Evaluating the Clinical Utility of Routine Sentinel Lymph Node Biopsy and the Value of Adjuvant Chemotherapy in Elderly Patients Diagnosed With Oestrogen Receptor Positive, Clinically Node Negative Breast Cancer

Background: Sentinel lymph node biopsy (SLNB) provides staging information and guides adjuvant therapy in early breast cancer (EBC). Routine SLNB in oncogeriatricians with low-risk EBC remains controversial. Aims: To evaluate axillary management in elderly patients diagnosed with oestrogen receptor positive (ER+), clinically lymph node negative (cLN−) EBC, and to assess whether SLNB affects further axillary management or adjuvant chemotherapy (ACTX) decision making. Methods: Female patients aged > 65 years, diagnosed with ER+, human epidermal growth factor receptor-2 negative (HER2−), and cLN− breast cancer (BC), who underwent surgery and SLNB were included. Clinicopathological predictors of ACTX and completion axillary lymph node dissection (CALND) were determined. Kaplan-Meier analyses assessed survival outcomes. Results: A total of 253 patients were included (median age: 72 years, range: 66-90), all underwent SLNB; 50 (19.8%) had lymphatic metastasis on SLNB (SLNB+). Of these, 19 proceeded to CALND (38.0%), 10 (52.6%) of whom had further axillary disease (ALND+). 20 of the 50 SLNB+ patients received ACTX (40.0%) as did 31 of the 203 SLNB− patients (15.2%) (P < .001). Oncotype DX (ODX) testing was utilized in 82 cases (32.8%). Younger age (P < .001), SLNB+ (P < .001) and ODX score (P = .003) were all associated with ACTX prescription. ODX > 25 (OR: 4.37, 95% CI: 1.38-13.80, P = .012) independently predicted receiving ACTX. Receiving ACTX and proceeding to CALND did not improve disease-free (P = .485 and P = .345) or overall survival (P = .981 and P = .646). Conclusions: Routine SNLB may not be necessary in elderly patients diagnosed with ER+, cLN− EBC. Future oncogeriatric practice is likely to see genomic testing guiding ACTX prescription in this group.


Introduction
Breast cancer (BC) is the second most common cancer in women worldwide with one-third of BCs occurring in patients above the age of 65. 1 Historically, surgical management of BC involved extensive resection with radical mastectomy and axillary lymph node dissection (ALND) with considerable associated morbidity. 2 However, the majority of women now present with early-stage disease (ie, early BC [EBC]), and surgical practice has evolved such that tumour resection via either breast conserving surgery (BCS) or mastectomy with simultaneous sentinel lymph node biopsy (SLNB) for patients with clinically lymph node negative (cLN−) disease is now performed routinely without compromising oncologic outcome. 3 This paradigm shift towards increasingly conservative surgical approaches to BC has been driven by a desire to minimize morbidity, while maintaining optimal oncologic outcomes. 4 In the case of the axilla, the development of SLNB facilitated accurate minimally invasive nodal staging to inform therapeutic decision-making based on the presence or absence of axillary metastases. 5 The accuracy of SLNB as a staging tool was supported by evidence from the National Surgical Adjuvant Bowel and Breast Project (NSABP) B-32 study which demonstrated that patients with a negative SLNB (SLNB−) could be spared ALND. 6 Further evolution towards conservative axillary management came about following reports from the American College of Surgeons Oncology Group (ACOSOG) Z0011 trial which demonstrated that ALND does not improve survival or local control in BC with less than 3 positive sentinel lymph nodes (LNs). 7,8 Furthermore, although traditional clinicopathologic features such as tumour size and axillary nodal disease were the primary drivers of adjuvant therapy decision 2 Breast Cancer: Basic and Clinical Research making, the importance of tumour biology has gained increasing relevance in this regard over the last decade. The NSABP B-14, NSABP B-20, Southwest Oncology Group-8814 (or SWOG-8814), and Trial Assigning Individualized Options for Treatment (TAILORx) studies have demonstrated the potential to deliver rationalized adjuvant chemoendocrine prescription based on recurrence scores calculated from genomic assays 5,[9][10][11][12] This has led to genomic assays, such as Oncotype DX (ODX) Recurrence Score (RS) (Genomic Health Inc., Redwood City, California) and MammaPrint (Agendia Precision Oncology, Amsterdam, The Netherlands), being incorporated as a component of modern BC management in ER+ cases.
Contemporary BC surgical management has been informed by evidence from well-designed, prospective randomized studies; however, most trials have focused their inclusion criteria on younger participants, possibly due to their greater potential survival benefit as well as their overall suitability for inclusion based on their functional status. [13][14][15] This selection bias has led to consistent underrepresentation of elderly patients in clinical trials and a less robust evidence base regarding oncological best practice for these patients. The 'Choosing Wisely' campaign was established in 2012 in the United States in an attempt to address this discrepancy in treatment and the associated vulnerability for geriatric patients. 16 Data published from 'Choosing Wisely' suggests a more refined approach to oncogeriatric patient management in those with ER+, HER2−, cLN− BC may be beneficial where possible. 17 Although routine SLNB has revolutionized surgical management of the axilla for the vast majority of EBC patients, it is not without associated morbidity, and a shift to an even more conservative approach may now be considered for older patients if SLNB as an axillary staging investigation is not providing information that will inform/change further axillary or systemic treatment. This is likely to be the case considering the fact that the primary driver of adjuvant systemic therapy in contemporary BC management is tumour biology. 18,19 Considering the changes in practice in the US following the Society of Surgical Oncology (SSO) Choosing Wisely recommendations, 17

Breast cancer diagnosis and work up
All patients underwent triple assessment as part of their EBC workup. Clinical examination was performed by a consultant breast surgeon. Standard radiological assessment consisted of mammography and ultrasound of the breast and ipsilateral axilla. Core biopsy was performed for clinically or radiologically suspicious lesions. All biopsied breast tissue was assessed by a consultant histopathologist with expertise in breast pathology. All breast tissue specimens were analysed in the accredited pathology laboratory at the tertiary referral centre. Staging was performed as per the American Joint Committee on Cancer (AJCC), version 8 Guidelines (2017). 21 Histopathological or immunohistochemical tumour evaluation and staging ER and progesterone receptor (PgR) status were determined in accordance to American Society of Clinical Oncology/College of American Pathology guidelines and reported using Allred scoring. [22][23][24] HER2 status was analysed using immunohistochemical analyses, 25 and fluorescence in-situ hybridization (or FISH) was requested in the case of equivocal 2+ results. 26 Tumour grade was determined using the Elston Ellis modification of the Scarff-Bloom-Richardson grading system. 27,28 Lymphatic invasion was assessed using IHC staining with D2-40 and vascular invasion using CD34. 29,30 Perineural invasion was determined using IHC staining with S-100 and a broad-spectrum keratin stain (AE1/AE3). 31 MIB1 antibody testing was used to Ki-67 proliferation. 32

Multidisciplinary care
All cases were discussed at a BC multidisciplinary meeting attended by consultant breast surgeons, histopathologists, radiologist, medical oncologists, and radiation oncologists. Decisions regarding patient specific treatment are determined from clinical, radiological, and pathological factors, as well as each patient's performance status in line with international guidelines and best practice. Adjuvant prescription of chemoendocrine therapies for some patients diagnosed with ER+, HER2−, cLN− after 2007 were guided by ODX genomic testing following decisions allowing public reimbursement for the 3 assay in oncological practice by the Irish government. 33 ODX testing was carried out in Genomic Health Inc, Redwood City, CA, USA. 34 Neoadjuvant treatment regimens, primary oncological surgeries and adjuvant treatment regimens were carried out in the tertiary referral centre. Patients returned to the tertiary referral centre for examination by a specialized breast surgeon postoperatively and returned for annual clinical and mammographic surveillance after BC treatment.

Patient follow up and clinical outcomes
Individual patient follow-up and survival data was recorded through a prospectively maintained database. All data was cross-referenced with patient electronic and medical records. Patient mortality was confirmed from data obtained from National Registries. The median and mean lengths of followup were assessed using the reverse Kaplan-Meier method. 35 Definition of DFS was 'freedom from invasive disease recurrence, a second primary cancer or death'. Patients with ⩾ 3 LNs positive for disease following SLNB, were considered for CALND, as described in the ASOCOG Z0011 study. 7

Statistical analysis
Details regarding treatment characteristics within each age group were determined and descriptive statistics were used to inform clinicopathological associations with treatment characteristics (ACTX or CALND). Binary logistic regression analysis was used to ascertain patient and tumour characteristics predicting being in receipt of ACTX or CALND expressed in crude odds ratios (OR) with 95% confidence intervals (CIs). Variables with P < .050 in univariable analysis were included in the multivariable logistic regression analysis, which was then used to identify variables that contributed independently to being spared ACTX or CALND. Kaplan-Meier curves and the Log-rank (Mantel-Cox) test were used to associate survival with treatment characteristics. All tests of significance were 2-tailed, with P < .050 indicating statistical significance. Data were analysed using Statistical Package for Social Sciences (SPSS) version 26. All patients were diagnosed with ER+, HER2−, cLN− primary BC, 121 of which were T1 (47.8%) and 132 of whom had T2 disease (52.2%). The majority had grade 2 tumours (80.6%, 204/253) and 203 had invasive ductal carcinoma histological subtype (80.2%). All tumours were ER+ (100.0%), and the mean ER score was 7.8 ± 0.1 (range: 3-8, median: 8). Overall, 214 tumours had positive PgR expression (84.6%), and the mean PgR score was 5.6 ± 2.1 (range: 0-8, median: 7). Clinicopathological data are outlined in Table 1.

Discussion
Traditionally, histopathological features such as tumour size and SLNB+ were predominant factors in guiding decisionmaking with regard to systemic therapies and the requirement for aggressive locoregional treatment through ALND. However, the majority of elderly patients develop tumours of favourable biology, that is, ER+ve. 36,37 In the era of personalized medicine, genomic analysis aims to tailor multimodal therapeutic strategies on the basis of case-specific clinicopathological and molecular properties, instead of standard histopathologic and immunohistochemical assessment. 5,[9][10][11][12]38 There is also a growing recognition that elderly patients are frequently overlooked as participants in oncological research studies, indirectly leaving this cohort vulnerable to over-investigation and overtreatment. Thus, recent studies have proposed 'de-escalation' of local therapies and omission of SLNB in elderly patients with EBC, in particular those with favourable prognosis, such as ER+ disease. 39,40 The most important finding in this series of elderly patients diagnosed with ER+ EBC is that in spite of SLNB being clinically indicated in all 253 cases, decision making relating to ACTX was more likely to be guided by genomic testing (ODX) than histopathological results of SLNB. Being aged < 75 at diagnosis, having SLNB+, and ODX score > 25 were all factors predictive of ACTX prescription, however, ODX score > 25 was the sole independent predictor following multivariable analysis. These findings are supported by the results of other large volume studies, 41,42 and are reflected in the modern multimodal treatment paradigm for early ER+ disease, where clinical decision-making regarding tailoring of chemoendocrine therapeutic strategies is primarily determined by genomic tissue analyses. 18,19 Consequently, the estimated risk of distant recurrence, BC-specific mortality and resulting benefit from ACTX is individualized to the patient and their specific tumour biology, with considerations surrounding functional status evaluated by the multidisciplinary team.
Other results of this univariable analysis are somewhat predictable: Patients aged > 75 years at diagnosis were less likely to undergo ACTX than their counterparts, while those with SLNB+ were likely to be prescribed systemic chemotherapy ( Table 7). The loss of physical functionality and resultant disability makes elderly patients less likely to tolerate the toxicities associated with ACTX: data from Given and colleagues suggests that up to 60% of patients aged 65 to 74 years report some form physical disability and unsuitability for systemic therapies. 43 These results validate the rationale for prudent use of ACTX in the elderly. Histopathological evaluation of the SLNB has traditionally been one of the catalysts for guiding adjuvant systemic therapies, and therefore it is unsurprising that SLNB+ predicted progression to ACTX. With improvements in clinical staging diagnostic modalities, and the increasing availability and affordability of genomic testing, eliminating the routine requirement for invasive axillary sampling in lowrisk, elderly BC patients is mooted in modern practice. 44 The primary reason for performing SLNB is to allow for accurate staging and to determine the requirement for ACTX and CALND, however survival outcomes following these interventions after SLNB in elderly patients is less well described. In this series, receipt of ACTX failed to improve DFS or OS for elderly patients with ER+ BC, irrespective of  nodal status. The authors acknowledge the possibility of a Type II error due to the small study population and few event rates, as is anticipated with EBC, as well as the increase in competing risks for death in older patients. However, this also highlights the challenges faced in optimizing therapy to maximize life expectancy and the quality of life, while minimizing morbidity for this patient group. Data suggests that negative implications of ACTX prescription often outweigh the perceived benefits in elderly and comorbid patients, 45 with only 15% of patients with cLN− examinations aged > 70 years experienced cancerrelated mortality. These statistics demonstrate the difficulty in achieving a survival benefit from prescription of toxic chemotherapy in this group, without risking iatrogenic toxicities or organ failure. 46 Consequently, the authors propose judicious use of ACTX in elderly patients with EBC, with prescription stratified through individualized risk following genomic analysis, obviating the need for SLNB to guide therapeutic decisionmaking through axillary staging, in this subset of patients.
As previously described, breakthrough trials have reduced indications for surgical resection through CALND in EBC, [6][7][8] and modern studies suggest exemption in elderly patients, given the associated morbidity, minimal survival benefit and oncological control following complete clearance of the axilla. 44,47 This recent vogue is further supported by randomized data illustrating no added survival advantage from performing CALND in patients aged > 70 years with clinically occult axillary lymph nodes. 48 Eighty percent of patients in this study had node negative disease, however, for those with node positive disease, proceeding to CALND failed to enhance locoregional control or survival. These results, in tandem with the clinical utility of genomic testing in predicting ACTX prescription, suggest routine SLNB may be unnecessary within this elderly cohort. Focusing on reducing invasive initial diagnostic staging may prove more beneficial in reducing the harmful effects of over-investigation and overtreatment; sonographic axillary staging may be non-inferior to SLNB, with regard to excluding clinically significant metastatic disease in the axilla and requirement for CALND. 49,50 Thus, less-invasive staging and management of the axilla may reduce morbidity without any compromise in oncologic outcomes for the elderly, with the further possibility of additional cost-savings over current management strategies. [51][52][53]   This study suffers the inherent limitations of being a single centre, retrospective cohort study implying potential selection, ascertainment and confounding bias. Competing risks affecting survival within the elderly cohort as well as the relatively small numbers powering this study may limit conclusions which can be drawn in relation to survival. Furthermore, the failure to derive clinicopathological predictors of patients undergoing CALND is likely explained by a Type II statistical error due to a limited number of participants proceeding to axillary clearance. However, the data reflects real world practice in a multidisciplinary setting and so the results are relevant to clinicians involved in contemporary BC management.

Conclusions
Sentinel lymph node biopsy seems less valuable in guiding axillary surgical management and prescription of ACTX than previously perceived in those aged >65 years with ER+ EBC. Genomic testing offers a personalized and less-invasive means of guiding ACTX prescription in this group, and this may better serve this cohort that are at particular risk of over-investigation and treatment. Moreover, CALND and ACTX prescription indicated through SLNB failed to improve survival outcomes for patients in this series, irrespective of nodal status. These findings suggest the omission of SLNB may not be detrimental to clinical outcomes for elderly patients diagnosed with EBC of favourable tumour biology.

Author Contributions
Study design was performed by AJL. DB and MGD collected data included in this analysis. MGD and EJR performed statistical analysis and drafting of the manuscript. KM and PFM provided corrections. MJK and AJL reviewed manuscript drafts.