**Neoadjuvant Chemotherapy in Breast Cancer: An Advanced Personalized Multidisciplinary Prehabilitation Model (APMP-M) to Optimize Outcomes**

**Alba Di Leone 1, \* , Daniela Terribile 1 , Stefano Magno 1 , Alejandro Martin Sanchez 1 , Lorenzo Scardina 1 , Elena Jane Mason 1 , Sabatino D'Archi 1 , Claudia Maggiore 2 , Cristina Rossi 2 , Annalisa Di Micco 2 , Stefania Carnevale 3 , Ida Paris 4 , Fabio Marazzi 5 , Valeria Masiello 5 , Armando Orlandi 6 , Antonella Palazzo 6 , Alessandra Fabi 7 , Riccardo Masetti <sup>1</sup> and Gianluca Franceschini 1**

	- <sup>2</sup> Centre of Integrative Oncology—Multidisciplinary Breast Centre—Dipartimento Scienze della Salute della Donna e del Bambino e di Sanità Pubblica, Fondazione Policlinico Universitario A. Gemelli I RCCS, 00168 Rome, Italy; claud.maggiore@gmail.com (C.M.); cristina.rossi13@yahoo.it (C.R.); annalisadimicco@nutrimentidimindfulness.it (A.D.M.)
	- <sup>3</sup> UOS Psicologia Clinica, Fondazione Policlinico Universitario A. Gemelli IRCCS, 00168 Rome, Italy; dott.stefaniacarnevale@gmail.com
	- <sup>4</sup> Department of Woman and Child Health and Public Health, Woman Health Area, Fondazione Policlinico Universitario A. Gemelli I RCCS, 00168 Rome, Italy; ida.paris@policlinicogemelli.it
	- <sup>5</sup> UOC di Radioterapia Oncologica, Dipartimento di Diagnostica per Immagini, Radioterapia Oncologica ed Ematologia, Fondazione Policlinico Universitario A. Gemelli I RCCS, 00168 Rome, Italy; fabio.marazzi@policlinicogemelli.it (F.M.); valeria.masiello@policlinicogemelli.it (V.M.)
	- <sup>6</sup> Comprehensive Cancer Center, Multidisciplinary Breast Unit, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Largo Agostino Gemelli, 8, 00168 Rome, Italy; armando.orlandi@policlinicogemelli.it (A.O.); antonella.palazzo@policlinicogemelli.it (A.P.)
	- <sup>7</sup> Medicina di Precisione in Senologia, Dipartimento Scienze della Salute della Donna e del Bambino e di Sanità Pubblica, Fondazione Policlinico Universitario A. Gemelli IRCCS, 00168 Rome, Italy; alessandra.fabi@policlinicogemelli.it

**Abstract:** Neoadjuvant chemotherapy is increasingly being employed in the management of breast cancer patients. Efforts and resources have been devoted over the years to the search for an optimal strategy that can improve outcomes in the neoadjuvant setting. Today, a multidisciplinary approach with the application of evidence-based medicine is considered the gold standard for the improvement of oncological results and patient satisfaction. However, several clinical complications and psychological issues due to various factors can arise during neoadjuvant therapy and undermine outcomes. To ensure that health care needs are adequately addressed, clinicians must consider that women with breast cancer have a high risk of developing "unmet needs" during treatment, and often require a clinical intervention or additional care resources to limit possible complications and psychological issues that can occur during neoadjuvant treatment. This work describes a multidisciplinary model developed at "Fondazione Policlinico Universitario Agostino Gemelli" (FPG) in Rome in an effort to optimize treatment, ease the application of evidence-based medicine, and improve patient quality of life in the neoadjuvant setting. In developing our model, our main goal was to adequately meet patient needs while preventing high levels of distress.

**Keywords:** breast cancer; neoadjuvant chemotherapy; multidisciplinary treatment; evidence-based medicine; personalized treatment; oncological outcomes; patient quality of life

**Citation:** Di Leone, A.; Terribile, D.; Magno, S.; Sanchez, A.M.; Scardina, L.; Mason, E.J.; D'Archi, S.; Maggiore, C.; Rossi, C.; Di Micco, A.; et al. Neoadjuvant Chemotherapy in Breast Cancer: An Advanced Personalized Multidisciplinary Prehabilitation Model (APMP-M) to Optimize Outcomes. *J. Pers. Med.* **2021**, *11*, 324. https://doi.org/10.3390/jpm11050324

Academic Editor: Raghu Sinha

Received: 18 March 2021 Accepted: 12 April 2021 Published: 21 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

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

#### **1. Introduction**

Breast cancer patients that exhibit high tumor-to-breast volume ratio, lymph nodepositive disease, and aggressive biological features (high grade, hormone receptor-negative, HER2-positive, triple negative characterization) are more often candidates for neoadjuvant chemotherapy (NAC). Although large clinical trials have shown no differences in terms of overall and disease-free survival between adjuvant and neoadjuvant systemic therapy, NAC may provide important advantages [1,2]: tumor chemosensitivity can be assessed in vivo by monitoring the response to therapy, potentially allowing for the switching of therapies in case of non-responsiveness; downstaging of tumors often allows clinicians to favor breast-conserving surgery (BCS) over mastectomy and contain excision volumes, thus improving cosmetic results; downstaging of the axilla can allow for the avoidance of lymph node dissection in selected patients, reducing surgical morbidity [3] (Figure 1). Therapeutic regimens include anthracyclines (epirubicin, 100 mg/m<sup>2</sup> ), cyclophosphamide (500 mg/m<sup>2</sup> ; triweekly for 4 cycles) and taxanes (docetaxel, 70 mg/m<sup>2</sup> ; triweekly for 4 cycles); or carboplatin (100 mg/m<sup>2</sup> ; weekly for 12 cycles); taxanes are combined with targeted trastuzumab therapy in case of HER2-positivity.

**Figure 1.** Decision-making process for neoadjuvant chemotherapy [4].

Specific evidence-based guidelines have been released to ensure that each patient treated in the neoadjuvant setting may receive the most effective, evidence-based chemotherapy regimen, in a personalized, multidisciplinary setting (Figure 2).

**Figure 2.** Evidence-based medicine in neoadjuvant chemotherapy.

specific "unmet needs" that pa- Less attention has been devoted to addressing the specific "unmet needs" that patients may experience during treatment [5]. The benefits of a multimodal prehabilitation model are still emerging in recent studies, particularly during the preoperative period. During this window of opportunity, patients may be more receptive to health behavior changes in a structured support network [6].

In this paper, we present the details of an advanced, personalized, multidisciplinary prehabilitation protocol, which we have adopted in our Multidisciplinary Breast Center at Fondazione Policlinico Universitario Agostino Gemelli (FPG) in Rome since 1 May 2018 for patients scheduled to receive NAC.

This protocol allows patients to access not only the most appropriate, evidence-based chemotherapy regimen, but also specific interventions aimed at protecting their quality of life via the inclusion of lifestyle and nutrition counselling, along with psychological distress- and integrative oncology (IO)-complementary interventions [7].

#### **2. Materials and Methods**

Our breast unit treats approximately 1000 new breast cancer cases every year. Between 1 May 2018 and 31 December 2020, 250 patients were referred to our center for neoadjuvant treatment. The mean patient age was 53 (range 25–74), and 130 patients were premenopausal.

– Our broad-based interdisciplinary team includes ten breast surgeons, two medical oncologists, two breast pathologists, five breast radiologists, three breast radiologic technicians, three psycho-oncologists, two nutritionists, two integrative oncology physicians, six certified breast care nurses, and one data manager, all exclusively devoted to the management of patients with breast disease. Other team members that devote at least 50% of their activity to breast pathology include three plastic surgeons, three additional medical oncologists, two radiation oncologists, two oncogeriatricians, two gynecologists, one geneticist, one cardiologist, and two palliative care specialists. All specialists regularly attend weekly multidisciplinary meetings (MDMs), in which all new cases of breast cancer are discussed [8]. In this setting, patients are also evaluated for enrollment in clinical trials [9].

During MDMs, the case of every patient is discussed in detail, and an individualized treatment plan is programmed in adherence to the latest practice guidelines. Out of 250 patients, 98 were scheduled for an appointment with the geneticist, 14 were referred for fertility counseling, and an appointment with a geriatrician was arranged for 34 elderly patients.

#### **3. Breast Unit and Outpatient Neoadjuvant Care Prehabilitation Clinic**

All patients receiving an indication to NAC are referred to the "outpatient neoadjuvant care prehabilitation clinic", where they are jointly taken care of by a "neoadjuvant oncologic treatment team" and a "neoadjuvant supportive care team". The first team explains the care plan designed by the multidisciplinary panel, and brings into focus the important aspects of their respective areas of expertise. At the end of the interview, the patients are directed to a follow-up examination by the supportive care team for a complete psychological, nutritional, and lifestyle assessment that will serve as a baseline for the upcoming treatment. As a result, the treatment is tailored to every patient in a multidisciplinary, holistic fashion [10]. When possible, every appointment is scheduled on the same day, to limit patient discomfort in returning to the hospital several times in the same week. to NAC are referred to the "outpatient neoadjuvant care prehabilitation clinic" where they are jointly taken care of by a "neoadjuvant oncologic treatment team" and a "neoadjuvant supportive care team". The first team ex-

#### **4. The Neoadjuvant Oncologic Treatment Team**

In this setting, patients are welcomed by a team of experts consisting of a breast surgeon together with the patient's referring oncologist and breast nurse (Figure 3) [11]. geon together with the patient's

This treatment team is in charge of reviewing the diagnostic workup, discussing the therapeutic plan, and explaining, scheduling, and monitoring additional interventions that may be relevant according to the age and specific medical features of each individual patient.

**Figure 3.** The neoadjuvant oncologic treatment team.

 

As a first step, the oncologic team reviews the diagnostic workup and schedules any additional appointments that may be required to complete it.

Every patient undergoing NAC in our breast center must have completed a full diagnostic panel that includes [12]:


in order to ensure a correct pre-surgical localization in case of pathologic complete response or regression to a non-palpable lesion;


**Figure 4.** Frontal (**a**) and lateral (**b**) view of pre-neoadjuvant chemotherapy (NAC) breast with tumor projection and measurement (cm).

> The team then reviews with the patient the global therapeutic plan [14]. The oncologist discusses with the patient the details of the chemotherapeutic regimen (previously defined at the MDM) and a date for the first session of NAC is set. An appointment for central venous catheter placement is also provided.

> Based on age, general conditions, family history, and pathologic features of the tumor, the following additional interventions are discussed and eventually scheduled.

#### *4.1. Cardiovascular Assessment*

As conventional chemotherapy and targeted therapies are associated with an increased risk of cardiac damage [15], each patient scheduled for NAC undergoes preliminary cardiovascular assessment.

The development of cardiotoxicity, even if asymptomatic, not only adversely affects patient cardiac prognosis, but may significantly limit the proper completion of therapeutic protocols, especially if additional anticancer treatments become necessary after recovery/relapse of the disease [16]. Cardiovascular disease is now the second leading cause of long-term morbidity and mortality among cancer survivors, and the leading cause of death among female breast cancer survivors [17].

Our protocol ensures that a cardio-oncologist evaluates the patient via electrocardiogram and echocardiography before beginning treatment, and then periodically in relation to personal risk and ongoing pharmacological treatment. An adequate preliminary stratification of cardiotoxicity risk and the early identification and treatment of subclinical

cardiac damage may help to avoid withdrawal of chemotherapy and prevent irreversible cardiovascular dysfunction.

#### *4.2. Genetic Counselling*

Because of recent media and popular culture coverage, general knowledge about breast cancer genetics has increased in recent years [18]. Genetic test results have also become increasingly relevant in selecting the most effective systemic therapy, thanks to the advent of PARP inhibitors for treatment of BRCA1/2-associated breast cancers. Genetic assessment has become equally relevant for the optimization of radiation therapy, with emerging concerns about radiation safety for the carriers of certain pathogenic mutations (e.g., TP53) [19].

In our model, indications to genetic testing are discussed for every patient during the MDM, taking into account patient age, family history, and the clinical features of the disease. If, according to current Italian and American guidelines [20,21], genetic testing is considered appropriate, an interview with the clinical geneticist is immediately scheduled [22]. The advantage of this approach is that patients who then undergo NAC have approximately six months to complete a full, well-rounded genetic evaluation before the scheduling of surgery. This allows us to tailor surgical choices based on the test results, avoiding the unnecessary double surgery that could derive from a positive test result obtained after breast-conserving surgery [23].

#### *4.3. Multiparametric Geriatric Assessment in Elderly Patients*

Elderly patients represent a very heterogeneous community in terms of life expectancy, comorbidities, and cognitive and social function, therefore it is crucial not to deny treatment based on age alone. In this framework, a multiparametric geriatric assessment is always appropriate, and is a convenient supplement in the evaluation of every elderly patient treated for breast cancer, as it can move the needle on proposed treatment.

A recent study by Okonji et al. reported that nearly 50% of fit elderly women with high-risk disease are undertreated [24]. The neoadjuvant use of chemotherapy is further neglected, with studies reporting higher toxicity rates and lower incidence of complete pathological response in patients aged over 65 [25]. However, although elderly patients are generally underrepresented in clinical trials, those with non-triple negative breast cancer show a prognosis comparable to younger patients in terms of overall survival [26]. An individualized care model must therefore be applied to select the elderly sub-population that could benefit from neoadjuvant chemotherapy, and monitor it closely during treatment to prevent toxicity in these fragile patients [27].

In our protocol, patients aged 70 years or older, whether or not they exhibit relevant comorbidities, are scheduled for a pre-treatment comprehensive geriatric assessment. The assessment is performed by a dedicated geriatrician with experience in breast oncology, who actively participates in our multidisciplinary team. Comorbidities, cognitive and psychological disorders, physical performance, risk factors, nutritional status, and general autonomy are comprehensively evaluated, and NAC is scheduled only in the event of oncogeriatric clearance. A second assessment is also scheduled at the end of chemotherapy.

#### *4.4. Gynecologic and Fertility Counselling in Younger Patients*

Chemotherapy and/or ovarian suppression can cause early (permanent or temporary) menopausal symptoms and reduce fertility. Many women are concerned about these issues, and it is important to provide them with proper counseling and treatment [28]. As regards menopausal symptoms, a large number of patients find these difficult to cope with, with a significant negative impact on their quality of life. Our gynecologists manage these symptoms using both traditional medicine and integrative care.

Fertility care should follow a multidisciplinary team-based approach, with strict interaction between medical oncologists, surgeons, and fertility specialists [29,30]. In our multidisciplinary prehabilitation care model, the main goal is to preserve the opportunity for family planning, offering oncofertility services in a timely manner without delaying chemotherapy.

The breast nurse follows the patient on this pathway and during the subsequent procedures for ovarian function and/or fertility preservation.

#### **5. The Neoadjuvant Supportive Care Team**

After completing the assessment with the "oncological treatment team", every patient is directed to a meeting with the "neoadjuvant supportive care team", which includes a nutritionist, a psycho-oncologist, and an integrative oncology expert (Figure 5) [31]. After completing the assessment with the "oncological treatment team", every patient is directed to a meeting with the "neoadjuvant supportive care team",

**Figure 5.** The neoadjuvant supportive care team.

– A lifestyle interview is conducted, and anthropometric parameters and body composition analysis are measured via segmental multi frequency–bioelectrical impedance analysis (SMF-BIA).

– Nutritional and physical activity screenings are performed in our unit just before the beginning and at the conclusion of oncologic treatments, and the impact of each type of intervention, from surgery to chemotherapy, on BMI, body composition, and metabolism is monitored during therapy. In this regard, patients are asked to keep a diary and send it regularly via email, and periodic video interviews are scheduled [32–34].

#### *5.1. Lifestyle and Nutrition Counseling*

Physical activity (PA), nutrition, body weight, and metabolism all play a key role in almost every aspect of cancer onset, progression, and management [35] (WCRF -World Cancer Research Fund 2018). However, nutritional screening is seldom performed even in high-quality breast units, and data on its value are still scarce [36–38].

– Specific recommendations about diet and physical activity based on the most recent scientific evidence [35] are given to all patients, with the aim of relieving chemotherapy toxicity and improving quality of life and oncological outcomes [39,40]. Moreover, during and after treatments, patients are supported by a personalized nutritional approach and motivated to practice PA in order to decrease their disease recurrence risk [39,41]. PA during cancer treatments represents a powerful asset to improve therapy-induced conditions such as anxiety, depression, sleep disorders, lymphedema, cancer- and therapy-related fatigue, bone health, and overall quality of life [42–46]. –

#### *5.2. Psychological Counselling*

Chemotherapy generates a distress that, over time, can severely affect patient quality of life [47,48]. A recent study showed that post-NAC patients have a significantly higher level of distress compared to patients receiving chemotherapy after surgery [49]. Understanding the needs of patients undergoing NAC enables us to address the communication process more appropriately, provide psychological support, and build clinical and rehabilitation interventions in a more personalized way [47]. In line with NCCN guidelines, the diagnostic and therapeutic pathways of patients scheduled for NAC include a pre– post treatment psychological evaluation. Specific, psycho-oncological support should be given to patients undergoing chemotherapy. At the beginning of NAC, patients undergo a clinical psychological interview aimed at assessing their risk of oncological distress, and identifying both the dysfunctional psychological factors and the protective psychosocial factors that could affect treatment. The goal is to improve adaptation to the oncological disease and promote adherence to therapeutic treatments. In addition to the interview, a psychometric assessment is carried out through screening and the employment of clinical tools such as the Distress Thermometer (DT) [50], the Hospital Anxiety Depression Scale (HADS) [51] and the General Self-Efficacy Scale (GSES) [52]. –

In our breast center, we aim to validate a semi-structural interview, which leads to a holistic and trans-disciplinary measurement of the psychological state of the patients. The assessment allows us to identify patients who may benefit from a psychological support intervention, individual psychotherapy, or group therapy [53] (Figure 6).

**Figure 6.** Psychological interview.

#### Emotional Eating Prevention during NAC

The impact of psychosocial factors such as worry, perceived risk, and perceived treatment efficacy on diet has been understudied in breast cancer patients [54]. The relationship between distress, weight change, and nutrition has been the subject of a fairly recent psycho-oncological study trend, with major studies conducted on patients at the end of their therapies. Our model proposes an integrative approach to identify emotional eating, a dietary pattern wherein people use food to help them deal with stressful situations and in response to negative emotions. Overweight individuals have been found to exhibit less effective coping skills in response to negative emotions, which leads them to emotionally eat more frequently [55]. Psychological disciplines can help to identify healthy and harmful habits, and promote changes in attitudes and healthy behaviors.

Psychological intervention based on the activation of self-efficacy in dietary behavior could favor the ability to adapt to oncological therapies through active participation in treatment, redefinition of problems, and the evaluation of alternative solutions. At the same time, the intervention acts in support of lifestyle changes and involves the activation of specific psychoeducational groups for patients who need to change their dietary behavior.

#### *5.3. Integrative Oncology during Neoadjuvant Therapy*

Our patients receive information about evidence-based complementary therapies available, in order to optimize the management of symptoms related either to the disease itself or to treatment toxicity: most frequently gastro-intestinal disorders, hot flashes, fatigue, insomnia, mucositis, peripheral neuropathy, anxiety, and mood disorders.

In accordance with the SIO (Society of Integrative Oncology) clinical guidelines for breast cancer patients [56] recently endorsed by the American Society of Clinical Oncology (ASCO) [7], personalized integrative care plans at the FPG Center for Integrative Oncology include mind–body interventions such as acupuncture, mindfulness-based protocols, qi gong, massage therapy, and other group programs like music therapy, art therapy, and therapeutic writing workshops.

#### 5.3.1. Acupuncture

Acupuncture, well known as a branch of traditional Chinese medicine, represents a reliable, cost-effective, and safe procedure for symptom management, if performed properly and by a specialized practitioner. The NCCN recommends acupuncture for pain, fatigue, nausea/vomiting, and hot flashes [57]. For some of these symptoms, such as nausea/vomiting, acupuncture can be used as a valid option for patients who wish to avoid pharmacological treatment. For other conditions, including fatigue, hot flashes, and chemotherapy-induced peripheral neuropathy (CIPN), acupuncture should be considered when conventional treatments are ineffective, not available, or burdened by remarkable side effects.

In many patients experiencing hot flashes due to chemotherapy-induced ovarian failure and/or estrogen-blocking treatments, a course of 6–12 acupuncture treatments is associated with therapeutic effects that persist for six months or longer and do not appear to require prolonged treatment [58,59].

Chemotherapy-induced peripheral neuropathy is a challenging pain symptom to manage, and has been a hot topic for acupuncturists for a long time. A recent Cochrane review concluded that there was insufficient evidence to support or reject the use of acupuncture for neuropathic pain [60]. To date, some randomized phase II studies on the effects of acupuncture are very promising [61–63]. Recently published ESMO (European Society of Medical Oncology) guidelines on therapy-induced neurotoxicity [64] state that "Acupuncture might be considered in selected patients to treat CIPN symptoms "(grade II, C). Currently, our Center for Integrative Oncology is taking part in a multicenter clinical trial on acupuncture for chemotherapy-induced peripheral neuropathy (CIPN) in breast cancer.

#### 5.3.2. Mindfulness

Mindfulness is defined as present-moment nonjudgmental awareness, and its practice can take the form of formal meditation, or more informal practices, such as simply remembering to be present as one undertakes day-to-day tasks. Mindfulness-based stress reduction (MBSR) has been shown to reduce distress and improve psychological well-being in patients with cancer [65–67]. Preliminary evidence suggests that MBSR may produce effects comparable to pharmacologic treatment for primary insomnia [68] and positively impact sleep quality and quantity in patients with cancer [69–71]. Randomized trials of mindfulness-based stress reduction report decreased fatigue, depression, anxiety, and fear of recurrence [72,73].

In addition, improvements have been noted in sleep [72,74], quality of life, and psychosocial adjustments [70], as well as in the long-term adverse effects associated with treatment [65].

#### **6. Conclusions**

Specific clinical complications and psychological issues due to the disease and therapies can occur during the course of neoadjuvant therapy, undermining outcomes.

We have therefore developed a multidisciplinary model to ease the application of evidence-based oncologic protocols, ensure patient-centered optimal treatment, prevent distress, and improve patient quality of life. Our model of intervention can encourage clinicians to personalize supportive care medicine and direct it towards precision medicine. The development of an appropriate clinical pathway, with multidisciplinary competence and the performance of standardized tasks, is essential in order to obtain a successful treatment and make the patient co-responsible for optimized results within the neoadjuvant setting. However, multidisciplinary prehabilitation trials in breast cancer patients undertaking NAC are necessary to confirm the efficacy of this model.

**Author Contributions:** Conceptualization, A.D.L., S.M. and D.T.; methodology, A.M.S. and A.P.; software, A.D.M. and S.C.; validation, F.M., V.M. and A.O.; formal analysis, C.R.; investigation, C.M.; resources, L.S.; data curation, E.J.M.; writing—original draft preparation, A.D.L.; writing review and editing, G.F. and I.P.; visualization, A.F.; supervision, R.M.; project administration, S.D. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Fondazione Universitaria Policlinico Agostino Gemelli IRCCS.

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** In this section, you can acknowledge any support given which is not covered by the author contribution or funding sections. This may include administrative and technical support, or donations in kind (e.g., materials used for experiments).

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

#### **References**


### *Article* **Conventional CT versus Dedicated CT Angiography in DIEP Flap Planning: A Feasibility Study**

**Anna D'Angelo 1, \* ,† , Alessandro Cina 1,† , Giulia Macrì 2 , Paolo Belli 1 , Sara Mercogliano 1 , Pierluigi Barbieri 1 , Cristina Grippo 3 , Gianluca Franceschini 4 , Sabatino D'Archi 4 , Elena Jane Mason 4 , Giuseppe Visconti 2 , Liliana Barone Adesi 2 , Marzia Salgarello <sup>2</sup> and Riccardo Manfredi 1**


**Abstract:** The deep inferior epigastric perforator (DIEP) flap is used with increasing frequency in post-mastectomy breast reconstruction. Preoperative mapping with CT angiography (CTa) is crucial in reducing surgical complications and optimizing surgical techniques. Our study's goal was to investigate the accuracy of conventional CT (cCT), performed during disease staging, compared to CTa in preoperative DIEP flap planning. In this retrospective, single-center study, we enrolled patients scheduled for mastectomy and DIEP flap breast reconstruction, subjected to cCT within 24 months after CTa. We included 35 patients in the study. cCT accuracy was 95% (CI 0.80–0.98) in assessing the three largest perforators, 100% (CI 0.89–100) in assessing the dominant perforator, 93% (CI 0.71–0.94) in assessing the perforator intramuscular course, and 90.6% (CI 0.79–0.98) in assessing superficial venous communications. Superficial inferior epigastric artery (SIEA) caliber was recognized in 90% of cases (CI 0.84–0.99), with an excellent assessment of superficial inferior epigastric vein (SIEV) integrity (96% of cases, CI 0.84–0.99), and a lower accuracy in the evaluation of deep inferior epigastric artery (DIEA) branching type (85% of cases, CI 0.69–0.93). The mean X-ray dose spared would have been 788 ± 255 mGy/cm. Our study shows that cCT is as accurate as CTa in DIEP flap surgery planning.

**Keywords:** breast cancer; conventional CT and CT angiography; DIEP flap planning

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

#### **1. Introduction**

The deep inferior epigastric perforator (DIEP) flap is, nowadays, considered the "gold standard" in autologous breast reconstruction [1]. Subcutaneous tissue and skin are transferred from the abdomen to the thorax in order to guarantee a more natural appearance of the reconstructed breast, compared to heterologous approach [2,3] (Figures 1 and 2). A low donor site morbidity with an aesthetical abdomen improvement is an important factor for choosing DIEP flap in autologous breast reconstruction. The inconsistent anatomy of the abdominal perforators leads to a more challenging and time-consuming technique compared to a (muscle sparing) Transverse Rectus Abdominis Muscle (TRAM) flap [4,5].

**Citation:** D'Angelo, A.; Cina, A.; Macrì, G.; Belli, P.; Mercogliano, S.; Barbieri, P.; Grippo, C.; Franceschini, G.; D'Archi, S.; Mason, E.J.; et al. Conventional CT versus Dedicated CT Angiography in DIEP Flap Planning: A Feasibility Study. *J. Pers. Med.* **2021**, *11*, 277. https://doi.org/ 10.3390/jpm11040277

Academic Editor: Hisham Fansa

Received: 22 February 2021 Accepted: 31 March 2021 Published: 7 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

111

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 2 of 9

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 2 of 9

Muscle (TRAM) flap [4,5].

Muscle (TRAM) flap [4,5].

inconsistent anatomy of the abdominal perforators leads to a more challenging and timeconsuming technique compared to a (muscle sparing) Transverse Rectus Abdominis

inconsistent anatomy of the abdominal perforators leads to a more challenging and timeconsuming technique compared to a (muscle sparing) Transverse Rectus Abdominis

**Figure 2.** Preoperative planning (**a**) of a DIEP flap reconstruction for right breast carcinoma, requiring nipple-sparing mastectomy. Eight-month postoperative result (**b**). **Figure 2.** Preoperative planning (**a**) of a DIEP flap reconstruction for right breast carcinoma, requiring nipple-sparing mastectomy. Eight-month postoperative result (**b**). **Figure 2.** Preoperative planning (**a**) of a DIEP flap reconstruction for right breast carcinoma, requiring nipple-sparing mastectomy. Eight-month postoperative result (**b**).

Preoperative planning is crucial [6] in order to identify perforator vessels originating from the deep inferior epigastric vascular system, and to evaluate superficial inferior epigastric vessels. DIEP flap survival depends on adequate blood supply, which is guaranteed by perforator vessels that are amply variable in terms of number, anatomical location, intramuscular course, caliber, and tortuosity. Preoperative assessment includes visualization of the deep inferior epigastric artery (DIEA) and evaluation of its intramuscular course and

branching pattern. The latter is described by Taylor's classification, which defines three types of DIEA branching above the arcuate line: in type I the artery ascends as a single intramuscular vessel; in type II, the artery divides, at the arcuate line, into two vessels with an intramuscular course; in type III, the artery divides, at the arcuate line, into three vessels with an intramuscular course [7].

The DIEA originates from the external iliac artery, above the inguinal ligament, and crosses the lateral margin of the rectus abdominis muscle 3–4 cm below the arcuate line, with an average pedicle length of 10.3 cm and an average vessel diameter of 3.6 mm [8]. It then normally divides into two branches, lateral and medial; in case of a central course (28%), multiple small branches with centrally located perforators can be detected [9].

Perforators arise on each side of the midline from the anterior rectus fascia in a central rectangular area, which extends craniocaudally from 2 cm above to 6 cm below the umbilicus, and laterally between 1 cm and 6 cm from the midline. A thorough preoperative anatomical study also allows an assessment of the communications between the superficial and deep systems. The caliber of the superficial inferior epigastric artery (SIEA) should be compared to that of the dominant perforator, in order to select the best pedicle for the flap. In addition, assessing the integrity of the superficial inferior epigastric veins (SIEVs) could be helpful, in case of a flap additional venous discharge requirement [8].

Different perforator locations are associated with a harder or easier dissection, and sometimes lead to extensive splitting of the muscle; compared to lateral vessels, medial perforators offer better flap perfusion but a harder dissection due to a long intramuscular course. Perforator dissection is carried out along the deep inferior epigastric pedicle up to its origin from the external iliac artery. The DIEP flap should be adapted and shaped to the single patient and type of breast reconstruction, with an optimized anatomical preoperative study that allows the identification of personal anatomical characteristics in order to accelerate dissection and flap harvesting, as well as to avoid vascularization deficiencies. An accurate preoperative planning with evaluation of single anatomical variants allows a decrease in decrease operating time and theatre utilization, with a consequent benefit in terms of surgical waiting lists and staff optimization.

Among the available imaging techniques, which include Magnetic Resonance Imaging (MRI) and color-Doppler ultrasound (US) [3,10], Computed Tomography Angiography (CTa), with the injection of contrast medium, has become the gold standard in planning surgery [11,12] thanks to its ability to map out the vascular anatomy and, consequently, select the best DIEP flap to harvest. Its high accuracy has been proved in studies performed, both on cadavers [8] and post-surgery. CTa also allows 3D surface and vascular treerendering [9], which can bring huge benefits to cross-sectional imaging and represents a valid visual tool for surgeons. The primary role of CTa in preoperative assessment is, therefore, motivated by its wide availability, fast acquisition time, high reproducibility, and great sensitivity in the identification of perforator vessels with calibers larger than 1mm. Still, CTa is associated with possible complications, such as allergic reactions to contrast medium, nephrotoxicity in patients with impaired renal function, and exposure to ionizing radiation in patients often already subjected to multiple CT scans to stage primary breast cancer [13].

Our goal was to investigate the accuracy of conventional CT (cCT), performed during breast cancer disease staging, compared to CTa in obtaining information required for DIEP flap surgical planning. We evaluated the accuracy of both techniques in identifying "dominant" perforator arteries, measuring their caliber and intramuscular course length, assessing superficial venous communications (SVC) and DIEA branching type according to Taylor's classification, identifying the caliber of SIEA, and assessing SIEV integrity. In addition, the total X-ray dose that could have been potentially spared by avoiding CTa was evaluated.

#### **2. Materials and Methods**

From January 2010 to February 2019, 344 patients programmed to receive mastectomies with immediate or delayed DIEP flap reconstruction, referred to our Institute, were enrolled in the study. Inclusion criteria were: cCT performed during disease staging with standardized technique (slice thickness of 1.25 mm in the portal venous-phase) in our Institution within 24 months after CTa. Exclusion criteria were: abdominal surgery between the two examinations or cCT performed in other Institutions.

This retrospective single-center study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board and Ethics Committee of Fondazione Policlinico Universitario Agostino Gemelli IRCCS on 11 June 2020. Anyone involved in the research agreed to participate and agreed to have the results of the research about them published.

#### *2.1. CTa and cCT Technique*

CTa and cCT were performed using a 64-slice multidetector CT (LightSpeed VCT, GE Healthcare, Waukeska, WI, USA), table travel per rotation was 23 mm (gantry rotation time 0.4 s) and field of view (FOV) was 40 cm in order to match patient width, matrix side 512 × 512. Tube voltage was 120 kVp, with Smart mAs (GE Healthcare) dose enabled (noise index set to 22). For CTa, the arterial-phase images were acquired at a 0.65 mm slice thickness; to minimize radiation exposure, a small field of view (FOV), which only includes the area of interest, is scanned: from the origin of the inferior epigastric artery at the level of the groin to a level approximately 3 cm above the umbilicus in a caudal-cranial direction. We administered, intravenously, 100 mL of iodinated contrast medium (Ultravist, Bayer Schering Pharma AG, Berlin, Germany) with a concentration of 370 mgL/mL (18-G cannula) at 4 mL/s flow rate, followed by 60 mL saline flush. A large-gauge (18 G) peripheral intravenous line was preferred to allow rapid infusion of contrast (4–5 mL/s) and, thus, an optimal opacification of small epigastric vessels. The arterial peak of enhancement was captured using bolus tracking (Smart Prep, GE Healthcare, Wuakuesha, WI, USA), so as to begin image acquisition upon the contrast medium arrival in the region of interest (ROI) on the common femoral artery; acquisition should be obtained with a minimum possible delay after contrast arrival is detected, with blood attenuation within ROI of 100–120 Hounsfield units (HU). During the exam, since a whole scan can be accomplished in one held breath, and the effect of breathing motion on the abdomen and pelvis may be relevant, patients are required to hold their breath and are supine, with their arms placed according to the programmed sugery (upwards for immediate breast reconstruction, downwards in case of delayed reconstruction).

For cCT, the venous-phase images were acquired at a 1.25 mm slice thickness, in cranio-caudal direction, with patients in a supine position with their arms lying upwards. Following our department's routine for oncologic staging cCT, 1.6 mL/kg of contrast medium (Ultravist 370 mgL/mL) was administered to patients at a rate of 3 mL/s, followed by 40 mL of saline solution at the same injection rate. The scan delay was empirically chosen at 70 s.

#### *2.2. Image Analysis*

Two radiologists with specific experience in flap surgery imaging reviewed, respectively, the cCT and aCT exams to assess the diagnostic accuracy of cCT in identifying: the main perforators, the "dominant" perforator, and the perforation site of the rectus abdominal fascia using volumetric reconstructions. The errors on x and y virtual coordinates were then calculated (Figure 3). The reader also evaluated the course of the dominant perforator, assigning a value from 1 to 4 ("1" extramuscular, "2" intramuscular for a length <2 cm, "3" <4 cm and "4" >4 cm); the branching of the DIEA according to Taylor's classification (Figure 4); the caliber of the SIEA compared to the dominant perforator (from 1 to 3, "1" <dominant, "2" =dominant, "3" >dominant) (Figure 5); the integrity of the SIEV (from 1 to 3, "1" intact, "2" attracted, "3" interrupted); and the presence of superficial venous

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 5 of 9

Taylor's classification (Figure 4); the caliber of the SIEA compared to the dominant perforator (from 1 to 3, "1" <dominant, "2" =dominant, "3" >dominant) (Figure 5); the

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 5 of 9

("0" if absent, "1" scarce, "2" moderate, "3" clearly evident) (Figure 6).

communications between the right and left hemi-abdomen ("0" if absent, "1" scarce, "2" moderate, "3" clearly evident) (Figure 6). integrity of the SIEV (from 1 to 3, "1" intact, "2" attracted, "3" interrupted); and the presence of superficial venous communications between the right and left hemi-abdomen ("0" if absent, "1" scarce, "2" moderate, "3" clearly evident) (Figure 6). integrity of the SIEV (from 1 to 3, "1" intact, "2" attracted, "3" interrupted); and the presence of superficial venous communications between the right and left hemi-abdomen

dominant perforator, assigning a value from 1 to 4 ("1" extramuscular, "2" intramuscular for a length <2 cm, "3" <4 cm and "4" >4 cm); the branching of the DIEA according to Taylor's classification (Figure 4); the caliber of the SIEA compared to the dominant perforator (from 1 to 3, "1" <dominant, "2" =dominant, "3" >dominant) (Figure 5); the

**Figure 3.** Dominant perforator's emergence from the anterior rectus abdominis fascia (red arrows) in cCT (**a**) and CTa (**b**) axial sub-volume maximum intensity projection (MIP) reconstructions. Images (**c**,**d**) show mapping of the dominant perforator on a VR reconstruction of the abdominal surface via a virtual coordinate system centered on a zero point, corresponding to the umbilicus in cCT (**c**) and CTa (**d**). **Figure 3.** Dominant perforator's emergence from the anterior rectus abdominis fascia (red arrows) incCT (**a**) and CTa (**b**) axial sub-volume maximum intensity projection (MIP) reconstructions. Images (**c**,**d**) show mapping of the dominant perforator on a VR reconstruction of the abdominal surface via a virtual coordinate system centered on a zero point, corresponding to the umbilicus in cCT (**c**) and CTa (**d**). **Figure 3.** Dominant perforator's emergence from the anterior rectus abdominis fascia (red arrows) in cCT (**a**) and CTa (**b**) axial sub-volume maximum intensity projection (MIP) reconstructions. Images (**c**,**d**) show mapping of the dominant perforator on a VR reconstruction of the abdominal surface via a virtual coordinate system centered on a zero point, corresponding to the umbilicus in cCT (**c**) and CTa (**d**).

**Figure 4.** Identification of the deep inferior epigastric artery branching according to Taylor's classification. cCT (**a**) and CTa (**b**) oblique-coronal sub-volume maximum intensity projection reconstruction (MIP) of the superficial abdominal wall revealed a bifurcated artery on the right hemi-abdomen (red arrows) and a single on the left (white arrows). **Figure 4.** Identification of the deep inferior epigastric artery branching according to Taylor's classification. cCT (**a**) and CTa (**b**) oblique-coronal sub-volume maximum intensity projection reconstruction (MIP) of the superficial abdominal wall revealed a bifurcated artery on the right **Figure 4.** Identification of the deep inferior epigastric artery branching according to Taylor's classification. cCT (**a**) and CTa (**b**) oblique-coronal sub-volume maximum intensity projection reconstruction (MIP) of the superficial abdominal wall revealed a bifurcated artery on the right hemi-abdomen (red arrows) and a single on the left (white arrows).

hemi-abdomen (red arrows) and a single on the left (white arrows).

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 6 of 9

**Figure 5.** Assessment of the SIEA caliber compared to the dominant perforator. cCT (**a**) and CTa (**b**) sub-volume sagittal MIP reconstructions show a SIEA (red arrows) with a 2 score (equal to the dominant perforator). **Figure 5.** Assessment of the SIEA caliber compared to the dominant perforator. cCT (**a**) and CTa (**b**) sub-volume sagittal MIP reconstructions show a SIEA (red arrows) with a 2 score (equal to the dominant perforator). **Figure 5.** Assessment of the SIEA caliber compared to the dominant perforator. cCT (**a**) and CTa (**b**) sub-volume sagittal MIP reconstructions show a SIEA (red arrows) with a 2 score (equal to the dominant perforator).

**Figure 6.** Assessment of superficial venous communications running between the right and left portion of the abdomen. Coronal sub-volume maximum intensity projection (MIP) reconstructions of the superficial abdominal wall for cCT (**a**) and CTa (**b**) show a large venous trunk on the right hemi-abdomen (red arrows), with a 3 score. Superficial inferior epigastric vein integrity was absent on the left (red circles). **Figure 6.** Assessment of superficial venous communications running between the right and left portion of the abdomen. Coronal sub-volume maximum intensity projection (MIP) reconstructions of the superficial abdominal wall for cCT (**a**) and CTa (**b**) show a large venous trunk on the right hemi-abdomen (red arrows), with a 3 score. Superficial inferior epigastric vein integrity was absent on the left (red circles). **Figure 6.** Assessment of superficial venous communications running between the right and left portion of the abdomen. Coronal sub-volume maximum intensity projection (MIP) reconstructions of the superficial abdominal wall for cCT (**a**) and CTa (**b**) show a large venous trunk on the right hemi-abdomen (red arrows), with a 3 score. Superficial inferior epigastric vein integrity was absent on the left (red circles).

#### *2.3. Statistical Analysis 2.3. Statistical Analysis 2.3. Statistical Analysis*

needed.

needed.

**3. Results** 

**3. Results** 

A Shapiro–Wilk test was employed to assess the parametric vs nonparametric distribution of variables. Continuous variables were described by mean and standard deviation. The accuracy of cCT was tested, with CTa employed as a standard of reference. Confidence intervals were reported at 95%. For inferential statistics, a Student *t*-test and Wilcoxon rank–sum test were employed, respectively, for parametric and nonparametric variables. Setting a type II error (1 − β) of 0.9 and a Type I error rate of 0.05, and assuming as clinically relevant a 0.9 accuracy of the cCT vs. CTa, a sample size of 35 patients was A Shapiro–Wilk test was employed to assess the parametric vs nonparametric distribution of variables. Continuous variables were described by mean and standard deviation. The accuracy of cCT was tested, with CTa employed as a standard of reference. Confidence intervals were reported at 95%. For inferential statistics, a Student *t*-test and Wilcoxon rank–sum test were employed, respectively, for parametric and nonparametric variables. Setting a type II error (1 − β) of 0.9 and a Type I error rate of 0.05, and assuming as clinically relevant a 0.9 accuracy of the cCT vs. CTa, a sample size of 35 patients was A Shapiro–Wilk test was employed to assess the parametric vs nonparametric distribution of variables. Continuous variables were described by mean and standard deviation. The accuracy of cCT was tested, with CTa employed as a standard of reference. Confidence intervals were reported at 95%. For inferential statistics, a Student *t*-test and Wilcoxon rank–sum test were employed, respectively, for parametric and nonparametric variables. Setting a type II error (1 − β) of 0.9 and a Type I error rate of 0.05, and assuming as clinically relevant a 0.9 accuracy of the cCT vs. CTa, a sample size of 35 patients was needed.

#### **3. Results**

We enrolled 35 patients with a mean age of 40 years (range 27–73 years) and a mean BMI of 25,2 kg/m<sup>2</sup> (range 21.2–32.3). No statistically significant differences were observed in patient characteristics. The accuracy of cCT in assessing the three largest perforators was 95% (CI 0.80–0.98). The dominant perforator was identified by cCT in all cases (100%, CI 0.89–100). cCT correctly identified the perforator intramuscular course in 93% of cases (CI 0.71–0.94) and the superficial venous communications in 90.6% of patients (CI 0.79–0.98). The SIEA caliber was correctly assessed by cCT in 90% of cases (CI 0.84–0.99). cCT was less accurate in the evaluation of DIEA branching type (85% of cases, CI 0.69–0.93), but had an excellent assessment of the integrity of SIEV (96% of cases, CI 0.84–0.99). The mean error in topographic localization was 4.8 ± 3.8 mm along the Y axis and 2.6 ± 3.8 mm along the X axis. If CTa had been spared before surgery, relying on cCT for DIEP planning, the mean X-ray dose potentially avoided would have been 788 ± 255 mGy/cm. Data reported are shown in Table 1.

**Table 1.** Performance of cCT versus CTa.


#### **4. Discussion**

The results of our study show that cCT, performed routinely during breast cancer disease staging, is as accurate as CTa in obtaining information required for DIEP flap planning. CTa, first described by Rozen in 2008 [8], has been suggested as the gold standard in preoperative assessment of perforating vessels. Other modalities, such as MRI [14] and color-Doppler US, have been compared to CTa. Preoperative breast MRI performed for breast malignancy characterization can be extended to the lower abdomen, but still allows visualization and localization of only some of the perforator vessels, as it possesses a lower spatial resolution compared to CT angiography [15]. The prone position required to perform breast MRI modifies the natural anatomy of the abdomen, and, together with artefacts due to respiratory movements and enhanced vascular assessment, constitutes a limitation to using MRI, as reported in our previous study [11]. Color-Doppler US, though it offers a more accurate spatial resolution than CTa, is an operator dependent procedure and requires advanced training to obtain a satisfying mapping of perforators [16]. To our knowledge, no previous studies investigated the role of CTa versus cCT. Our results show an excellent diagnostic accuracy of cCT in identifying the three largest perforators, the perforator intramuscular course, SCVs, the dominant perforator, SIEA caliber, and SIEV integrity. cCT was less accurate in the evaluation of DIEA branching type, probably because of lower contrast resolution during the venous phase, different contrast medium injection speed, and the cranial-caudal direction of acquisition. The mean error in topographic localization of the dominant perforator was 4.8 ± 3.8 mm along the Y axis and 2.6 ± 3.8 mm along the X axis, probably because of the different arm position in delayed surgical reconstruction and the presence of clothes (knickers) when the cCT is performed. Results from both techniques were compared with intraoperative findings: all preoperatively assessed dominant perforators were confirmed intraoperatively, without significant differences in terms of expected position.

Our study suggests that performing cCT alone, in the preoperative assessment of DIEP-flap candidates, is safe and feasible. Furthermore, everyday clinical practice could benefit from the adoption of this technique in several ways: preoperative assessment is

faster without the necessity of programming a CTa exam; there is also the matter of reduced healthcare costs and patient discomfort, both in terms of psychological stress and x-ray or contrast medium exposure. This evidence is particularly significant when dealing with patients who are already exposed, because of their underlying disease, to multiple CT examinations. If CTa had been withheld before surgery, relying on cCT alone for DIEP planning, the patient would have been spared a mean X-ray dose of 788 ± 255 mGy/cm. Furthermore, this technique is easily applicable to most centers around the world, including facilities without access to CTa, as it does not require a dedicated acquisition protocol or a radiologist specialized in vascular anatomy.

Our study has some limitations, the major of which being that we could not assess interobserver variability between CTa and cCT because only one experienced radiologist was present for each method. Furthermore, DIEP flap procedure total surgical time was not taken into account in this study, although it was widely analyzed in our previous manuscript [3].

#### **5. Conclusions**

We found that cCT, although not intentionally performed for preoperative surgical assessment, nonetheless provided an accurate visualization of the best perforator and of the main abdominal vessels involved in DIEP planning, thus, overcoming the limits of US in terms of reproducibility and operator dependence, and of MRI in terms of spatial resolution, costs, and artifacts related to the prone position. In this way, patients scheduled for DIEP flap surgery with a recent cCT could avoid further assessment with CTa. In conclusion, in order to strongly reduce radiation exposure, time, and costs in DIEP flap planning, a previous recent cCT may be a valuable option due to high concordance with CTa findings.

**Author Contributions:** Conceptualization, A.C., A.D. and G.M.; methodology, P.B. (Paolo Belli); software, S.D.; validation, R.M., M.S. and G.F.; formal analysis, E.J.M. and C.G.; investigation, S.M.; resources, P.B. (Pierluigi Barbieri); data curation, L.B.A. and G.V.; writing—original draft preparation, A.D., G.M., C.G. and A.C.; writing—review and editing, E.J.M., S.D.; visualization, S.M.; supervision, A.C.; project administration, R.M. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board and Ethics Committee of Fondazione Policlinico Universitario Agostino Gemelli IRCCS; Università Cattolica del Sacro Cuore, Rome, Italy.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in the study are available on request from the corresponding author.

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

#### **References**


### *Article* **Development of a Digital Research Assistant for the Management of Patients' Enrollment in Oncology Clinical Trials within a Research Hospital**

**Alfredo Cesario 1 , Irene Simone 1 , Ida Paris 2 , Luca Boldrini 3 , Armando Orlandi 4 , Gianluca Franceschini 5 , Filippo Lococo 6 , Emilio Bria 4 , Stefano Magno 7 , Antonino Mulè 5 , Angela Santoro 5 , Andrea Damiani 3 , Daniele Bianchi 8 , Daniele Picchi 8,9 , Guido Rasi 10 , Gennaro Daniele 10,11 , Alessandra Fabi 12 , Paolo Sergi 8 , Giampaolo Tortora 4 , Riccardo Masetti 5,13 , Vincenzo Valentini 3 , Marika D'Oria 1, \* and Giovanni Scambia 2,14**


**Abstract:** Clinical trials in cancer treatment are imperative in enhancing patients' survival and quality of life outcomes. The lack of communication among professionals may produce a non-optimization of patients' accrual in clinical trials. We developed a specific platform, called "Digital Research Assistant" (DRA), to report real-time every available clinical trial and support clinician. Healthcare professionals involved in breast cancer working group agreed nine minimal fields of interest to preliminarily classify the characteristics of patients' records (including omic data, such as genomic mutations). A progressive web app (PWA) was developed to implement a cross-platform software that was scalable on several electronic devices to share the patients' records and clinical trials. A specialist is able to use and populate the platform. An AI algorithm helps in the matchmaking between patient's data and clinical trial's inclusion criteria to personalize patient enrollment. At the same time, an easy configuration allows the application of the DRA in different oncology working groups (from breast cancer to lung cancer). The DRA might represent a valid research tool supporting clinicians

**Citation:** Cesario, A.; Simone, I.; Paris, I.; Boldrini, L.; Orlandi, A.; Franceschini, G.; Lococo, F.; Bria, E.; Magno, S.; Mulè, A.; et al. Development of a Digital Research Assistant for the Management of Patients' Enrollment in Oncology Clinical Trials within a Research Hospital. *J. Pers. Med.* **2021**, *11*, 244. https://doi.org/10.3390/jpm11040244

Academic Editor: Enrico Capobianco

Received: 5 March 2021 Accepted: 24 March 2021 Published: 27 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

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

and scientists, in order to optimize the enrollment of patients in clinical trials. User Experience and Technology The acceptance of participants using the DRA is topic of a future analysis.

**Keywords:** clinical trial; patient enrollment; artificial intelligence; machine learning; breast cancer; lung cancer; oncology; web app; personalized medicine

#### **1. Introduction**

Cancer care is a complex pathway that is based on a multidisciplinary collaboration among professionals who share the latest evidence and pool their expertise and information through regular communication flows [1]. Multidisciplinary data sharing is an essential approach for tracing patients' pathways, optimizing therapeutic opportunities, and improving healthcare quality. This approach increases evidence-based practice and avoids treating patients outside standardized protocols and recommended guidelines [2,3].

Clinical trials are imperative for testing novel cancer treatments, advancing the knowledge of care, and determining the best strategies to enhance patients' survival and quality of life outcomes [4,5]. Nevertheless, the possible lack of communication and real-time synchronization among professionals may produce a fragmentation of services and practices, potentially resulting in the non-optimization of patients' accrual in clinical trials and the limitation of their access to innovative therapeutic solutions [4–6].

One possible solution can be represented by data sharing approaches, facilitating the enrollment of patients in clinical trials that allow for increasing the chances of recovery, testing novel treatments, and improving knowledge of disease. Less than 5% of the patients are currently enrolled in clinical trials due to logistical issues, a lack of resources, and difficulty in data sharing [4,7–9].

Our research hospital has a notable oncological vocation, with nearly 60,000 patients annually accessing our facility with its complex organization in clinical, surgical, and service departments that welcome and manage all of the needs of the cancer patients. Specifically, the Comprehensive Cancer Center coordinates and optimizes all of the cancer related activities, guaranteeing the functionality of specific multidisciplinary working groups and the access to innovative therapies through enrollment in clinical trials or comprehensive interpretation of big data at the institutional and network levels [9–11]. In order to reduce daily communication inconveniences [12,13], a specific platform, called "Digital Research Assistant" (DRA), was developed to report real-time every available clinical trial active within our research hospital and assist clinicians in properly matching patients with the more appropriate studies.

The aim of this paper is to show how the DRA was implemented for breast cancer clinical trials to map all of the active studies on this specific disease and encouraging proper patients' enrollment. Its scalability was also evaluated presenting the lung cancer case study.

#### **2. Materials and Methods**

#### *2.1. Ideation*

A project manager and two Information and Communication Technology (ICT) professionals started a pilot project with the Breast and Lung Cancer institutional Working Groups, following a user-centered designed approach [14] (Figure 1).

**Figure 1.** User-centered designed approach. Context: the program manager identifies who are the primary users of the product, how and why they will use it, what are their needs, and which environment they will use the tool. Requirements: when the context is defined, the program manager identifies the detailed requirements of the product, according to the needs of the user. Design solutions and development: once goals and requirements are settled, the ICT professionals and the project manager design and develop the tool for its usability. Evaluate Product: product designers (in this case, ICT professionals) run usability tests to obtain users' feedback on the product.

Healthcare professionals of the involved working groups agreed on nine minimal fields of interest to preliminarily classify the characteristics of patients' records in the platform (Table 1) and obtain a quick evaluation of the patients and its possible link to the active and open clinical trials, using breast cancer as a case study.


**Table 1.** Fields chosen by the professionals of the Breast Cancer Working Group, in order to classify the patients inserted in the platform.

> TNM classification and corresponding stage were obtained through the input of numerical values according to the 8th edition of TNM classification of malignant tumor [15];


gesterone (PgR: positive or negative) receptors, epidermal growth factor receptor 2 (HER2: positive or negative) and the proliferation index (Ki67: from 1% to 100%):


Particularly, prognosis and treatment are determined by the stage (TNM classification) of the tumor at the time of diagnosis, but also by the histological/molecular subtype that is obtained with biopsies or in the definitive pathological examination.

#### *2.2. Implementation*

The DRA was created with the aim to meet several essential clinical and research points:


The infrastructure was designed and developed by separating the front-end (i.e., the exposed services) from the back-end (information content) in order to ensure data protection and security (Figure 2).

**Figure 2.** Hardware infrastructure.

Infrastructure versatility was then tested using a different case study (lung cancer), to confirm the possibility to easily adapt the platform for indications other than the one used for the first development.

#### 2.2.1. Technologies and Software

A progressive web app (PWA) was developed to implement a cross-platform software scalable on several electronic devices (i.e., PC, tablet, smartphone). Differently from classic web apps, a PWA that is installed on mobile devices acts as if it was a native app of the device itself, allowing:


Business logic was developed using the Microsoft, NetCore 3.1 Framework. This software was structured with APIs that make access to data in secure mode with a https protocol scalable and decoupled from the front-end. The app is scalable in terms of the evolution and reutilization of the code, as well as maximization of loading information on the network.

The angular open source framework version 9.07 was used to develop the front-end, directly running from the browser after being downloaded from the web server. This choice was taken to have an advance in terms of efficiency, saving the exchange of information between client and server every time that there is a request for action by the user.

The SignalR open source framework was used to guarantee a real-time update of data, even when the app is open on a browser. This technology automatically updates information modified by other users, using a two-way channel between the client (browser) and the server (web app). In order to ensure the communication of changes in information to clients, even when they are not connected to the web app, neither it is open, push notifications have been activated using the Google Firebase engine. This open source service allows for sending messages through a web service that transmits notifications to the users of the service. Finally, Microsoft Sql Server 2016 Enterprise Edition was the DBMS used to define the relational model related to this architecture.

#### 2.2.2. Accessibility

System access is possible through a hybrid authentication architecture (Figure 3) that allows specialists and healthcare professionals located in various research centers to use the platform:


**Figure 3.** Authentication architecture.

The app admits three profiles:


• Study Manager: enabled to use the same functions of the "User" profile, as well as to manage the creation and modification of trials.

"System Administrators" manage the access of internal users (with "User" and "Study Manager" profiles) enabling them to use the app. An HR representative of the research hospital supervises the list of users.

#### **3. Results**

*3.1. Functionalities and Configuration*

From the side menu, the following functionalities are available:

	- # Users
	- # Clinical Trial (or Study)
	- # Type of Clinical Trial (or Study)
	- # Phase
	- # Settings

Other functionalities include system management and configurations, which are dedicated to "System Administrator" profiles.

3.1.1. Patients' Management and Enrollment (Matching) to Clinical Trials

Under the operational functions, it is possible to see a real-time updated patients list from the activated module to:



**Figure 4.** List of patients uploaded in the system (in Italian). Names are examples and do not correspond to real cases.

Search filters are available for a better user experience. In particular, the legend includes four entries:


Patient enrollment changes according to the logged profile. If the profile is "Study Manager", then the enrollment occurs immediately, otherwise a "User" sends a request to the "Study Manager" of the selected trial, which allows or denies access to the patient in the study. When a patient is accepted, or directly recruited, the "Study Manager" inserts the starting and ending date of the trial.

#### 3.1.2. Clinical Trial Configuration

A list of Clinical Trials with their status (i.e., active, suspended, closed) is displayed for all of the profiles (Figure 5). To configure a Clinical Trial, the "User", or the "Study Manager" can enter the information related to the study in which patients can be enrolled (Figure 6). These information are shared with other users, especially those that are interested in the same pathology.


**Figure 5.** List of Clinical Trials (in Italian). Names are examples and do not correspond to real cases.


**Figure 6.** New Clinical Trial form (in Italian).

3.1.3. Phase of the Clinical Trial Configuration

"User" and "Study Manager" profiles can input and modify the Phase of the clinical trial, visible on the selection menu while configuring a trial (Figures 7 and 8).



**Figure 8.** New Phase insert form (in Italian).

Matchmaking option. An algorithm then configures the clinical trial. By defining the inclusion-exclusion criteria of a patient in a trial enrollment, these criteria become the rules of the algorithm that allows matchmaking between an eligible patient and a trial. When the "User" inserts a new patient, it is possible to click on the action "assign the patient to a study". This action shows the list of clinical trials for which the patient is eligible. If the patient has characteristics that are coherent with the study, he/she can be enrolled. In particular, omic characteristics (such as genomic mutations) may help achieve a Personalized Medicine approach in oncological clinical trial enrollment.

As an enrichment of the services offered by the platform, a connection with the GEmelli NEtwoRk for Analysis and Tests in Oncology and medical Research "Generator"— Real World Data facility is offered to the clinician. Gemelli Generator Real-World Data is a research facility whose aim is the integration of the vast amount of patient data that are available in the Gemelli Data warehouse (about 700 million data items as measured at the end of December 2020). The generator takes care of the integration of these data items, in anonymized form, into specific datamarts, based on appropriate terminological systems, quality-checked and normalized with regard to the information originated from different, heterogeneous data sources, like traditional electronic health records (EHRs), omics data, Patient-Reported Experience Measures (PREMS), and Patient-Reported Outcome Measures (PROMS).

Machine Learning and Artificial Intelligence-based methods are at the heart of the Generator infrastructure, allowing for researchers to develop state of the art models, clustering, and decision support systems [16]. After the patient selection phase of the DRA, a simple user interface will give clinicians the opportunity to query the Generator datamarts for the availability of further covariates, referred to the selected patients, that can add more information to what is already present in the DRA core. Full integration between the two systems, at the ICT level, will guarantee an automatic and swift response in a privacy protected environment.

In this way, researchers can have a deeper view of the available data and formulate more study hypotheses, based on the large variety of information coming from heterogeneous data sources.

#### 3.1.4. Settings Configuration

"System Administrators" can insert or modify the characteristics of the setting attributes (that are chosen by the WG) related to the clinical trials (Figures 9 and 10), while "User" and "Study Manager" profiles can select them directly.

**Figure 9.** New Setting form (in Italian).


**Figure 10.** List of Settings descriptions (in Italian).

#### 3.1.5. User Requests

In the "User Requests" section, all of the requests and their status (accepted, refused, and pending for evaluation) are displayed as well as other users' information requests about a patient or a trial (Figure 11).


**Figure 11.** User Requests (in Italian). Names are examples and do not correspond to real cases.

#### 3.1.6. Requests Management

This section is only accessible to "Study Manager" profiles and allows accepting or declining a request (Figure 12). Each row shows a single request with the possibility of examining patient or trial information.


**Figure 12.** Requests Management (in Italian). Names are examples and do not correspond to real cases.

#### 3.1.7. Clinical Trials List

This section shows a list of all the clinical trials that the logged profile is responsible for. The "Study Manager" profile also allows editing information about the trials and examining the enrolled patients' full list (Figure 13).


**Figure 13.** Trials list (in Italian). Names are examples and do not correspond to real cases.

3.1.8. Possible Enrollment List

This section shows all of the patients preferred by "User" profiles, and preference can be deselected. It is possible to send or accept requests of admission in a trial (Figure 14).


**Figure 14.** Possible enrollment list (in Italian). Names are examples and do not correspond to real cases.

#### **4. Customization**

In this paper, we described the scale-up customization of the first DRA model on breast cancer to lung cancer thought for a high volume cancer care center. Table 2 shows all the varied characteristics of the patient, except for the "Age" field.

**Table 2.** Fields chosen by the professionals of the Lung Cancer Working Group, in order to classify patients inserted in the platform.


The parameters included in the "Minimum Fields" (Table 2) were selected while considering the main characteristics of lung cancer patients that can guide the accrual in clinical trials.


Prognosis and treatment are determined by disease stage (TNM classification), as identified by preliminary diagnostic investigations and histology. Surgery remains the main prognostic factor in the early-stage tumors and the completeness of resection (negative margin status) is a widely recognized factor influencing the long-term results in this setting. Otherwise, in locally advanced and metastatic stages, the molecular characterization represents the main determinant of long-term outcomes, based on the dramatic predictive role of featured biomarkers of activity/efficacy for molecular targeted agents and immunotherapy.

Table 3 shows the numerical data from a pilot test of the DRA database.

**Table 3.** Number of patients available in the Digital Research Assistant.


#### **5. Discussion and Conclusions**

Our solution exploited communication among professionals that are involved in oncological care and cancer research. The DRA we developed allows them to know all of the running clinical trials, guaranteeing all patients the best access to cure and research protocols, reducing the fragmentation of patients' access to the oncological care-path with the multiple therapeutic intersections available in high volume centers (radiotherapy, surgery, and new lines of systemic therapy managed by multiple specialists).

This digital tool appeared to be well performing for patients' data sharing within the single institution, but also in setting up networks with other cancer centers facilitating patients' enrollment also for peripheral centers. In fact, in high-patient volume centers, such as our institution, the DRA seems a possible efficient resource to face this issue [20–26]. At the same time, the platform is easily moldable to the needs of different oncology work groups, as evidenced by the easy customization starting from the model for breast cancer to arrive at that for lung cancer.

Taking these considerations together, the platform might represent a valid research tool supporting clinicians and scientists, working in both high- and low-volume centers and the enrollment success rates for each matchmaking run is currently the object of in depth analysis and it will be a topic of future publications. User Experience and Technology Acceptance of participants using the DRA is topic of a second dedicated analysis.

**Author Contributions:** Conceptualization, all authors; methodology, A.C., I.S., M.D., P.S., D.B., and D.P.; software, P.S., D.B., and D.P.; validation, A.C., I.S., I.P., L.B., A.O., G.F., F.L., E.B., S.M., A.M., G.T., R.M., V.V., G.S.; formal analysis, A.C., M.D.; investigation, A.C., I.S., M.D., and G.S.; writing—original draft preparation, D.B., M.D., L.B., I.P., A.O., G.F., F.L., E.B., S.M., A.D., A.S., and A.M.; writing—review and editing, D.B., M.D., L.B., I.P., A.O., G.F., F.L., E.B., S.M., A.D., A.S., R.M., G.D., A.F., G.R., and A.M.; visualization, all authors. supervision, G.S.; project administration, I.S. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data sharing not applicable.

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

#### **References**


### *Article* **Different Impact of Definitions of Sarcopenia in Defining Frailty Status in a Population of Older Women with Early Breast Cancer**

**Andrea Bellieni 1 , Domenico Fusco 1, \* , Alejandro Martin Sanchez 2 , Gianluca Franceschini 2 , Beatrice Di Capua 1 , Elena Allocca 3 , Enrico Di Stasio 4 , Fabio Marazzi 5 , Luca Tagliaferri 5 , Riccardo Masetti 2 , Roberto Bernabei <sup>6</sup> and Giuseppe Ferdinando Colloca 5**


**Abstract:** Sarcopenia is a geriatric syndrome characterized by losses of quantity and quality of skeletal muscle, which is associated with negative outcomes in older adults and in cancer patients. Different definitions of sarcopenia have been used, with quantitative data more frequently used in oncology, while functional measures have been advocated in the geriatric literature. Little is known about the correlation between frailty status as assessed by comprehensive geriatric assessment (CGA) and sarcopenia in cancer patients. We retrospectively analyzed data from 96 older women with early breast cancer who underwent CGAs and Dual X-ray Absorptiometry (DXA) scans for muscle mass assessment before cancer treatment at a single cancer center from 2016 to 2019 to explore the correlation between frailty status as assessed by CGA and sarcopenia using different definitions. Based on the results of the CGA, 35 patients (36.5%) were defined as frail. Using DXA Appendicular Skeletal Mass (ASM) or the Skeletal Muscle Index (SMI=ASM/heightˆ2), 41 patients were found to be sarcopenic (42.7%), with no significant difference in prevalence between frail and nonfrail subjects. Using the European Working Group on Sarcopenia in Older People (EWGSOP2) definition of sarcopenia (where both muscle function and mass are required), 58 patients were classified as "probably" sarcopenic; among these, 25 were sarcopenic and 17 "severely" sarcopenic. Only 13 patients satisfied both the requirements for being defined as sarcopenic and frail. Grade 3-4 treatment-related toxicities (according to Common Terminology Criteria for Adverse Events) were more common in sarcopenic and frail sarcopenic patients. Our data support the use of a definition of sarcopenia that includes both quantitative and functional data in order to identify frail patients who need tailored treatment.

**Keywords:** sarcopenia; physical performance; frailty; older cancer patients

**Citation:** Bellieni, A.; Fusco, D.; Sanchez, A.M.; Franceschini, G.; Di Capua, B.; Allocca, E.; Di Stasio, E.; Marazzi, F.; Tagliaferri, L.; Masetti, R.; et al. Different Impact of Definitions of Sarcopenia in Defining Frailty Status in a Population of Older Women with Early Breast Cancer. *J. Pers. Med.* **2021**, *11*, 243. https:// doi.org/10.3390/jpm11040243

Academic Editor: Anguraj Sadanandam

Received: 23 February 2021 Accepted: 22 March 2021 Published: 26 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

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

#### **1. Introduction**

Breast cancer is the most frequent cancer diagnosed in women and the leading cause of cancer death among women [1]. About 60% of these new diagnoses involve patients > 65 years of age and about 40% of patients are >70 [2].

Chronological age per se is a misleading criterion when deciding the best treatment for older women with breast cancer. A group of older patients with the same cancer of identical chronologic age can demonstrate wide heterogeneity concerning vitality, comorbidity, functional status, physiologic reserve, and psychosocial functioning [3–5]. Nonetheless, the accrual of older adults in cancer trials has been poor and undermined by several barriers through the years [6]. This is a severe matter of concern when evidence-based guidelines are applied to older populations, with negative consequences on survival [7]. Thus, a personalized approach based on individual patients' clinical conditions and functionality rather than age [8–11] should be considered the standard of care for older women with breast cancer.

To help guide treatment decisions, two geriatric medicine features have been incorporated in geriatric oncology: the concept of frailty and the comprehensive geriatric assessment. The Comprehensive Geriatric Assessment (CGA) represents the most efficient evaluation instrument, as recommended by the International Society of Oncological Geriatrics (SIOG) [12] and recently by the American Society of Clinical Oncology (ASCO) [13], to identify and define the frailty of the patient and his/her functional reserve [14]. Despite accumulating evidence regarding the value of the geriatric assessment in terms of encompassing older patients' diversity, a full CGA is considered rather time-consuming. Its effectiveness is far limited without interdepartment collaborative care and frailty-targeted optimized intervention programs to implement daily oncology practices [15–21]. CGA is the only method capable of assessing older cancer patients' frailty, predicting the risk of toxicity related to the treatments and the risk of mortality [22]. The CGA approach is considered essential to identify problems that are not immediately evident. Several studies have demonstrated the ability of CGA to identify otherwise unrecognized conditions of vulnerability to support the decision-making of the specialist (oncologist, radiotherapist, surgeon) when estimating the risk of toxicity to prevent said toxicity and preserve the functional performance of patients [23–26].

CGA can help to identify several geriatric syndromes [27]. Among all of them, sarcopenia has played an increasing role [28]. Sarcopenia is now considered one of the biological mechanisms underlying the concept of frailty. A reduction, compared to physiological criteria, in skeletal muscle mass characterizes this, with essential structural changes in muscle quality, and typically manifests itself with an alteration in function and/or a reduction in strength [29,30]. Several studies have shown the association between sarcopenia and functional decline, disability, frailty, falls, risk of fractures, multiple hospitalizations, and death [31,32]. A high prevalence of sarcopenia has been described in cancer patients, and its occurrence is associated with an increased risk of treatment toxicity, increased postoperative complications, increased sensitivity to antiblastic treatments, and a higher mortality rate, regardless of cancer stage [33]. It should also be stressed that cancer and cancer treatments may themselves be responsible for increasing disability, thereby accelerating the functional decline trajectory.

Several definitions of sarcopenia have been proposed. Initially, low muscle mass was considered the only criterion for diagnosis [34]. This is also the case for the vast majority of reports on cancer populations, with different indexes and cut-offs proposed. By contrast, in the geriatric field, the role of physical performance and muscle strength has been stressed as a necessary complement to the definition. The original operational definition of sarcopenia by the European Working Group on Sarcopenia in Older People (EWGSOP) [35] in 2010 was a significant change at the time, adding the muscle function to the former definitions, which were based only on detection of low muscle mass [28,29,31,32]. In its 2018 definition (Table 1), EWGSOP2 uses low muscle strength as a primary parameter of sarcopenia. It is considered a more reliable measure of muscle function and a better predictor of adverse

outcomes [36,37]. Specifically, sarcopenia is probable when low muscle strength is detected. A sarcopenia diagnosis is confirmed by the presence of low muscle quantity or quality. When low muscle strength, low muscle quantity/quality, and low physical performance are all detected, sarcopenia is considered severe (Table 1). Techniques for evaluating muscle quantity are available in many but not all clinical settings. As instruments and methods for assessing muscle quality are developed and refined in the future, this parameter is expected to grow in importance as a defining feature of sarcopenia. Physical performance was formerly considered part of the core definition of sarcopenia. In the revised guidelines, it is used to categorize the severity of sarcopenia.

**Table 1.** Definition of sarcopenia by European Working Group on Sarcopenia in Older People (EWGSOP2) guidelines <sup>1</sup> .


<sup>1</sup> Cruz-Jentoft et al., (2019). Sarcopenia: revised European consensus on definition and diagnosis. Age and Ageing, 48(1):16–31.

The present study aimed to assess sarcopenia's prevalence using different definitions in a population of older women with breast cancer and investigate possible correlations between sarcopenia and frailty status and the impact of these conditions on toxicities from oncological treatments.

#### **2. Materials and Methods**

We retrospectively analyzed data on the comprehensive geriatric evaluation of older women admitted at the Breast Surgery Unit of the Fondazione Policlinico Universitario A. Gemelli IRCCS, starting in January 2016 and ending in December 2019. All breast cancer patients aged ≥ 70 with a histological confirmed early breast cancer (stage 0–III, according to TNM) underwent CGA. The patients were selected weekly during the multidisciplinary tumor board (TBM), based on the registry criteria, and sent for geriatric evaluation. The only exclusion criteria were: life expectancy less than six months and refusal to participate in the study. Anthropometric measures (weight, height, BMI), the socio-family context, and support of all the patients were recorded and investigated. The patients underwent a medical examination, including medical history and physical examination. The primary socio-demographic data, the comorbidities, and the information on the oncological history and the anatomo-pathological and cancer immunohistochemical features, in accordance with the data present in the patients' medical records, were detected. Anthropometric measures (weight, height, body mass index) were collected for all patients. The comprehensive geriatric assessment (CGA) was based on recommendations from SIOG and national clinical guidelines [38]. The following areas were evaluated: performance status by Eastern Cooperative Oncology Group (ECOG) [39,40], comorbidity burden by the Charlson Comorbidity Index [41], functional status by Activity of Daily Living (Katz ADL) [42] and by Instrumental Activities of Daily Living (Lawton IADL) [43], cognition by Mini-Mental State Examination (MMSE) [44], nutritional status by Mini Nutritional Assessment (MNA) [45],

mood by Geriatric Depression Scale (GDS) [46], physical performance by Short Physical Performance Battery (SPPB), gait speed and time up-and-go test (TUGT) [47–49], muscular strength by handgrip (Jamar dynamometer) [50] and chair stand test [51]. Patients were asked about the presence of common geriatric syndromes, such as falls or incontinence. Only patients who completed a Dual X-ray Absorptiometry (DXA) scan for muscle mass evaluation were included for the present study.

#### *2.1. Sarcopenia and Frailty Definitions*

The definition of sarcopenia by EWGSOP2 [28] was applied, using cut-offs proposed by the guidelines mentioned above (Table 1). Muscle mass was measured by (DXA) total body (Hologic Horizon) Appendicular Skeletal Mass (ASM), calculated as the sum of arm and limb lean mass measured through DXA and expressed in kg. Frailty was defined by Balducci's criteria [52,53] considered as the detection of deficits in two or more domains of the CGA.

#### *2.2. Toxicities*

We retrospectively analyzed hospital electronic medical records of the patients included in the present study after a 12-month follow-up period in order to detect toxicities as they were reported by treating clinicians. Toxicities were evaluated using Common Terminology Criteria for Adverse Events (CTCAE) v5.0.

#### *2.3. Analysis*

All evaluations were performed by geriatricians belonging to the geriatric oncology team of the Fondazione Policlinico Universitario A. Gemelli IRCCS and who were specialized in the field of geriatric oncology and appropriately trained within the training courses of the International Society of Oncological Geriatrics (SIOG) [54]. Once the data collection was completed, all analyses were carried out using IBM SPSS 23. The collected data were synthesized using means and standard deviations for continuous variables and absolute and percentage frequencies for categorical variables. Statistical significance was conventionally set at *p* < 0.05.

#### **3. Results**

From January 2016 to December 2019, over 300 elderly patients aged ≥ 70 years belonging to the Breast Surgery Unit of Fondazione Policlinico Universitario A. Gemelli IRCCS (Rome), were evaluated.

Using the inclusion criteria, 96 patients were enrolled. The medium age of the examined sample was 76.9 (70 ÷ 89; SD 4.586), with an average level of comorbidity measured by the Charlson Comorbidity Index (CCI) of 6.7 (5 ÷ 13; SD—1.904), while ECOG performance status was mainly between 0 and 1 (89.6% of patients). Invasive ductal carcinoma was the most common histotype (75%), followed by lobular carcinoma (14.6%) (Table 2).

All patients underwent surgery: 84.38% (81) received a conservative treatment (quadrantectomy), representing 12.3% of cases (10 patients) with total lymphadenectomy, while 41.9% of cases (34 patients) received the removal of the sentinel lymph node. A total of 14.58% (14) received a full mastectomy, of whom three also underwent total lymphadenectomy. Less than 20% of patients received adjuvant chemotherapy, while almost two-thirds received adjuvant radiotherapy. In total, 85.4% of the patients were prescribed hormone suppressive therapy (with an aromatase inhibitor), based on hormone receptor status.


**Table 2.** Characteristics of the study population.

ECOG: Eastern Cooperative Oncology Group.

Based on CGA results, 35 patients (36.5% of the sample) were defined as frail, according to Balducci's criteria, and 61 (63.5) as nonfrail (Table 3).



CCI = Charlson Comorbidity Index; ADL = Activities of Daily Living; IADL = Instrumental Activities of Daily Living; MMSE = Mini-Mental State Examination; MNA = Mini Nutritional Assessment; GDS = Geriatric Depression Scale; SPPB = Short Physical Performance Battery; TUGT = Time Up and Go Test; BMI = Body Mass Index; SMI = Skeletal Muscle Index; ASM = Appendicular Skeletal Muscle mass.

Frail patients were older compared to nonfrail ones (79 years, SD 4.994; 75.67, SD 3.88; *p* = 0.000) and had a slightly higher burden of comorbidities (mean CCI of frail patients was 7.71 against 6.11 for nonfrail patients, *p* = 0.10) and a higher level of disability (ADL mean 5 vs. 5.72 for nonfrail; IADL mean 5.4 vs. 7.64 for nonfrail; *p* = 0.000), and were at higher risk of malnutrition (MNA mean 23.12 vs. 25.87; *p* = 0.001).

The cognitive level of frail patients assessed by the MMSE screening test was almost 2 points lower than the other patients (25.09 frail patients; 27.88 nonfrail patients; *p* = 0.001) and they had a higher frequency of depressive symptoms than the nonfrail ones (average GDS 6.19 vs. 3.47 for nonfrail; *p* = 0.000). Polypharmacy, defined as taking five or more medications daily, was the case for 74.3% of frail patients and 49.2% of nonfrail patients.

Using the DXA parameters (either appendicular skeletal mass (ASM) or Skeletal Muscle Index {SMI = ASM/heightˆ2]), 41 of 96 patients undergoing evaluation by DXA were found to be sarcopenic (42.7% of the sample examined) and 55 nonsarcopenic (57.3%). The average SMI of the sample was 6.47 (4.91 ÷ 9.73; SD 0.893). There were no significant differences in the prevalence of sarcopenia between frail and nonfrail patients (see Table 3)

According to the revised EWGSOP2 [28] criteria, 58 patients could be classified as "probably" sarcopenic with low muscle strength, defined as a chair stand test > 15 s for five rises (average value 17.09 s; 15.07 ÷ 26.7; SD 4.175). Among them, only 25 (out of 58) had a confirmed diagnosis of sarcopenia (either ASM < 15 kg or SMI DXA < 5.5 kg/m<sup>2</sup> ) with an average ASM of 13.46 kg and an average SMI value of 5.6 kg/m<sup>2</sup> . In total, 17 (out of 25) patients could be defined as severely sarcopenic with an SPPB score ≤ 8 (mean value 4.7) (Figure 1).

Frail sarcopenic patients had a mean ASM of 12.89 kg (SD 1.087) and a mean SMI value of 5.49 kg/m<sup>2</sup> (SD 0.376). Frail nonsarcopenic patients had a mean ASM of 17.39 kg (SD 2.47) and a mean SMI value of 7.11 kg/m<sup>2</sup> (SD 0.991).

**Figure 1.** Prevalence of sarcopenia according EWGSOP2 definition [16]. SMI = =Skeletal Muscle Index; ASM = Appendicular Skeletal Muscle mass; SPPB = Short Physical Performance Battery

Figure 2 shows the overlap between sarcopenia and frailty (as assessed by the results of CGA). Only 13 patients satisfied both the requirements for being defined sarcopenic ("confirmed" sarcopenia along to EWGSOP2) and frail (using modified Balducci's criteria derived from CGA). Among the sarcopenic population, the proportion of patients that are also frail increases, moving from "probable" sarcopenia to "severe" sarcopenia (proportion of frail patients is 55.2% for "probable", 56.5% for "confirmed", and 72.2% for "severe" sarcopenia).

**Figure 2.** Prevalence of frailty and sarcopenia in the study sample.

In a one-year follow-up, the whole sample reported 52 cases of treatment toxicities (54.16%). According to the Common Terminology Criteria for Adverse Events (CTCAE), in the frail group 17 out of 35 patients developed toxicities of any types: five patients had grade 3–4 toxicities (14%). Among sarcopenic patients, 12 out of 23 patients developed toxicities of any types; five patients experienced grade 3–4 toxicities (22%).

Among patients reporting toxicities, frail patients reported Grade 3–4 toxicities (according to CTCAE) more frequently than nonfrail (29% vs. 14%) ones, while sarcopenic patients reported G3–G4 toxicities more than nonsarcopenic patients (42% vs. 13%) (Figure 3).

**Figure 3.** Percentage of patients who experienced Grade 1–2 and Grade 3–4 toxicities (according to Common Terminology Criteria for Adverse Events, CTCAE) in our population.

#### **4. Discussion**

In the aging scenario of the general population and the increasing number of diagnosed and treated cancers in older adults, it has become critical to identify, understand, and assess the so-called geriatric syndromes. Among these, more attention is being placed on sarcopenia. For this reason, it has become essential to know the differences between sarcopenia and the loss of muscle mass related to the normal process of muscle aging or other pathological conditions such as cachexia [30].

In our sample, we identified different frequencies of sarcopenia depending on the definition used. Sarcopenia can be defined as a pathological loss of skeletal muscle mass

characterized by essential structural changes in muscle quality, which occurs in older adults and shows functional impairment and/or strength reduction [55]. Aging is related to a decline in muscle mass and strength [56] but only when this decline becomes pathological (sarcopenia) does this process lead to adverse health outcomes [57].

In patients with cancer, many studies showed how the loss of muscle mass is a prevalent condition independent of disease stage and body mass [58]. This is due to many factors leading to the deterioration of muscles: inflammation, cancer-derived catabolic factors, malnutrition, reduced physical activity, and the effect of cytotoxic and targeted treatments on muscle mass and quality [10,59].

Loss of muscle mass can precede the cancer and further complicate its course, predisposing patients to a shorter time of tumor progression, increased chemotherapy-related toxicity, postoperative complications, poor functional status, hospitalization, increased length of hospital stay, high 30-day readmission rate, and mortality [60]. While the loss of muscle mass has been proven to be an independent predictor of adverse outcomes at all ages and for several cancers, such as breast cancer, hepatocellular carcinoma, and advanced urothelial cancer [60,61], the presence of sarcopenia in older cancer patients, who could be at higher risk for this condition, has been associated with therapy-related toxicities and increased adverse outcomes [33], [61–63].

In oncology, the skeletal muscle index (SMI) is the main parameter used to evaluate sarcopenia. Several epidemiologic studies have determined the prevalence of sarcopenia using cut-off values determined by CT scans or DXA when muscle mass is normalized for height [64,65]. Muscle function has rarely been taken into consideration. This is mainly due to most existing studies' retrospective natures, relying on large CT scan datasets for oncological reasons (disease staging or surgical evaluation). At the same time, physical function tests are seldom conducted in routine clinical practice.

In recent years, first in the geriatric field and then in other settings, the definition of sarcopenia has shifted from an evaluation of muscle mass to a qualitative assessment of muscle function. Physical performance is a powerful predictor of adverse outcomes. This concept has been an integral part of the "physical frailty phenotype" construct [32]. Indeed, sarcopenia can be considered the most relevant biological determinant of physical frailty. Moreover, measurement of muscle mass and quality has many technological limitations, while the selection of specific cut-offs is a matter of debate [66], while muscle function is much easier to measure, at least in geriatric clinics (where hang grip or chair standing tests are routinely conducted). This new definition of sarcopenia correlates with many adverse outcomes (institutionalization risk, toxicity, mortality). This aspect has meant that sarcopenia, from a simple geriatric syndrome, has become one of the fundamental bases of modern geriatrics.

Still, in the nongeriatric setting, there is often confusion between frailty and sarcopenia, so we designed this study to try to identify those factors that can identify patients at greater risk of adverse outcomes.

Our study shows different ways to define sarcopenia and that the quantitative data (i.e., muscle mass measurement) alone is not sufficient. We detected a high prevalence of low muscle mass in the whole sample (almost 42.7%), almost equally distributed in frail and nonfrail patients. The proportion of patients with reduced muscle mass is in line with what has been reported in the literature [67]. However, when more stringent criteria that incorporate muscle function (such as EWGSOP 2) were used, only a limited proportion of patients (26%) could still be defined as "sarcopenic". Indeed, severely sarcopenic patients were almost always classified as frail on the results of CGA.

Even though sarcopenia has been regarded as a key component of the frailty status in older adults, it should be kept in mind that frailty is a multidimensional concept that goes beyond each of its features. Relying solely on what is usually defined as "sarcopenia" in oncological research (that is, low muscle mass) can be misleading, resulting in classifying many more patients as frail and possibly omitting valuable treatments. On the contrary, a stricter definition, where muscle loss is coupled with reduced muscle function (both strength and performance), could enable a better selection of patients.

In this study, 13.5% of patients have a correspondence between sarcopenia and frailty; this group has a very high probability of experiencing adverse outcomes. Identifying this subgroup will allow a real personalization of treatments in the near future and the identification of the risk of adverse events for apparently fit patients.

DXA is a noninvasive instrument for body composition assessments. Information on muscle and adipose tissue can also be gathered by other tools such as CT scans. Identifying reduced muscle mass should be promoted in oncology to avoid the adverse outcomes associated with this condition [28]. For example, chemotherapy could be personalized based on body composition, with doses adapted to the individual patient to limit toxicities [68]. Frailty is better identified by CGA which allows the detection of unidentified problems and the correct malignancy prognosis estimation [69]. Thus, CGA avoids overand undertreatments in a scenario focus on tailored treatments.

In our sample, both frailty and sarcopenia are associated with treatment-related toxicities, especially with more severe (G3–G4) ones. However, this association is stronger for sarcopenic or sarcopenic-frail patients than it is for frail patients, although these differences are not statistically significant. It should be kept in mind that treating clinicians were aware of frailty status, so that more potentially toxic treatments were spared to frailer patients. On the other side, it is also possible that sarcopenia was not routinely considered when planning surgery or adjuvant therapies. This could have resulted in more adverse effects both in sarcopenic nonfrail patients and in sarcopenic frail patients, which indeed showed a comparable frequency of high-grade treatment-related toxicities.

The novelty of this study is in having identified for the first time in the same group of breast cancer patients the various degrees of sarcopenia and frailty through the available gold standards and a subgroup at high risk of adverse events (toxicity, reduced compliance, etc.) between the two.

Some limitations of our study have to be acknowledged. Firstly, given the crosssectional nature of our data, it was not possible to make inferences on otherwise clinically significant outcomes associated with frailty and sarcopenia, such as survival or loss of functional independence. Secondly, the small sample size prevents us from generalizing our results to other clinical situations. Indeed, we believe that these data should prompt further research on the association between frailty, sarcopenia, and body composition, hopefully with a longer follow patients report to identify which parameter constitutes the best clinical deterioration predictor. More research is also needed on possible interventions to counteract sarcopenia, restoring muscle mass, and function. Physical exercise is a promising intervention that could prevent functional decline in older adults [70]. More data on the effect of structured physical activity in older adults with cancer are needed.

#### **5. Conclusions**

Our data support the use of a comprehensive definition of sarcopenia that takes into account both physical performance and muscle mass in order to identify older women with breast cancer at higher risk of clinical deterioration and treatment-related toxicities. A multidimensional geriatric assessment in this population is strongly recommended and evaluation of muscle mass and function should be regarded as an essential part of it, with the aim of offering patients the best personalized treatment.

**Author Contributions:** Conceptualization, A.B., D.F. and G.F.C.; methodology, A.B., D.F., A.M.S., B.D.C. and E.A.; data curation, A.B., D.F. and E.D.S.; writing—original draft preparation, A.B., D.F., B.D.C. and E.A.; writing—review and editing, A.M.S., G.F., F.M., L.T., R.M., R.B. and G.F.C.; supervision, D.F. and G.F.C.; project administration, R.B. All authors have read and agreed to the published version of the manuscript.

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

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

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

#### **References**


### *Article* **Liver Metastasectomy for Metastatic Breast Cancer Patients: A Single Institution Retrospective Analysis**

**Armando Orlandi 1, \* , Letizia Pontolillo 1,2 , Caterina Mele 3 , Mariangela Pasqualoni 1,2 , Sergio Pannunzio 1,2 , Maria Chiara Cannizzaro 1,2 , Claudia Cutigni 1,2 , Antonella Palazzo 1 , Giovanna Garufi 1,2 , Maria Vellone 2,3 , Francesco Ardito 2,3 , Gianluca Franceschini 2,4 , Alejandro Martin Sanchez 4 , Alessandra Cassano 1,2 , Felice Giuliante 2,3 , Emilio Bria 1,2 and Giampaolo Tortora 1,2**


**Abstract:** The liver represents the first metastatic site in 5–12% of metastatic breast cancer (MBC) cases. In absence of reliable evidence, liver metastasectomy (LM) could represent a possible therapeutic option for selected MBC patients (patients) in clinical practice. A retrospective analysis including MBC patients who had undergone an LM after a multidisciplinary Tumor Board discussion at the Hepatobiliary Surgery Unit of Fondazione Policlinico Universitario "Agostino Gemelli" IRCCS in Rome, between January 1994 and December 2019 was conducted. The primary endpoint was overall survival (OS) after a MBC-LM; the secondary endpoint was the disease-free interval (DFI) after surgery. Forty-nine MBC patients underwent LM, but clinical data were only available for 22 patients. After a median follow-up of 71 months, median OS and DFI were 67 months (95% CI 45–103) and 15 months (95% CI 11–46), respectively. At univariate analysis, the presence of a negative resection margin (R0) was the only factor that statistically significantly influenced OS (78 months *versus* 16 months; HR 0.083, *p* < 0.0001) and DFI (16 months *versus* 5 months; HR 0.17, *p* = 0.0058). A LM for MBC might represent a therapeutic option for selected patients. The radical nature of the surgical procedure performed in a high-flow center and after a multidisciplinary discussion appears essential for this therapeutic option.

**Keywords:** metastatic breast cancer; liver metastases; hepatic surgery; personalized medicine

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

#### **1. Introduction**

Metastatic breast cancer (MBC) is the first oncological cause of death in women despite the advances in therapeutic strategies, with a 5-year survival of only ~25% [1,2]. The liver represents the first metastatic site in 5–12% of MBC [3] cases. Despite the transient response to chemo or endocrine therapy, most patients experience disease progression after 1–2 years [4]. While current evidence supports a liver metastasectomy (LM) for advanced colorectal cancer in improving survival [5,6] on the basis that hepatic parenchyma filters circulating tumor cells (CTC) from the primary neoplastic site to systemic circulation, LM is considered a possible therapeutic option for selected MBC patients in clinical

**Citation:** Orlandi, A.; Pontolillo, L.; Mele, C.; Pasqualoni, M.; Pannunzio, S.; Cannizzaro, M.C.; Cutigni, C.; Palazzo, A.; Garufi, G.; Vellone, M.; et al. Liver Metastasectomy for Metastatic Breast Cancer Patients: A Single Institution Retrospective Analysis. *J. Pers. Med.* **2021**, *11*, 187. https://doi.org/10.3390/jpm11030187

Academic Editor: Valeria Bertagnolo

Received: 27 December 2020 Accepted: 4 March 2021 Published: 8 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

147

practice, in the absence of prospective evidence. Several studies reported controversial results about the survival rate after hepatic loco-regional treatment in MBC with liver metastases with a 3-year and 5-year survival rate that ranged between 49–94% and 5–78% respectively [3,7–28]. A recent review of Bale et al. [29] showed that a primary tumor's characteristics such as small tumor size, nodes negativity, low grade, and early-stage may be associated with a better outcome after liver surgery. In addition, they evidenced as an independent positive prognostic factor a long interval between the primary diagnosis and the detection of breast cancer liver metastasis (BCLM) more than 1 year, liver-limited disease (with the exception of isolated pulmonary and bone metastasis), response to preoperative systemic therapy before hepatic surgery, and the BCLM expression of estrogen receptor (ER) and progesterone receptor (PgR). The major limits of the studies in the literature are represented by the small number of patients enrolled and the presence of multiple confounding factors for the heterogeneity of the biology of the primary tumor, the presence of synchronous and metachronous metastases, the presence of extrahepatic disease, and the types of systemic treatments used. However, in all studies, patients with a low burdendisease benefited from R0 resections of BCLM with an improvement in survival rate [7,30]. Therefore, we report data about our experience of MBC patients who underwent liver metastasis surgery.

#### **2. Materials and Methods**

#### *2.1. Study Design and Participants*

A retrospective analysis including MBC patients who had undergone LM after a multidisciplinary Tumor Board discussion at the Hepatobiliary Surgery Unit of the Fondazione Policlinico Universitario "Agostino Gemelli" IRCCS in Rome, between January 1994 and December 2019, was conducted.

Eligible patients were aged 18 years or older, had a histological diagnosis of invasive BC and synchronous or metachronous LM. All immunophenotype BC were eligible in the study: luminal (ER and/or PgR positive), epidermal growth factor receptor 2 (HER2) positive and triple-negative (TNBC: ER, PgR, and HER2 negative). In all patients, disease assessment was determined by computerized tomography (CT) scan and magnetic resonance imaging (MRI) of the liver. The presence of extrahepatic disease was allowed provided that these sites were stable or in response to previous systemic treatments before hepatic surgery. The evaluation of expression of ER, PgR, and HER2 was done respecting the ASCO-CAP guidelines. Using the pathology report after hepatic surgery, the presence or the absence of disease at the resection margin (R0: no disease at the resected surgical margin, R1: the presence of disease at the resected surgical margin) was determined. For each patient, demographic data were collected including gender and age. Clinicopathological data on menopausal status (defined retrospectively after a woman has experienced 12 months of amenorrhea without any other pathological or physiological cause), metastatic sites, hepatic metastases presentation, number of systemic therapy pre-hepatic surgery, histotype (ductal *versus* lobular), immunophenotype, and resection margins were also collected. The study was approved by the Institutional Review Boards.

#### *2.2. Study Endpoints*

The primary endpoint was overall survival (OS) after LM, defined as the time from LM to death; the secondary endpoint was the disease-free interval (DFI) after LM, defined as the time from surgery to recurrence (in patients with liver-only disease) or progression of the disease (in patients with extrahepatic metastases which was stable or in response to previous treatment before LM). An exploratory analysis was performed to evaluate the survival impact of demographic and clinicopathological factors: age (<50 *versus* ≥50 years old), menopausal state (pre-menopausal *versus* menopausal), metastatic sites (only liver *versus* other), hepatic metastases presentation (synchronous *versus* metachronous), number of liver metastases (1 *versus* > 1), number of systemic therapy pre-hepatic surgery (none *versus* ≥ 1), histotype (ductal *versus* lobular), immunophenotype (luminal *versus* TNBC *versus* HER2+), and hepatic resection margins (R0 *versus* R1).

#### *2.3. Statistical Analysis*

Statistical analyses were performed using MedCalc software version 14 (MedCalc Software Ltd, Ostend, Belgium). Survival curves were calculated according to the Kaplan-Meier method and differences in survival were assessed with the log-rank test. Independent predictors of disease-specific survival and recurrence were identified by Cox proportional hazard analysis. Statistical significance was defined as a *p*-value < 0.05. As the study was explorative, an estimate of the sample size was not calculated.

#### **3. Results**

#### *3.1. Demographic and Clinicopathological Characteristics of Patients*

During the study period, a total of 49 patients, all female, underwent LM at our Hospital. Clinical data were available for 22. Patient age at the time of surgery ranged from 34 to 71 years with a median age of 48 years. Ten patients were premenopausal, 12 postmenopausal. Nineteen patients had isolated liver disease, 3 patients had multiorgan metastasis. Among patients with multi-organ metastasis, 2 had bone metastasis and 1 adrenal metastasis. Liver metastasis was metachronous for 17 patients and synchronous for 5 patients. Seven patients underwent surgery upfront, while 15 patients received one line of systemic treatment prior to surgery; the best response to systemic treatment was a partial response (PR) for 11 patients, 3 patients had a stable disease (SD), and only one had a progression disease (PD), with a disease control rate (DCR: PR + SD) of 93%. The histotype was ductal carcinoma for 21 patients, only 1 was lobular; 14 patients had a luminal tumor, 3 patients were HER2+, and 5 patients were TNBC. Nine patients underwent anatomical liver resection (resection of segments in 7 patients and resection of the left hepatic lobes in 2 patients were done) and 13 patients received metastasectomies (not anatomical liver resection). The resection margin was negative (R0) in 20 patients and positive in 2 patients. Among the 11 patients who had obtained a partial response, 4 patients had a pathological complete response (only fibrosis was found in the absence of neoplastic cells). Postoperative mortality (mortality within one month after hepatic surgery) was 0%. Complications occurred only in two patients: 1 patient presented perihepatic abscess and 1 patient with perihepatic abscess and a pulmonary embolism; both cases were resolved with medical therapy. All patients received at least one line of systemic therapy in the post-surgery setting: as maintenance of the previous treatment (hormonal therapy for luminal therapy, trastuzumab +/− hormonal therapy for HER2+ and the same chemotherapy in TNBC) and a new line of therapy after recurrence/progression of the disease.

Demographic and clinicopathological characteristics of patients are listed in Table 1.


**Table 1.** Demographic and clinicopathological characteristics of patients (*n* = 22) and correlation with the disease-free interval (DFI) and overall survival (OS).


**Table 1.** *Cont.*

DFS: disease-free survival after liver resection; OS: overall survival after liver resection. N.: number. N.R.: not reached. Italics and bold: statistical significant *p*-value.

#### *3.2. Survival Outcomes*

At the data cut-off analysis of May 2020, 11 patients were still alive and 7 patients were free of progression disease after hepatic surgery. Of the 15 patients who experienced recurrence, 8 have had disease progression with liver metastases, 3 with liver and bone metastases, 3 with lung metastases, and 1 with brain metastases. After a median follow-up of 71 months, median OS was 67 months (95% CI 45–103) (Figure 1) while median DFI was 15 months (95% CI 11–46) (Figure 2), respectively.

At univariate analysis, the presence of a negative resection margin was the only factor that statistically significantly influenced OS (78 *versus* 16 months; HR 0.083, *p* < 0.0001) (Figure 3) and the DFI (16 *versus* 5 months; HR 0.17, *p* = 0.0058) (Figure 4). None of the other factors were significantly associated with OS and the DFI; their association with the DFI and OS is shown in Table 1. A trend toward significance (the boundary of *p*-value < 0.2) was observed in the OS analysis for metastatic sites (only liver *versus* other sites, 103 *versus* 50 months, *p* = 0.14) while a prior systemic therapy showed a trend in favor also for the DFI (none *versus* ≥ 1, 14 *versus* 46 months, *p* = 0.1). The multivariate analysis confirmed the negative resection margin as the only factor which statistically significantly influenced OS (*p* = 0.0034) and DFS (*p* = 0.024).

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 5 of 10

**Figure 1.** OS in the study population (*n* = 22): median OS was 67 months (95% CI 45–103). **Figure 1.** OS in the study population (*n* = 22): median OS was 67 months (95% CI 45–103). **Figure 1.** OS in the study population (*n* = 22): median OS was 67 months (95% CI 45–103).

**Figure 2.** DFS in the study population (*n* = 22): median DFI was 15 months (95% CI 11–46). **Figure 2.** DFS in the study population (*n* = 22): median DFI was 15 months (95% CI 11–46).

**Figure 2.** DFS in the study population (*n* = 22): median DFI was 15 months (95% CI 11–46). At univariate analysis, the presence of a negative resection margin was the only factor that statistically significantly influenced OS (78 *versus* 16 months; HR 0.083, *p* < 0.0001) (Figure 3) and the DFI (16 *versus* 5 months; HR 0.17, *p* = 0.0058) (Figure 4). None of the other factors were significantly associated with OS and the DFI; their association with the DFI and OS is shown in Table 1. A trend toward significance (the boundary of *p*-value < 0.2) was observed in the OS analysis for metastatic sites (only liver *versus* other sites, 103 *versus* 50 months, *p* = 0.14) while a prior systemic therapy showed a trend in favor also for At univariate analysis, the presence of a negative resection margin was the only factor that statistically significantly influenced OS (78 *versus* 16 months; HR 0.083, *p* < 0.0001) (Figure 3) and the DFI (16 *versus* 5 months; HR 0.17, *p* = 0.0058) (Figure 4). None of the other factors were significantly associated with OS and the DFI; their association with the DFI and OS is shown in Table 1. A trend toward significance (the boundary of *p*-value < 0.2) was observed in the OS analysis for metastatic sites (only liver *versus* other sites, 103 *versus* 50 months, *p* = 0.14) while a prior systemic therapy showed a trend in favor also for Clinicopathological characteristics of patients with R0 resection are listed in Table 2. Of the 20 patients with an R0 resection, 13 patients had a single lesion while 7 had two metastases. Radiological dimensions of the liver lesions are available for 13 of the 20 patients and ranged from 9 mm to 80 mm; histological dimensions were available for 15 patients and ranged from 4 to 35 mm. Fourteen patients had received one line of systemic treatment before the surgery: 4 patients had a complete response (CR), 6 patients had a partial response (PR), 3 patients had a stable disease (SD), and one patient experienced a progression of the disease (PD) before hepatic surgery; the DCR was 92.8%. Of about the 20 patients with an R0 resection, 19 had an immunophenotype of liver metastases

consistent with primary tumor: 13 patients had a luminal immunophenotype (one of which was HER2 positive at diagnosis), 3 were HER2 positive, and 4 patients were TNBC. the DFI (none *versus*  1, 14 *versus* 46 months, *p* = 0.1). The multivariate analysis confirmed the negative resection margin as the only factor which statistically significantly influenced the DFI (none *versus*  1, 14 *versus* 46 months, *p* = 0.1). The multivariate analysis confirmed

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 6 of 10

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 6 of 10

Of the 2 patients with R1 resection, one had multiple (six) liver lesions and one a single metastasis, both were luminal consistent with primary tumor immunophenotype. OS (*p* = 0.0034) and DFS (*p* = 0.024). the negative resection margin as the only factor which statistically significantly influenced OS (*p* = 0.0034) and DFS (*p* = 0.024).

**Figure 3.** OS according to resection margin of liver metastases (R0 versus R1: 78 *versus* 16 months; HR 0.083, *p* < 0.0001). **Figure 3.** OS according to resection margin of liver metastases (R0 versus R1: 78 *versus* 16 months; HR 0.083, *p* < 0.0001). **Figure 3.** OS according to resection margin of liver metastases (R0 versus R1: 78 *versus* 16 months; HR 0.083, *p* < 0.0001).

**Figure 4.** DFI according to resection margin of liver metastases (R0 versus R1: 16 *versus* 5 months; HR 0.17, *p* = 0.0058).


**Table 2.** Characteristic of patients with liver metastasectomy with negative resection margin (R0).

N.: number. \*: number of patients who received treatments before surgery. CR: complete response, PR: partial response, SD: stable disease.

#### **4. Discussion**

Despite an improvement in the systemic treatment of MBC, the median survival of patients with metastatic disease is between 18 and 24 months [31]. Resection of breast cancer liver metastasis may represent a therapeutic option for selected patients. The radical nature of the surgical procedure performed in a high-flow center and after a multidisciplinary discussion appears essential for this therapeutic option [32] like in other neoplastic diseases [33].

In a recent systematic review of resection of MBC-LM, the median OS was 35.1 months and the median DFS was 21.5 months [34]. At the same time, in a case-matched analysis, the resection group had an impressive median OS of 82 months versus a median OS of 31 months in the systemic group, so the authors concluded that the combination of surgery with systemic treatment results in an improved OS [7].

The median OS and the DFI in our population were 67 months and 15 months respectively. Thus, our study seems to confirm a possible survival benefit in patients undergoing liver surgery of metastases especially in patients with an R0 resection. In fact, in our study, the presence of a negative resection margin was the only factor that statistically significantly influenced OS (78 *versus* 16 months; HR 0.083, *p* < 0.0001) and DFI (16 *versus* 5 months; HR 0.17, *p* = 0.0058). Of the 20 patients with an R0 resection, 13 patients had a single lesion while 7 had two metastases, this implies that a careful selection of patients with limited liver disease is important to obtain an adequate surgical result.

Fourteen patients received one line of treatment before surgery with a DCR of 93%; therefore, it also emerged in this evaluation that the selection of patients with a metastatic disease under control by systemic treatment can allow an important result. However, it is equally important to note that also the patients with PD during systemic therapy before liver surgery achieves an R0 resection, demonstrating how liver resection can also be proposed as a salvage treatment in highly selected cases. Moreover, surgical complications only occurred in two patients in the absence of post-surgery mortality, these data suggest that liver metastasectomy could be a safe procedure. At the same time, in our population, surgical radicality was achieved in almost all patients who were eligible for the study, and the R0 margin was found as the only prognostically relevant factor influencing both the DFI and OS. Taken together, these results confirm the importance of performing LM exclusively in high-flow centers, post a multidisciplinary discussion since, under this condition, the removal of liver metastases from breast cancer can significantly influence the survival of patients without significant side effects.

The main bias of the study is the smallness of the sample, also due to the limited availibility data of all the population of patients with MBC and undergoing liver surgery in our institution (49 versus 22). The sample size might have affected the results, limiting the statistical significance impact for some factors analyzed that appear to have a role in survival. In addition, the DFI and OS may have been influenced by the subsequent systemic therapies, but due to the heterogeneity of the population and the number and characteristics of treatment received post-surgery, it is not possible to evaluate their impact on survival outcomes. The trend benefit of the low tumor burden (only liver *versus* other sites) is in line with other results and with the suggestion of international guidelines to justify a multidisciplinary and more aggressive therapy in patients with limited metastatic disease in order to obtain a greater chance of healing [1]. The correlation with the menopausal state with better prognosis, on the other hand, could be related to a lower biological aggressiveness of the disease. Additionally, the trend benefit of the use of systemic presurgery treatment, as it has been shown in previous studies [3,8,18], seems to have a role in the increasing survival eradicating or debulking microscopic lesions. In contrast to other studies [25], in our population, we did not note an improved outcome for patients with luminal disease.

#### **5. Conclusions**

Despite the limitations imposed by a retrospective analysis on a small sample, our study confirms the possible positive role of R0 surgical excision of liver metastases from MBC if performed in a high-flow center after multidisciplinary evaluation. The prospective confirmation of this data appears to be increasingly necessary in order to consolidate the use of locoregional treatments in oligometastatic breast cancer disease, in particular, to identify the subgroup of patients who can benefit from surgical treatment.

**Author Contributions:** Conceptualization, A.O.; methodology, A.O., E.B. and G.T.; Writing—review and editing, all authors; supervision, A.O., E.B. and G.T. All authors have read and agreed to the published version of the manuscript

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

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Fondazione Policlinico Universitario "A. Gemelli" IRCCS Roma (Prot N. 0011515/20-12/03/2020).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

**Acknowledgments:** E.B. is currently supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC) under Investigator Grant (IG) No. IG20583. GT is supported by AIRC, IG18599, AIRC 5 × 1000 21052. E.B. is currently supported by Institutional funds of Università Cattolica del Sacro Cuore (UCSC-project D1-2018/2019).

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

#### **References**


### *Article* **Sentinel Node Biopsy after Neoadjuvant Chemotherapy for Breast Cancer: Preliminary Experience with Clinically Node Negative Patients after Systemic Treatment**

**Alejandro Martin Sanchez 1, \* , Daniela Terribile 1,2 , Antonio Franco 1 , Annamaria Martullo 1 , Armando Orlandi 1,3 , Stefano Magno 1 , Alba Di Leone 1 , Francesca Moschella 1 , Maria Natale 1 , Sabatino D'Archi 1 , Lorenzo Scardina 1 , Elena J. Mason 1 , Flavia De Lauretis 1 , Fabio Marazzi 4 , Riccardo Masetti 1,2 and Gianluca Franceschini 1,2**


**Abstract:** Sentinel lymph node biopsy (SLNB) following neoadjuvant treatment (NACT) has been questioned by many studies that reported heterogeneous identification (IR) and false negative rates (FNR). As a result, some patients receive axillary lymph node dissection (ALND) regardless of response to NACT, leading to a potential overtreatment. To better assess reliability and clinical significance of SLNB status on ycN0 patients, we retrospectively analyzed oncological outcomes of 399 patients treated between January 2016 and December 2019 that were either cN0-ycN0 (219 patients) or cN1/2-ycN0 (180 patients). The Endpoints of our study were to assess, furthermore than IR: oncological outcomes as Overall Survival (OS); Distant Disease Free Survival (DDFS); and Regional Disease Free Survival (RDFS) according to SLNB status. SLN identification rate was 96.8% (98.2% in patients cN0-ycN0 and 95.2% in patients cN+-ycN0). A median number of three lymph nodes were identified and removed. Among cN0-ycN0 patients, 149 (68%) were confirmed ypN0(sn), whereas regarding cN1/2-ycN0 cases 86 (47.8%) confirmed an effective downstaging to ypN0. Three year OS, DDFS and RDFS were significantly related to SLNB positivity. Our data seemed to confirm SLNB feasibility following NACT in ycN0 patients, furthermore reinforcing its predictive role in a short observation timing.

**Keywords:** neoadjuvant chemotherapy; sentinel lymph node; breast cancer; systemic treatment; locally advanced breast cancer; mini-invasive treatment

#### **1. Introduction**

Sentinel lymph node biopsy (SLNB) is considered the gold standard for axillary staging in early breast cancer patients with clinically negative lymph nodes (cN0), as it reduces potential complications of axillary dissection (ALND) [1–4].

**Citation:** Sanchez, A.M.; Terribile, D.; Franco, A.; Martullo, A.; Orlandi, A.; Magno, S.; Di Leone, A.; Moschella, F.; Natale, M.; D'Archi, S.; et al. Sentinel Node Biopsy after Neoadjuvant Chemotherapy for Breast Cancer: Preliminary Experience with Clinically Node Negative Patients after Systemic Treatment. *J. Pers. Med.* **2021**, *11*, 172. https://doi.org/ 10.3390/jpm11030172

Academic Editor: Raghu Sinha

Received: 21 January 2021 Accepted: 22 February 2021 Published: 2 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

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

Since many studies have shown a great variation in identification (IR) and false negative rates (FNR), the reliability of SLNB after neoadjuvant chemotherapy (NACT) remains questionable [5–10].

As a result, several patients continue to undergo complete axillary dissection, regardless of axillary staging and response to neoadjuvant chemotherapy, leading to a potential overtreatment of both cN0 and cN1/2 patients who remained or became ycN0 after NACT [3].

For this particular subgroup of patients, recent studies reported acceptable IR and FNR, suggesting that SLNB could be feasible in cN0-ycN0 patients and also in women who are cN1/2 before chemotherapy and achieve an ycN0 status [10–12].

Moreover, for cN0-ycN0 patients, a recently published retrospective study correlates a metastatic SLN with a significant worsening of oncological outcomes, such as Distant Disease Free Survival (DDFS), proving that SLNB is not only feasible after NACT, but that in this setting it could be a good predictive tool to better assess patients at risk [13].

The aim of this analysis, besides reporting our personal workout model for patients receiving neoadjuvant regimens, is to better assess the feasibility and prognostic significance (according to status) of SLNB in ycN0 patients.

With this purpose, we retrospectively analyzed clinical and oncological results obtained from cT1-4 breast cancer patients who were either cN0 or cN1/2 prior to neoadjuvant treatment and became or remained cN0 at the end of the systemic therapy (ycN0).

The endpoints of our study were to evaluate, furthermore than IR: oncological outcomes as Overall Survival (OS); Distant Disease Free Survival (DDFS); and Regional Disease Free Survival (RDFS) according to SLNB status.

#### **2. Materials and Methods**

From the prospectively maintained database of the Multidisciplinary Breast Center of the Fondazione Policlinico Universitario Agostino Gemelli IRCCS in Rome, we identified patients with locally advanced breast cancer (cT1-cT4 patients, cN0-cN1/2) who had received neoadjuvant chemotherapy and remained or became ycN0, subsequently undergoing breast surgery and SLNB, between 2016 and 2019.

We excluded from our analysis ycN0 patients in whom a SLN was not identified during surgical procedure, who consequently underwent direct ALND.

Endpoints of our study were:


#### **3. Clinical Workout**

The indication for neoadjuvant treatment (chemotherapy or endocrine therapy) and surgical management of the axilla were discussed during a multidisciplinary meeting (MDM) of breast surgeons, medical oncologists, radiation oncologists, radiologists, pathologists and geneticists.

According to national and international guidelines Associazione Italiana di Oncologia Medica (AIOM) 2019 and National Comprehensive Cancer Network (NCCN) 2020), patients underwent NACT in the following cases:


Pre-neoadjuvant clinical staging: Locoregional staging was assessed by clinical examination, breast and axillary ultrasound, mammography, breast magnetic resonance, or core biopsy of both breast lesion and suspected axillary lymph nodes.

The systemic staging was assessed by total body computed tomography scan or positron emission tomography and bone scintigraphy.

Neoadjuvant regimens: NACT regimen depended on stage and tumor characteristics. We used the following chemotherapy schemes:

	- Sequential scheme: Anthracyclines plus Cyclophosphamide on day 1 every 21 days for 4 cycles (4 AC); followed by docetaxel on day 1 every 21 days for 4 cycles or paclitaxel on day 1 every week for 12 cycles.
	- 6 TAC: docetaxel plus Doxorubicin plus Cyclophosphamide on day 1 every 21 days for 6 cycles.
	- 6 TCH: docetaxel plus Carboplatin plus Herceptin on day 1 every 21 days for 6 cycles.
	- Sequential scheme: Anthracyclines plus Cyclophosphamide on day 1 every 21 days for 4 cycles (4 AC); followed by docetaxel on day 1 every 21 days for 4 cycles or paclitaxel on day 1 every week for 12 cycles plus Herceptin on day 1 every 21 days for 18 cycles.

Hormone therapy with aromatase inhibitor was delivered to elder and fragile postmenopausal patients with locally advanced breast cancer expressing hormone receptors (ER, PgR) and low Ki-67 (Luminal A and Luminal B). Neoadjuvant protocol was administered for at least six months.

Clinical assessments during and after NACT: Before each cycle of chemotherapy, patients underwent treatment response monitoring with a clinical examination and "in office" breast/axillary ultrasound.

Patients with no evidence of clinical response or with disease progression were the subject of multidisciplinary discussion about a change in NACT scheme or immediate surgery.

One month after NACT finalization, loco-regional staging was repeated (clinical examination, breast and axillary ultrasound, mammography, breast magnetic resonance).

Breast surgical treatment: Surgical management was discussed during a dedicated MDM, taking into account the clinical restaging and patient's preferences.

Patients with a favorable ratio between breast volume and residual lesion were addressed by conservative techniques:


In case of unfavorable ratio between breast volume and residual tumor size, multicentric cancer, inflammatory cancer and contraindications to adjuvant radiotherapy patients were judged eligible for mastectomy techniques and immediate breast reconstruction (implant or autologous reconstruction):


Axillary assessment: Axillary workout is summarized in Figure 1.

**Figure 1.** Axillary workout in the neoadjuvant setting.

SLNB was performed using blue dye technique (Patent Blue V or Methylene blue, 2–5 cc) injected sub-dermally, 15–30 min before surgery. Blue-stained axillary lymph nodes were defined as SLNs. Axillary lymph nodes whose consistency and dimension were considered suspicious were also removed and analyzed.

Pathologic examination of the SLN was macroscopic, cytologic and histologic.

The intraoperative cytology examination of the lymph nodes was performed by dissecting them in two parts along the major axis of the capsule if larger than 0.5 cm. After SLN division, a slide was affixed or dragged on the cut surface of both halves and stained with Harris hematoxylin solution.

In case of suspected cytology, lymph node halves were frozen to −22 ◦C, serially divided in ultrathin sections and stained with Harris hematoxylin solution.

For definitive pathologic assessment, SLN was included and examined along with two consecutive sections stained with Hematoxilyn and Eosin (HE) and, subsequently, with five sequences of three consecutive sections, 200 microns spaced. The middle section of each series was colored with CAM5.2, and those remaining with HE.

All non-sentinel nodes were examined with standard procedure, as mentioned for SLN intraoperative histologic assessment.

For patients with ypN+ (micro or micro-metastatic) disease, axillary dissection was directly performed. I and II level lymph nodes were always removed, while III level lymph nodes were removed only in case of intraoperative detection of clinically suspicious nodes at lower levels.

Patients with isolated tumor cells positivity at SLNB were treated as ypN0 and did not receive ALND.

Adjuvant treatments: were determined on the basis of patient's age, pre-neoadjuvant clinical staging, surgical intervention, pathological staging and tumoral biology.

Adjuvant chemotherapy: Patients who did not make a pathological complete response to neoadjuvant treatment were treated according to different adjuvant regimens.


Adjuvant radiotherapy: was tailored to the type of surgical intervention and pathological staging. Radiation was delivered using 3D conformal schemes and intensity modulated radiotherapy on linear accelerator using 6-10-15 MV photons.

Axillary radiation was considered for patients with pathologically positive lymph nodes and subsequent ALND with less than 10 nodes removed, ypN3 tumor staging, extracapsular invasion or isolated tumor cells (ITC) in SLNs.

#### **4. Statistical Analysis**

Statistical analysis was performed with SPSS version 26.0 for Windows. Results are expressed as mean, median and range. Fisher exact test was used for categorical variables. A *p* < 0.05 was considered statistically significant.

Kaplan-Meier curves were used to plot OS, DDFS and RDFS. Oncological outcomes were calculated over a median follow up of 24 months (2–48).

Only factors significantly associated with this outcome in the univariate analyses were included in multivariate models. Multivariate analyses were performed on all patients, and separately for cases cN0 and cN1/2 prior to neoadjuvant treatment.

#### **5. Results**

Between January 2016 and December 2019, 4478 patients with invasive breast cancer were treated in our multi-disciplinary center.

From our prospectively maintained database we extracted 412 patients with cT1 cT4 and cN0-cN1/2 diseases, who became or remained ycN0 at the end of neoadjuvant treatment and underwent surgical treatment.

We excluded from this study four cN0–ycN0 patients and nine cN+-ycN0 patients that underwent immediate ALND following a non-identification of a SLN during axillary procedure. The overall SLN identification rate was 96.8% (98.2% in cN0-ycN0 patients and 95.2% in cN+-ycN0 patients).

Regarding the remaining 399 cases that underwent SLNB, in 117 (29.4%) the main indication for NACT was to reduce the tumor diameter and achieve a conservative breast treatment instead of a mastectomy; 104 (26.0%) had the presence of clinically involved lymph nodes (cN+) and 76 cases (19.0%) both concomitant situations.

Furthermore, in 102 cases (25.6%) patients younger than 50 years with unfavorable risk factors received NACT mainly to ensure a prompt systemic treatment, independently of T/N status.

Among patients that underwent NACT, 219 patients that were cN0 at the time of diagnosis remained ycN0, while 180 cN1/2 patients benefit of chemotherapy and down staged to ycN0 status (patients characteristics prior to neoadjuvant chemotherapy, according to axillary clinical status before systemic treatment are summarized in Table 1).


**Table 1.** Clinical characteristics of 399 patients according to cN status prior to neoadjuvant treatment.

Clinical restaging after NACT: Neoadjuvant regimes and clinical response are summarized in Table 2. Concerning clinical response, we observed an overall complete clinical response in 144 patients (36.1%), a partial response in 228 patients (57.1%) no response in 12 patients (3%) and a progression to T stage, with breast skin or pectoralis major fascia involvement, in 15 patients (3.8%).

Among women treated with hormone-based NACT, we observed four cases of clinical complete response to treatment (16.7%), 15 cases of partial response (62.4%) and five cases of no response (20.9).


**Table 2.** Schemes of delivered neoadjuvant treatments according to axillary clinical stage at diagnosis and related clinical response.

Breast surgery: 246 patients received conservative OPS and 153 women were given mastectomy (134 patients (87.6%) were treated with conservative mastectomy followed by implant/autologous reconstruction).

Among elderly and fragile patients treated with hormone-based NACT, four patients (16.7%) with initial skin involvement did not experience any response to NACT and were consequently treated with a simple mastectomy (Table 3).

**Table 3.** Surgical treatment and adjuvant radiotherapy.


\* Patient's refusal or early progression of systemic disease.

Axillary treatment: During SLNB surgical procedure a mean number of 2.7 lymph nodes (3, 1–7) were identified and removed.

Overall, SLNB was negative in 235 cases (58.9%). Among 219 cN0-ycN0 patients, 149 (68%) were confirmed ypN0(sn). Of the 180 cN1/2-ycN0 cases, 86 (47.8%) confirmed an effective downstaging to an ypN0 (sn) status, 76 patients (42.2%) remained macrometastatic, 10 patients (5.6%) decreased to a micro-metastatic involvement and eight patients (4.4%) patient revealed a residual ITC positivity.

Among patients given ALND, a mean number of 12.8 lymph nodes (11.5, 6–30) were removed. Axillary pathological staging is summarized in Table 4.

Pathological Characteristics: Pathological characteristics are shown in Table 5. Complete breast remission (ypT0) occurred in 132 (33.1%) women. Tumors were luminal A-like, luminal B-like, HER2 positive and triple negative in 75 (18.8%), 121 (30.4%), 10 (2.5%) and 33 (8.2%) cases, respectively. Tumor subtype was not assessable in 160 (40.1%) patients with complete pathological remission in the breast or very limited residual disease.


**Table 4.** Pathological characteristics according to cN status prior to neoadjuvant treatment.

\* evidence of isolated cancer cells in the lymph node. \*\* evidence of microscopic residual of tumor (<0.2 mm) in the lymph node. \*\*\* ypN0, ypN1mic and ypNi+.


Oncological outcomes: After a median of 35.6 months (2–55), axillary failure (AF) occurred in 2 cN0 patients with a negative SLN and in 4 cN1/2 patients with a negative SLNB.

In our experience, AF occurred in patients that were diagnosed in unfavorable conditions, such as multifocal (two patients; 33%), cT3 (two patients; 33%) and T4b (one patient; 16.5%) tumors. Moreover, we observed an AF in 1 patient with ITC SLN positivity, who refused both ALND and adjuvant axillary radiotherapy.

Furthermore, in 3/6 patients, AF was diagnosed concurrently to a distant relapse (50%).


OS, DDFS and RDFS were significantly related to SLNB positivity, even overall, and according to axillary staging before NACT (cumulative incidence of regional relapses, as well as OS and DDFS curves, are shown in Figure 2).

At uni- and multivariate analysis (Table 5), positive SLN was confirmed as an independent prognostic factor for DDFS, as well as triple negative immunophenotype and T pathological complete response, even in the cN0-ycN0 and in the c1/2-ycN0 group.

**Figure 2.** Cumulative Overall Survival, Regional Recurrence and Distant Disease Free Survival. **Figure 2.** Cumulative Overall Survival, Regional Recurrence and Distant Disease Free Survival.

#### **6. Discussion**

In an effort to minimize the clinical impact of breast cancer, several improvements have been made with regards to breast surgical treatment following NACT [14–16].

Conversely, the axillary approach remains a controversial field: SLNB is considered the gold standard for axillary staging in early breast cancer patients with clinically negative lymph nodes, confining ALND to a very limited group of patients.

The purpose of this de-escalation in surgery is to reduce axillary morbidity (seroma formation, loss of arm sensitivity, shoulder dysfunction and lymphedema) by restricting or avoiding axillary dissection without proven oncological advantages.

However, to be reliable SLNB should always be over 90% in identification rate (IR) and below 10% in false negative rate (FNR), conditions that could be easily met in early breast cancer treatment, whereas after NACT initial experiences reported questionable results [17,18].

Despite these initial observations, a progressive set-up of the axillary workout before and after NACT led to metanalysis of retrospective studies that seem to validate SLNB after NACT, reporting acceptable IR and FNR, comparable to those reported for the early breast cancer setting [19–21].

Moreover, recent evidence also seemed to validate SLNB in ycN0 patients, for whom positivity would also play an important role as a significant prognostic factor [13].

We analyzed records regarding 399 consecutively treated patients. We achieved, even with a single agent technique, an acceptable IR for cN0-ycN0 and for cN1/2-ycN0 patients (98.2% and 95.2%, respectively).

These data are in line with previously published results in theoretically more favorable conditions, such as its execution in early breast cancer setting, and obtained with

the use of double tracer (radiotracer + blue dye), thus strengthening the observation of Fringuelli et al. [22] that NACT does not influence axillary lymphatic drainage and consequently axillary mapping success, furthermore clearly confirming that identification rate of SLNs actually improves with the surgical experience of the operating team, especially for single tracer technique, as reported by Zhang in the neoadjuvant setting [23].

Regarding prognostic power of SLNB after NACT, the European Institute of Oncology recently published a paper in which Galimberti et al. analyzed 396 cT1-4, cN0/1/2 patients who became or remained cN0 after neoadjuvant treatment and underwent SLNB.

Their data confirmed SLN status as a significant prognostic factor in cN0-ycN0 patients, a finding that seems to be consistent with the known prognostic significance of axillary involvement in the early breast cancer setting. However, at multivariate analysis, SLNB lost its prognostic power in the cN1/2-ycN0 group, suggesting that an axillary involvement before NACT could potentially jeopardize a reliable mini-invasive radiation.

Our data (although a result of limited and preliminary observations) confirm Galimberti's conclusions: in 219 cN0-ycN0 patients SLNB was safely performed. In this subgroup, we observed two cases of axillary failure, and three-year OS, DDFS and RDFS were 95.5%, 92.2% and 94.2%, respectively. Moreover, multivariate analysis showed that strong prognostic factors such as triple negative immunophenotype, persistence of extended involvement of the axilla (ypN2/3) and positivity of SLNB maintain a statistically comparable prognostic role.

In this setting, our observations further reinforce not only the feasibility but also the low risk of false negative rates of SLNB, confirming its predictive role even after NACT.

Among 180 cN1/2–ycN0 patients, a mini-invasive axillary staging by means of SLNB should be taken into account both for the high identification rate and also for the observed axillary complete pathological response rate (in our experience 47.8%).

In such a setting, although there is a higher rate or axillary failure (4.6% versus 1.3% in cN0-ycN0 patients), we registered three-year OS, DDFS and RDFS rates that were comparable to those registered in cN0 patients (93%, 84.8% and 87.9%, respectively).

We also confirmed SLNB's prognostic power at multivariate analysis that, even in a more complex subgroup of patients, resulted in statistical comparability to other strong prognostic factors such as triple negative immunophenotype, and complete pathological response on the breast (ypT0).

This observation differs from Galimberti's conclusions for cN1/2–ycN0 patients, a phenomenon that can be related to our shorter follow up timing, suggesting that in the first three years axillary response could reflect a systemic control of disease, but also to our high rate of patients diagnosed with cN1 axillary status before NACT (144/180 (80%) of cases included in cN1/2 group were cN1).

This particular subset distribution could have influenced uni- and multivariate results, suggesting that SLNB remains a reliable prognostic tool in patients with a lower grade of initial axillary involvement, whereas in patients with a major axillary burden prior to neoadjuvant treatment, the disruption of the lymphatic architecture (caused both by both perinodal infiltration and chemo-therapic agents) could compromise the reliability of SLN in predicting axillary status and therefore compromise its predictive prognostic power.

#### **7. Conclusions**

Our results, although from a single institution and being a retrospective experience with a limited follow-up timing, strengthen the possibility of safely performing SLNB after NACT in cN0-ycN0 patients, and reinforce the need for further refinements in the mini-invasive axillary approach for cN1-2 patients who become node-negative after NACT. **Author Contributions:** A.M.S., A.F. and A.M.—Designed study, analysis and interpretation of data, drafted paper and revised it critically, approved the submitted version; D.T., A.O., S.M., A.D.L., F.M. (Francesca Moschella), M.N., S.D., L.S., E.J.M., F.D.L., F.M. (Fabio Marazzi): Designed study, analysed and interpretation of data, drafted paper; R.M. and G.F.: Designed study, drafted paper and revised it critically, approved the submitted version. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board and Ethics Committee of Fondazione Policlinico Universitario Agostino Gemelli IRCCS; Università Cattolica del Sacro Cuore, Rome, Italy.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** Disclosure of financial interests and potential conflict of interest. All authors submitting this manuscript confirm and attest that they have no conflict of interest. There are no source of support in any form nor funding for this work. There are no financial relationships for this work.

#### **References**

