19 risk mitigation measures specifically pertain to treatments used for urothelial bladder cancer. Current NCCN guidelines recommend neoadjuvant chemotherapy in muscle invasive bladder cancer [7]. In the choice of regimen, the two most commonly used regimens are dose-dense methotrexate, vinblastine, doxorubicin, cisplatin (ddMVAC) and gemcitabine/cisplatin (GC) [7,11,12]. During this time, we have advocated for using gemcitabine/cisplatin over ddMVAC for the following reasons. Although there is some discussion suggesting that ddMVAC may have a trend towards higher efficacy, it has yet to be definitively supported in a head-to-head prospective trial and retrospective studies have shown similar amounts of efficacy between these two regimens [11,12]. In addition, ddMVAC tends to be more myelosuppressive than GC, placing patients at higher risk for infections due to the neutropenia and symptomatic anemia requiring blood transfusions, which during this time have been especially challenging due to a steep drop in blood donations [11,12]. Finally, ddMVAC is given as a 14-day cycle whereas GC is given as a 21-day cycle. The 14-day cycle of ddMVAC allows a patient to proceed sooner to radical cystectomy compared to GC, but during this time we would recommend GC because it allows the oncologist to space out the patient visits and can help adhere better to the principle of social distancing [11,12]. Another measure we have taken is to implement weekly telephone checks with patients undergoing active systemic therapy. This allows us to determine if a patient is having any significant chemotherapy-related adverse effects or other acute medical issues, for which they could potentially be treated as an outpatient before they progress to needing emergency room or acute inpatient care. For example, if a patient is experiencing significant dysuria due to a potential urinary tract infection, one can prescribe antibiotics empirically at their local pharmacy and help them avoid the need to seek emergency room care, which is most likely to be overcrowded during this pandemic. For those patients that are undergoing concurrent radiation and chemotherapy with curative intent, we have continued their treatments while taking the abovementioned general COVID-19 precautions (i.e., screening at entry, donning facemasks, and weekly telephone checks).

2.3.3. Increasing Off-Treatment Period between Cycles for Palliative Systemic Therapy in Metastatic Urothelial Carcinoma Patients

For those patients already undergoing palliative first-line systemic therapy, we have continued their treatments as those regimens provide overall survival benefit. In this situation, if chemotherapy needs to be started we would recommend, as discussed above, to prescribe anti-emetics and pain medications for the patient to have immediately available at home as an outpatient. Additionally, weekly telephone checks would be conducted to prevent any chemotherapy-related complications early. Another important factor to consider, as discussed above, is lengthening the period of time between treatments. Normally, gemcitabine/cisplatin or gemcitabine/carboplatin is administered as a two weeks on, 1 week off schedule. In this case, it is reasonable to do 2 weeks on, 2 weeks off to help spread out the treatment duration as much as possible to maximize social distancing. Second-line treatment usually involves the use of immune checkpoint inhibitors such as pembrolizumab or atezolizumab. Pembrolizumab is dosed every 3 weeks, but in order to maximize social distancing for the patient, it is reasonable to stretch it to every 4 weeks during this period since it is unlikely the cancer will grow significantly during the extra week off. In this setting, atezolizumab and nivolumab already has an Federal Drug Administration (FDA)-approved every-4-week dosing, which would also make it a viable alternative. The use of third-line treatment with enfortumab vedotin requires administration once every week for 3 weeks straight and then taking the fourth week off. Again, to provide more space between visits, it would be reasonable to increase the off-treatment period from 1 week to 2 weeks to provide the patient more social distancing.

Even during these difficult times, it is crucial to continue clinical trials to the best of our ability and help advance the field of oncology. In order to preserve needed resources for COVID-19 prevention and treatment within our institution, we have focused our efforts on continuing current open clinical trials and slowing down the pace of opening new trials.

#### **3. Conclusions**

COVID-19 has developed into an international pandemic affecting millions of individuals and has created a new set of obstacles for the bladder cancer community in both delivering and receiving care. Because patients with bladder cancer require treatment even in these difficult times, we have developed a framework that utilizes a team-based approach with shared decision making among all stakeholders involved (physicians, patients, caregivers) to optimize outcomes during this difficult time. It is our hope that the conceptual framework presented above and institutional experience can be adjusted to fit the available local resources for others that are looking to balance these two competing needs when treating patients with bladder cancer during the COVID-19 pandemic.

**Author Contributions:** Y.L. conceived the article; T.W., S.L., and T.J. wrote the manuscript with input from Y.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Cancer Institute of the National Institutes of Health under grant number P30CA033572 to City of Hope Comprehensive Cancer Center.

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

#### **Abbreviations**


#### **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* **Strategies to Improve Participation of Older Adults in Cancer Research**

**Jennifer Liu 1,**†**, Eutiquio Gutierrez 2,**†**, Abhay Tiwari 1, Simran Padam 1, Daneng Li 1, William Dale 3, Sumanta K. Pal 1, Daphne Stewart 1, Shanmugga Subbiah 1, Linda D. Bosserman 1, Cary Presant 1, Tanyanika Phillips 1, Kelly Yap 1, Addie Hill 1, Geetika Bhatt 1, Christina Yeon 1, Mary Cianfrocca 1, Yuan Yuan 1, Joanne Mortimer <sup>1</sup> and Mina S. Sedrak 1,\***


Received: 17 April 2020; Accepted: 19 May 2020; Published: 22 May 2020

**Abstract:** Cancer is a disease associated with aging. As the US population ages, the number of older adults with cancer is projected to dramatically increase. Despite this, older adults remain vastly underrepresented in research that sets the standards for cancer treatments and, consequently, clinicians struggle with how to interpret data from clinical trials and apply them to older adults in practice. A combination of system, clinician, and patient barriers bar opportunities for trial participation for many older patients, and strategies are needed to address these barriers at multiple fronts, five of which are offered here. This review highlights the need to (1) broaden eligibility criteria, (2) measure relevant end points, (3) expand standard trial designs, (4) increase resources (e.g., institutional support, interdisciplinary care, and telehealth), and (5) develop targeted interventions (e.g., behavioral interventions to promote patient enrollment). Implementing these solutions requires a substantial investment in engaging and collaborating with community-based practices, where the majority of older patients with cancer receive their care. Multifaceted strategies are needed to ensure that older patients with cancer, across diverse healthcare settings, receive the highest-quality, evidence-based care.

**Keywords:** geriatric oncology; older adults; cancer clinical trials; recruitment; community; team science

#### **1. Introduction**

Aging is a major risk factor for cancer, with 28% of cancers in the US diagnosed in adults aged 65–74, 18% in adults aged 75–84, and 8% in adults aged 85 and older [1]. Globally, the US has the third largest number of older adults [2], and as the US population ages, the number of older adults with cancer is increasing and will make up a growing share of the oncology population [3–5]. Despite this, older adults remain vastly underrepresented in the research that sets the standards for cancer treatments [6,7]. Consequently, most of what is known about cancer therapeutics is based on clinical trials conducted in younger and healthier patients [8–10]. Furthermore, despite a plethora of

literature documenting barriers to the accrual of older adults to cancer clinical trials (Figure 1), there are few evidence-based strategies to mitigate these barriers [11,12]. A recent systematic review revealed that among 8691 studies screened, only 12 relevant observational studies examined barriers hindering the participation of older adults in cancer clinical trials, and one (negative) randomized controlled trial evaluated an intervention to increase the enrollment of older adults in trials [8]. Moreover, few studies have focused on understanding the complex and multifactorial influences affecting the clinical trial participation of older patients with cancer in the community, where the majority of this population is often treated [13,14].

**Figure 1.** The barriers to older adult participation in cancer clinical trials are multifaceted and interrelated, with barriers existing at the system, clinician, patient, and caregiver levels. System-related barriers include trial design, overly stringent eligibility criteria, and lack of infrastructure support, appropriate and representative trials, and funding. Clinician-related barriers include concerns for side effects or toxicities, patient age, comorbid conditions, personal bias, costs to clinicians, and lack of time, support and staff, awareness of trials, and engagement amongst clinicians. Patient-related barriers include attitudes towards clinical trials, knowledge, side effects/toxicities, burden, and financial limitations. Caregiver-related barriers include caregiver burden and preference against participation.

Recognizing this problem, numerous organizations, including the Institute of Medicine (IOM), have cited the growing population of older adults with cancer and highlighted the need to generate high-quality, evidence-based data for the treatment of older adults [6,15–18]. The primary aim of this review is to lay out a multipronged approach that addresses barriers to the participation of older adults in cancer clinical trials. A secondary aim is to highlight how a collaborative clinical research network of academic and community-based oncology practices can facilitate the inclusion of older adults in cancer research.

#### **2. Make Eligibility Criteria More Inclusive and Less Restrictive**

Restrictive eligibility criteria often exclude a large proportion of patients from participating in clinical trials—a loss which sacrifices the generalizability of results to the overall patient population. Older adults, in particular, are often not offered the opportunity to participate in trials due to concerns for multiple comorbidities, organ dysfunction, treatment toxicity, and/or frailty [8,19–22]. Numerous trials explicitly restrict inclusion to patients with an Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 1 or a Karnofsky performance score (KPS) of ≥70 [23,24]. As a

result, trial data and findings are often derived from younger patients who are more fit and without organ dysfunction and comorbidities, as commonly seen in older adults [23].

A variety of efforts are underway to broaden the eligibility criteria to make trials more relevant to patients of all ages, including initiatives by the National Institutes of Health ("Inclusion Across the Lifespan policy"), American Society of Clinical Oncology (ASCO), and the Food and Drug Administration (FDA) [16,25,26]. Sponsors and investigators are encouraged to work together to follow these policy recommendations aimed at making the eligibility criteria less restrictive and more inclusive of demographically and clinically diverse patients, representative of the populations seen in community-based oncology settings. Investigators should review the eligibility criteria closely and forgo exclusion on the basis of lab values (e.g., creatinine), performance status, comorbid conditions, or second malignancy when designing clinical trials. Instead, a patient's functional or biological age should be taken into consideration, which may be a better indicator of how a patient will tolerate a certain treatment regimen.

Additionally, investigators should work closely with community clinicians and other stakeholders (i.e., patient advocates) to understand the patient populations seen in community-based practices and design eligibility criteria that are more inclusive, thus allowing trial results to be applicable to a broader patient population. Inclusive clinical trial systems within a collaborative research network ought to consider patients as they are, rather than as they should be. Trial eligibility should be evaluated and revised to ensure that investigators are broadening study access to new, successful cancer treatment regimens for older adults with cancer across diverse healthcare delivery settings.

#### **3. Capture Relevant End Points that Matter to Older Patients**

Sponsors and investigators should capture relevant end points that are important to the older patient, including measures of tolerability as well as clinical and biological aging consequences of cancer and its treatment [17,27–29].

Studies have shown that most older adults do not want to compromise their quality of life or function for survival benefits [30]. Hence, trials designed to collect information on this population must include patient-reported measures that are relevant to the patient's experience and move beyond clinician-reported toxicity. Therefore, data are needed to understand the immediate (short-term) and longitudinal (long-term) impact of cancer and cancer therapy on patient health and quality of life [29,31]. Given that each individual is unique and can respond differently to therapy, incorporating assessments that focus on patient-reported outcomes may allow investigators to better describe a patient's tolerability beyond the traditional numerical grades (NCI CTCAE) [32–34], and the use of mixed-methods approaches (e.g., surveys, interviews, focus groups) may allow investigators to better understand a patient's experience and priorities for treatment. Furthermore, measuring the frequency of toxicities over time can help investigators to characterize the longitudinal impact of treatment and better determine the relevance to the older patient population. This information can inform new ways treatment approaches and optimal strategies to reduce toxicities, avoid early treatment discontinuation, and achieve an effective dose intensity.

In addition to capturing measures of quality of life and tolerability, clinical measures of aging should also be considered as end points in cancer trials of older adults. This is important because aging is a heterogeneous process. While certain declines in organ function are universal as the human body ages, the consequences of this decline on everyday function proceed at a unique pace in each individual [35,36]. Therefore, chronologic age tells us relatively little about a specific individual [37]. A more detailed evaluation of an older patient is needed to capture factors that more effectively predict morbidity and mortality. A geriatric assessment (GA) may serve this purpose. The GA includes an evaluation of functional status, co-morbid medical conditions, cognitive function, nutritional status, social support, and psychological state, as well as a review of medications [38]. Additionally, geriatric screening tools (i.e., G8, VES-13), instruments that assess life expectancy (i.e., ePrognosis), and predictors for risks of chemotherapy toxicity (i.e., CARG or CRASH Score) can

be utilized in trials to better describe or risk-stratify this population [39–41]. The ASCO guidelines now recommend the use of these assessments in the evaluation and management of older patients with cancer, and several NCI-sponsored cooperative groups have demonstrated the feasibility of using these measures in clinical trials [40].

Biological measures of aging are also important to consider when designing studies for older patients with cancer. These measures may provide insights on mechanisms behind aging-related clinical consequences due to cancer treatment, such as the biological drivers of functional and cognitive decline. Several studies have hypothesized that cancer and cancer treatment may accentuate or accelerate the rate of aging, leading to a decrease in multisystem reserve and increased risk for cardiomyopathy, secondary malignant neoplasms, frailty, muscular weakness, and neurocognitive issues [42]. Biological processes including stem cell exhaustion, cellular senescence, telomere attrition, increased free radical production, and epigenetic modifications have all been shown to play a role in accelerated aging due to cancer therapies [42,43]. Understanding the hallmarks of aging and the implications of cancer therapies associated with accelerated aging may provide insights into the shortand long-term effects of certain therapies, and further research in this area is needed. Knowledge of the underlying mechanisms to accelerated aging could ultimately help us to identify new strategies for targeting these processes in order to ameliorate accelerated aging and improve the health and well-being of this growing population.

#### **4. Optimize Trial Designs for a Special Population**

Inclusion of older patients may, in some cases, impose the need for larger sample sizes, which is not always feasible or fundable. Sponsors and investigators should consider high-yield approaches to collect data on older patients to efficiently facilitate the conduct of larger-scale trials at a reasonable cost. For example, pragmatic or innovative trials (e.g., adaptive, extended, embedded, N-of-1 studies) designed specifically for older patients may be used to fill knowledge gaps and improve the evidence base guiding the treatment of older adults [29,44].

There are several innovative trial designs that may be leveraged when considering a special population such as older adults with cancer (Table 1 adapted from prior literature [16,39]). For example, adaptive trials can be leveraged to allow for modifications to be made as the study proceeds [45,46]. Based on interim data analysis, the underperforming treatment arm may be eliminated to allow for a larger proportion of participants to be assigned to the more effective treatment arm. This is ideal for older adults because it reduces the number of participants in the treatment group that is performing poorly [16,39,47]. While adapted trial designs can be applicable to both exploratory and confirmatory studies, prospective cohort studies can be used to generate data on current standard-of-care treatments in older patients and provide insight into patterns of care and decision making [16,39].

Extended trial design can also be used in cases where the results of a trial have been reported, but there was an insufficient number of older adults enrolled to draw conclusions. A cohort of older adults can be added to the superior treatment arm to fill knowledge gaps and obtain data on the older population [16]. Embedded studies, also known as correlative or ancillary studies, can also be used to include additional measures of interest that are specific to older adults (i.e., toxicity, GA domains) within the infrastructure of a parent study [39]. Lastly, an N-of-1 or single-subject trial presents a feasible and innovative approach to better understanding how to care for older adults by following one patient over time, to examine aging-related consequences that occur throughout the cancer continuum [48].



When designing clinical trials appropriate for the older patient population, incorporating older adults into the study design phase, such as through the utilization of community-based participatory research methodology, may facilitate their participation in clinical trials [49,50]. Moreover, it is important for sponsors and investigators to engage all key stakeholders, including patient advocates and community-based clinicians, in the process of trial design in order to understand the types of patients seen in community practices (Figure 2). These discussions can inform protocol design and ensure participants are more representative of the patient populations beyond the academic setting. Furthermore, these collaborations can foster improved accrual, conduct, and dissemination of the research, thus ultimately increasing generalizability and clinical relevance of the trial findings.

**Figure 2.** Strategies to connect academic centers and community-based practices. Ensuring collaboration between academic centers and community sites is essential in improving the participation of older adults in cancer research. First, academic centers should invest resources such as study personnel, regulatory oversight, and funding to community sites in order to help create trials and participation where most patients receive treatment. Second, both community sites and academic centers need to exchange tools, advice, and guidance to help increase the quality of trials being done and make sure that the needs of older adults are met. Lastly, community sites must have their clinicians participate in the designing of trials and eligibility criteria so that trials better fit the present patient population.

#### **5. Increase Institutional Support, Interdisciplinary Care, and Telehealth Use**

Structural barriers at the system or institutional level dominate trial decision-making, and successful participation in research requires substantial institutional guidance, resources, and infrastructure [51]. There are many barriers to the conduct and support of a clinical trial research program. Studies have shown that community-based practices often struggle with understanding the value of trials, covering the costs of supporting a research program, meeting program requirements, managing the clinic workflow changes as they pertain to clinician involvement, and sustaining hospital leadership support, among other barriers [13,52,53]. In order to overcome these barriers to entry, partnerships with larger practices or academic centers with existing infrastructure can facilitate the conduct of clinical trials at community practices. Additionally, leadership is needed at the institutional level to ensure that the mission, interests, and workflow of community sites are aligned with the academic medical centers in order to cover the ever-increasing requirements of implementing and maintaining a clinical research program, especially one that aims to increase the representation of older adults with cancer (Figure 2) [54].

In addition to institutional leadership, accessibility to specialized geriatric care may be an important facilitator in the recruitment and retention of older adults in cancer clinical trials. Clinical programs that meld geriatric and oncology communities to meet the complex needs of older adults may facilitate the improved management of older adults [17,30,39]. City of Hope, for example, has jointly launched two interdisciplinary clinical programs: (1) the Specialized Oncology Care & Research in the Elderly (SOCARE) clinic and (2) the Aging Wellness Clinic. Similar models have also been established at the Memorial Sloan Kettering Cancer Center, Thomas Jefferson University, and the University of

Rochester, among others [55–57]. These specialized programs offer interdisciplinary, individualized, and integrated treatment for older patients and survivors with cancer. However, these programs require resources and institutional support, and further research is warranted to understand how these care models lead to improved patient outcomes and participation in clinical trials.

Furthermore, the integration of technology in the form of virtual visits and telehealth encounters may facilitate the enrollment of older adults in cancer research. There are two potential strategies for this. First, technology may be leveraged to reduce the burden of the patient and caregiver—a known barrier to older participation in research [8,17,39]. Studies have shown that telehealth use is associated with high patient satisfaction, improved access to care, and reduced health disparities among older adults [58–62]. Challenges to telehealth use among the older patient population must be taken into consideration, including practical barriers to adoption, anxiety using technology, and financial burden, among other factors that promote the digital divide [53]. To address some of these barriers, simple and intuitive graphical interfaces, the availability of technical support, and the provision of telehealth services at no cost may reduce the barriers to entry and improve adoption of telehealth in this population [63]. However, further efforts are needed to better understand the limits and benefits of telehealth in geriatric oncology. Moreover, the COVID-19 pandemic has led to the rapid implementation of telehealth services for both standard care and clinical research [64,65]. This swift paradigm shift in how patients receive their care provides a unique opportunity to gain insight on how telehealth can further be used to improve patient participation in cancer clinical trials—particularly in older adults. However, the feasibility, adoption, and sustainability of telehealth in clinical trials is limited, and further research is needed [66].

Second, telehealth can be a key player in providing specialized geriatric care to sites that may not have access to multidisciplinary programs with geriatric expertise. For example, City of Hope has an ongoing pilot study to evaluate the feasibility of delivering a GA-driven intervention at a community-affiliated site using telehealth. In this way, telehealth may help fill the gap in care for older patients who otherwise would not have access to multidisciplinary, specialized geriatric-based care [67]. Whether this improves clinical trial accrual warrants further investigation.

#### **6. Leverage Principles of Behavioral Economics**

There is a growing interest in the general medical literature around the use of behavioral economics to shape clinician and patient behavior, with some even suggesting the use of "nudges" to facilitate patient enrollment in research [68]. A nudge is defined as a change in the way choices are presented or information is framed that alters a person's behavior in a predictable way without restricting choice [68,69]. If implemented properly, nudges are transparent, easy to opt out of if needed, and aligned with the best interest or welfare of the person being nudged. It is a principle of behavioral economics that leverages the fact that our decisions and behaviors are heavily influenced by the environment in which they occur, and interventions can be systematically developed to influence how individuals behave (e.g., weight loss, exercise, and statin utilization) [70–74].

Nudges may also be a novel way to facilitate patient participation in cancer clinical trials, especially among highly underrepresented groups such as older adults [75]. For example, surveys that assess patients' interest in participating in clinical research can help identify eligible patients who are likely to participate and provide insight into the types of trials that are desirable. Using these interest surveys, clinicians can identify patients who may be more inclined to participate in studies when presented the opportunity to do so due to their prior expressed interest (i.e., foot-in-the-door technique) [75]. Another nudging strategy that can be utilized to encourage the enrollment of older adults in cancer trials is the personalization of trials [75]. Explaining why a patient was specifically chosen to participate, how they may personally benefit, and how their participation will contribute to the scientific community and future patients may improve patient enrollment. Furthermore, nudges can be directed towards providers. Clinicians, for example, can be nudged by incorporating the trial information in the clinical pathway of electronic health records [69]. Including the trial information in

electronic health records may improve clinician awareness of potentially beneficial trials regarding their patients and help facilitate enrollment of older adults.

Evidence on the use of nudges to influence patient and clinician behavior in the context of clinical trials is still in its infancy. Further research is warranted to examine whether these strategies can be employed to influence enrollment and how they can be successfully implemented (e.g., feasible, acceptable, and sustainable) in both academic and community-based cancer practices.

#### **7. Conclusions**

The underrepresentation of older adults in cancer clinical trials is undeniably a multifaceted problem that requires a multifaceted solution. There have been promising steps toward improving trial participation among older adults, but there are still significant gaps in knowledge that hinder older patients from receiving high-quality, individualized, evidence-based care. The recommendations presented in this review range from small changes that can be adopted by individuals and research teams to large-scale, systemic changes that are needed at the institutional and/or policy level. Regardless, multiple steps are needed on multiple fronts across a collaborative network of academic and community-based cancer practices in order to have a cumulative and palpable effect. An investment in academic and community clinical trial partnerships could help to further cancer research, and more importantly, ensure that older adults with cancer have equal access to new treatments and advances in care.

**Author Contributions:** Conceptualization, J.L., E.G., A.T., S.P., and M.S.; Writing—Original Draft Preparation, J.L., E.G., A.T., S.P., and M.S.; Writing—Review and Editing, all authors; Visualization, J.L. and A.T.; Supervision, M.S.; Project Administration, M.S.; Funding Acquisition, M.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Institute of Aging (NIA R03AG064377), the National Cancer Institute (NCI K12CA001727), the Waisman Innovation Fund, and Circle 1500. The funder had no role in the design and conduct of the study; in the collection, management, analysis, and interpretation of the data; in the preparation, review, or approval of the manuscript; or in the decision to submit the manuscript for publication.

**Acknowledgments:** We dedicate this work to the late Arti Hurria, a leader in geriatric oncology and an advocate for expanding inclusion of older adults in cancer clinical trials. We hope this manuscript will help carry Hurria's legacy forward and foster new collaborations in a way that she would have advocated for.

**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/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Journal of Clinical Medicine* Editorial Office E-mail: jcm@mdpi.com www.mdpi.com/journal/jcm

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18