Omission of sentinel lymph node biopsy in early stage luminal breast cancer: impact on adjuvant CDK4/6 inhibitor recommendation and oncological outcomes.

1
Introduction
The SOUND (Sentinel Node versus Observation after Axillary Ultrasound) and INSEMA (Intergroup Sentinel Mamma) trials established that a sentinel lymph node biopsy (SLNB) can safely be omitted in many patients with small, clinically and imaging node-negative (cN0/iN0) breast cancer (BC) [1,2]. These results are particularly applicable to patients with hormone receptor (HR)-positive and human epidermal growth factor receptor 2 (HER2)-negative BC, who represented approximately 95 % of patients in both trials. Recruitment for the SOUND trial was completed in June 2017, and recruitment for the INSEMA trial ended in April 2019. Consequently, patients in both trials underwent adjuvant therapy before the approval of the first cyclin-dependent kinase 4/6 inhibitor (CDK4/6i) abemaciclib in the adjuvant setting in October 2021. Currently, two CDK4/6i, abemaciclib and ribociclib, are approved for the adjuvant treatment of stage II-III HR-positive, HER2-negative BC. These approvals are based on the findings of two trials, monarchE and NATALEE, which demonstrated oncological benefits in terms of invasive disease-free survival (iDFS) [3,4]. Notably, both trials included substantial proportions of patients with stage II disease (26.2 % in monarchE and 40.1 % in NATALEE). These patients equally benefited from adjuvant CDK4/6i treatment.
Eligibility criteria for abemaciclib and ribociclib differ, however some patients may be eligible for both treatments. Eligibility for abemaciclib requires at least one positive lymph node. Treatment with ribociclib is recommended in patients with stage II-III disease, including those with node-positivity. However, it may also be recommended in many node-negative patients. Importantly, it is currently unclear in how many patients the omission of a SLNB would result in missed eligibility for CDK4/6i treatment. No differences in adjuvant systemic treatment were seen in the SOUND trial, including patients with tumors up to 2 cm [1]. However, in the INSEMA trial, which included patients with tumors up to 5 cm, more patients received adjuvant chemotherapy after SLNB [2]. This raises concern about potential undertreatment if SLNB is omitted, particularly regarding CDK4/6i [5].
Consequently, we designed the present study to isolate how nodal staging contributes to adjuvant CDK4/6i recommendation in SOUND/INSEMA-eligible patients. Specifically, we address the question of how many SOUND/INSEMA-eligible patients would need to undergo a SLNB to identify one patient eligible for adjuvant CDK4/6i treatment. In addition, we investigated whether the omission of SLNB impacts the risk of recurrences.

2
Methods
2.1
Study design
This retrospective cohort study included consecutive patients with ≤cT2, node-negative (cN0/iN0), HR-positive, HER2-negative BC who underwent upfront final breast-conserving surgery and SLNB at the University Hospital of Basel, Switzerland, between January 2014 and December 2024. Data for this analysis was extracted on June 02, 2025.
All participants were identified from a prospectively maintained institutional database within the online good clinical practice conform data management system secuTrial©. The study was reported according to the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines [6].

2.2
Staging procedure
Tumor stage was reported according to the American Joint Committee on Cancer 8th Edition. Clinical tumor stage (cT) was determined by clinical examination and imaging including mammography and ultrasound. The clinical nodal stage was determined by clinical examination (cN) and ultrasound (iN) performed by dedicated and specially trained radiologists and senologists in all patients as standard of care over the study period. Typically, patients were referred after screening mammography with suspicious findings, or upon suspicious clinical examination. All patients completed staging with mammography and sonography of the breast and axilla, including core-needle biopsy of all suspicious lesions. Breast-MRI including axillary evaluation was performed after interdisciplinary discussion, typically in patients with dense breasts.

2.3
Surgical procedure
Sentinel lymph node biopsy was performed either before or after tumor removal. Standard incision was mid-axillary over 3–4 cm, inferior to the lower axillary hairline-border; less frequently, the same incision was used for tumor removal and SLNB. Sentinel lymph nodes (SLNs) were defined as either hot (technetium Tc99m), blue (blue dye), or a combination thereof. Use of dual tracer was recommended in earlier years of the study period. In recent years, Tc99m as single tracer without routine use of lymphoscintigraphy became institutional standard.

2.4
CDK4/6 inhibitor eligibility
Eligibility for adjuvant abemaciclib according to the original monarchE study protocol included patients with either ≥4 positive axillary lymph nodes, or 1–3 positive lymph nodes plus one additional criterion of grade 3 disease, tumor size ≥5 cm, or Ki67 ≥ 20 % [3].
Eligibility for adjuvant abemaciclib according to approval by the Food and Drug Administration (FDA) included patients with either ≥4 positive axillary lymph nodes, or 1–3 positive lymph nodes plus one additional criterion of grade 3 disease or tumor size ≥5 cm7.
Eligibility for adjuvant ribociclib was according to the original NATALEE study protocol and included patients with stage II-III, HR-positive, HER2-negative BC [4]. All patients with node-positivity but those with stage IB (T1N1mi) were considered eligible. Ribociclib eligibility based on node-criteria was defined as either pT1pN1 or pT2pN1, provided that no additional criteria would render the pT2 status per se eligible for ribociclib treatment (i.e., grade 3; or grade 2 plus Ki67 ≥ 20 % or Oncotype Dx RS of ≥26), irrespective of nodal status.

2.5
Number needed to treat based on recurrence estimates of adjuvant CDK4/6 inhibitors
The number needed to undergo SLNB to prevent one recurrence by use of an adjuvant CDK4/6i was estimated based on the number of patients with stage II disease needed to receive abemaciclib (NNT = 28, 5-year follow-up) or ribociclib (NNT = 63, 3-year follow-up, published), or ribociclib (NNT = 27, 5-year follow-up, reported) (Table 1) [3,4,8].
The necessary number of SLNB in patients meeting the SOUND/INSEMA eligibility criteria to prevent one iDFS event by use of an adjuvant CDK4/6i was estimated based on the extrapolated number of patients with stage II disease needed to receive abemaciclib (NNT = 28, 5-year follow-up) or ribociclib i) based on the published 3-year follow-up (NNT = 63), and ii) based on the reported 5-year follow-up (NNT = 27).

Ethical approval
This study was approved by the local ethics committee (Ethikkommission Nordwest-und Zentralschweiz, approval number 2016-01525) and written informed consent was obtained from all patients [6].

2.6
Aims
The primary aim of this study was to assess how many patients with BC meeting eligibility criteria for the omission of SLNB according to SOUND/INSEMA would have missed adjuvant CDK4/6i treatment recommendation due to positive SLNs. A secondary aim was to assess the impact of missed CDK4/6i eligibility after SLNB omission on iDFS.

2.7
Statistical analysis
Patient, tumor, and treatment characteristics were summarized using descriptive statistics. The results are presented as median values with interquartile ranges (IQR) for continuous variables, which were compared using Wilcoxon rank-sum test. Categorical variables are presented as absolute frequencies with percentages and compared between groups via Fisher's exact test. Multivariable logistic regression models were used to identify predictors of CDK4/6i eligibility and node-positivity. Logistic regression results were summarized using average marginal effects, representing the adjusted absolute risk difference (aARD, percentage-point change in probability) between comparison groups. All statistical analyses were performed using R (version 4.3.2).

3
Results
3.1
Patient characteristics
Of a total of 309 included patients, 78.3 % (242/309) were postmenopausal, 28.8 % (89/309) had a ≥pT2 tumor, and 14.6 % (45/309) had pN1 disease (Table 2).

3.2
CDK4/6 inhibitor eligibility
In total, 21.4 % (66/309) of patients qualified for adjuvant CDK4/6i treatment (Fig. 1). Eligibility for abemaciclib was present in 6.1 % (19/309, monarchE; Table 3), and 3.2 % (10/309, FDA approved). Eligibility for ribociclib was present in 20.4 % (63/309), with 7.8 % (24/309) being eligible based on node-criteria. Consequently, 17 (95 %CI 11–27), 31 (95 %CI 17–64), and 13 (95 %CI 9–20) patients would be required to undergo SLNB in order to identify one patient qualifying for abemaciclib (monarchE), abemaciclib (FDA approved), and ribociclib (node-criteria), respectively.
Patients found to be eligible for CDK4/6i treatment more frequently had a cT2 stage (54.6 % vs. 20.6 %, p < 0.001), and more aggressive tumor characteristics (i.e., Ki67 ≥ 20 % 60.6 % vs. 21.4 %; grade III 25.8 % vs. 8.2 %; and lymphovascular invasion 43.9 % vs. 10.7 %, Table 2). In patients eligible for CDK4/6i treatment, pathologic tumor stage (≥pT2 72.7 % vs. 16.9 %, p < 0.001) and pathologic nodal stage (pN1 59.1 % vs. 2.5 %, p < 0.001) were more advanced. Those found to be eligible for a CDK4/6i tended to have more lymph nodes removed (median 2 vs. 2, p = 0.069), which harbored a higher number of positive lymph nodes (median 1 vs. 0, p < 0.001).
On multivariable analysis pathological tumor stage ≥ pT2 (aARD 32.0 %, 95 %CI 19.1–44.9 %, p < 0.001), Ki67 ≥ 20 % (aARD 18.9 %, 95 %CI 9.4–28.5 %, p < 0.001), and lymphovascular invasion (aARD 19.5 %, 95 %CI 7.5–31.5 %, p < 0.001) were identified as predictors of CDK4/6i eligibility (Fig. 2). The only predictor of node-positivity was lymphovascular invasion (aARD 26.8 %, 95 %CI 11.9–41.7 %, p < 0.001).
A multivariable model considering only preoperatively available variables identified clinical tumor stage cT2 (aARD 22.3 %, 95 %CI 11.7–32.9 %, p < 0.001) and Ki67 ≥ 20 % (aARD 18.7 %, 95 %CI 6.8–30.6 %) as predictors of CDK4/6i eligibility (Fig. 3), Higher tumor grade showed a trend towards increased CDK4/6i eligibility (grade II vs. I aARD 11.4 %, 95 %CI 1.7–21.1 %; grade III vs. I aARD 16.4 %, 95 %CI -0.5 %–33.4 %, p = 0.054).

3.3
Impact of sentinel lymph node biopsy omission on invasive disease-free survival
The number of SLNB to prevent one iDFS event was estimated to be 456 (95 %CI 287–723), 866 (95 %CI 447–1677), 812 (95 %CI 541–1217), and 348 (95 %CI 232–522) for abemaciclib (monarchE), abemaciclib (FDA approved), ribociclib (3-year follow-up data), and ribociclib (5-year follow-up data), respectively.
Next, we estimated the impact of SLNB omission on iDFS events. SLNB omission according to INSEMA/SOUND eligibility criteria may result in an increase of 0.3 % (0.2–0.4), 0.2 % (95 %CI 0.1–0.3), 0.2 % (95 %CI 0.1–0.2), and 0.3 % (95 %CI 0.2–0.5) of recurrences, as adjuvant abemaciclib (monarchE), abemaciclib (FDA approved), ribociclib (3-year follow-up), or ribociclib (5-year follow-up) respectively, may not be recommended (Table 3). A sensitivity analysis was performed to depict the range of plausible treatment effects (Supplementary Table 1).

3.4
Subgroup analysis of patients with cT1 tumors
A subgroup analysis of patients with cT1 tumors (n = 223) showed that 13.0 % (29/223) of patients were node-positive, and 13.5 % (30/223) were eligible for treatment with a CDK4/6i. Eligibility for abemaciclib was 4.5 % (10/223, monarchE), and 2.2 % (5/223, FDA approved). Eligibility for ribociclib based on node-criteria was 8.1 % (18/223). Therefore, the number needed to undergo SLNB to identify CDK4/6i treatment candidacy was 23 (95 %CI 13–47), 45 (20–137), and 13 (95 %CI 9–21) for abemaciclib (monarchE), abemaciclib (FDA approved), and ribociclib, respectively. The number of SLNB necessary to prevent one iDFS event was estimated to be 625 (95 %CI 324–1206), 1249 (95 %CI 471–3314), 781 (95 %CI 488–1251), and 335 (95 %CI 209–536) for abemaciclib (monarchE), abemaciclib (FDA approved), ribociclib (3-year follow-up data), or ribociclib (5-year follow-up data), respectively. SLNB omission in this subgroup may result in an increase of 0.2 % (95 %CI 0.1–0.3 %), 0.1 % (95 %CI 0.1–0.2 %), 0.2 % (95 %CI 0.1–0.2 %), and 0.3 % (95 %CI 0.2–0.5 %) of iDFS events, based on missed treatment with abemaciclib (monarchE), abemaciclib (FDA approved), ribociclib (3-year follow-up data), or ribociclib (5-year follow-up) respectively.

4
Discussion
The SOUND and INSEMA trials have provided Level Ia evidence that the omission of a SLNB does not impact survival in mainly postmenopausal patients with small, HR-positive, HER2-negative BC [1,2]. However, both trials completed their patient recruitment prior to the approval of abemaciclib and ribociclib in the adjuvant setting. Both drugs have demonstrated superior oncological outcomes in the monarchE and NATALEE studies, respectively. Consequently, despite the similarity in adjuvant systemic treatment between the groups in the SOUND/INSEMA trials, the impact of SLNB omission on CDK4/6i eligibility remained to be elucidated. The present study enrolled patients who met the eligibility criteria for SOUND/INSEMA and underwent SLNB. The proportion of patients who met the criteria for adjuvant CDK4/6i treatment, as determined by nodal status, was found to range from 6.1 % to 7.8 %. Additionally, the impact of SLNB omission on oncological outcomes was estimated in this patient population based on published or reported NNTs from the monarchE and NATALEE trials [3,4,8]. The impact was found to be minimal, with a potential increase in iDFS events of 0.2–0.3 %. Therefore, SLNB omission should not be questioned to inform CDK4/6i indication.
The concept of SLNB omission is predicated on the recognition that SLNB carries a low but clinically relevant surgical morbidity rate of up to 10 %, which negatively impacts patient-reported outcomes [[9], [10], [11], [12], [13], [14]]. The pivotal trials that established SLNB as a standard procedure for patients with clinically node-negative disease yielded false-negative rates (FNR) of 10 %. Due to the very low rate of axillary recurrence, this FNR was subsequently deemed an acceptable threshold [[15], [16], [17]]. Although sonographic axillary evaluation is operator-dependent, highly specialized teams appear to approach the 10 % false-negative threshold in selected patients, as per SOUND/INSEMA eligibility criteria [[18], [19], [20], [21]]. Consequently, the potential morbidities and side effects of a surgical staging procedure must be evaluated against the backdrop of possible understaging and undertreatment, if oncologically pertinent. The present analysis demonstrates that fewer than 10 % of patients may be undertreated by omission of SLNB with respect to CDK4/6i treatment, with only a marginal impact on estimated oncological outcomes across all subgroups.
Omitting sentinel lymph node biopsy (SLNB) may lead to both radiotherapeutic overtreatment of the breast and undertreatment of regional nodal basins. While many SOUND/INSEMA-eligible patients might qualify for partial breast irradiation or even omission of radiotherapy, most in the SOUND trial (90 %) and all in INSEMA received whole-breast irradiation (WBI) [1,2]. The American Society of Clinical Oncology (ASCO) and the American Society of Breast Surgeons have addressed this, advocating for shared decision-making, multidisciplinary discussion, and against routine WBI after SLNB omission [22,23].
Compared to the cohorts of the SOUND and INSEMA trials, the cohort presented here differed in certain aspects. Patients in the present study more frequently exhibited cT2 (27.8 %) and ≥pT2 disease (28.8 %), compared to the cohorts of the SOUND (≥pT2 4.7 %) and INSEMA (cT2 9.6 %; ≥pT2 19.8 %) trials [1,2]. Grade III disease was present in 11.9 % of cases and Ki67 ≥ 20 % in 30.7 %, which aligns well with the results of the SOUND trial (17.9 % and 36.8)1. Notably, in the INSEMA trial, only 3.6 % of patients had grade III disease and 12.9 % a Ki67 ≥ 20 %, representing a potential selection bias [2]. Consequently, while more advanced tumor stages were eligible for the INSEMA study, patients with less aggressive biology were included. A subgroup analysis of patients with cT1 tumors confirmed the main results with a missed CDK4/6i eligibility between 4.5 and 8.1 % and a minimal impact on recurrence risk. Therefore, while SLNB omission may be most appropriate in women with cT1 disease according to the cohorts included in the SOUND and INSEMA trials, INSEMA reported the non-inferiority of SLNB omission across all investigated subgroups. Consequently, the cohort examined in this study may be more representative of patients in the real-world setting who opt to forgo any axillary surgery.
The rate of node-positivity was 14.6 % in the present cohort. These results are consistent with those of the INSEMA trial, in which 15.0 % had positive sentinel lymph nodes [2]. In patients with cT1 disease, node-positivity was 13.0 %, being in line with the 13.7 % reported in the SOUND trial [1]. The number of patients without indication for CDK4/6i treatment despite involved lymph nodes (i.e., micrometastatic disease without additional risk features) represented 2.3 % (7/309). The only common predictor for CDK4/6i eligibility and node-positivity was lymphovascular invasion, which is a well-known risk factor for node-positivity and a significant prognostic marker [[24], [25], [26]]. Presently, lymphovascular invasion serves as a relevant criterion for determining the indication for RNI [27,28]. In the context of SLNB omission, its value in relation to the decision on RNI should be explored in future studies. A recent study utilizing data from the US National Cancer Database revealed that the omission of SLNB in patients with cT1 cN0 BC (absence of information on iN-status) would affect adjuvant abemaciclib eligibility in 2.1 % and ribociclib eligibility based on node-criteria in 7.5 % of cases [29]. These results align with our estimates of 4.5 % and 8.1 % for missed abemaciclib and ribociclib eligibility, respectively. While the impact on premenopausal women was found to be stronger in the study by Wanis et al. [29], menopausal status was not found to predict CDK4/6i eligibility in the present study. To date, menopausal status is not included in any trial eligibility criteria for adjuvant CDK4/6i. However, the current guidelines, which have already adopted the results of the SOUND/INSEMA studies, advocate for SLNB omission exclusively in postmenopausal women ≥50 years [22,30]. Previously, the Choosing Wisely initiative had issued recommendations against the routine use of SLNB in patients aged 70 and above, which were however not followed in clinical practice [[31], [32], [33], [34], [35], [36]]. Subsequent real-world analyses will be necessary to evaluate whether the results of the SOUND/INSEMA trial do translate into clinical practice.
Abemaciclib and ribociclib have been demonstrated to improve oncological outcomes in the adjuvant setting also for patients with stage II disease [3,4,8]. However, the discussion points surrounding the indication for adjuvant CDK4/6i treatment in stage II disease encompasses the minimal absolute benefit, the physical and financial toxicities, and the inquiry into whether adjuvant CDK4/6i treatment may yield a diminished iDFS and overall survival impact in comparison to treatment in the metastatic setting [7,[37], [38], [39]].
Two ongoing prospective trials, namely BOOG 2013-08 (NCT02271828) [39] and NAUTILUS (NCT04303715) [40], are investigating similar cohorts of patients as SOUND and INSEMA. Prospective evidence of potential differences in CDK4/6i treatment is expected, especially from the NAUTILUS study, which opened in September 2020 and therefore just shortly before the approval of abemaciclib. In consideration of the results that have been presented, it is anticipated that missed CDK4/6i candidacy will minimally impact oncological endpoints in these two studies.
4.1
Limitations
The main limitation of this study is its retrospective single-center observational design. While 5-year follow-up data from monarchE has been published, point estimates for patients with stage II disease and 5-year follow-up from NATALEE were only recently reported orally [8]. Therefore, both 3-year and 5-year follow-up data are presented for adjuvant ribociclib. Furthermore, we analyzed CDK4/6i indication, which is different from the clinically relevant information on treatment initiation and adherence. However, this is the first study to investigate missed CDK4/6i indication due to SLNB omission and its potential impact on recurrence in patients eligible for SLNB omission. A strength is the routine use of axillary ultrasound throughout the study period, yielding included patients correctly eligible for the SOUND and INSEMA protocols.

5
Conclusions
According to SOUND/INSEMA trial protocols, fewer than 8 % of patients may not receive a recommendation for adjuvant treatment with a CDK4/6i due to missing SLN information. The impact on oncological outcomes is estimated to be minimal. Therefore, SLNB should not routinely be performed in patients who meet SOUND/INSEMA criteria to inform CDK4/6i treatment recommendations.

CRediT authorship contribution statement
Martin Heidinger: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization. Tibor A. Zwimpfer: Writing – review & editing, Validation, Investigation. Florian S. Halbeisen: Writing – review & editing, Formal analysis. Nadia Maggi: Writing – review & editing, Validation, Data curation. Marie Louise Frevert: Writing – review & editing, Validation, Data curation. Rama Kiblawi: Writing – review & editing, Validation, Data curation. Julie M. Loesch: Writing – review & editing, Validation, Data curation. Fabienne D. Schwab: Writing – review & editing, Validation, Data curation. Christian Kurzeder: Writing – review & editing, Validation. Giacomo Montagna: Writing – review & editing, Validation, Supervision, Investigation, Data curation. Walter P. Weber: Writing – review & editing, Validation, Supervision, Resources, Project administration, Investigation, Funding acquisition, Data curation, Conceptualization.

Ethical approval
This research project complies with the guidelines for human studies according to the Declaration of Helsinki as revised in 2013. The study protocol was approved by the Ethics Committee of Northwest and Central Switzerland (Ethikkommission Nordwest und Zentralschweiz, EKNZ; Approval number 2016-01525), and all subjects have given their written informed consent.

Data statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Sources of funding and support
This work was funded by the Department of Surgery, University Hospital of Basel and 10.13039/501100009078Hand in Hand Anstalt.

Role of the funding source
None of the funders had any role in the design of the study, in the collection, analysis, and interpretation of the data, and in writing the manuscript.

Conflicts of interest
Tibor A. Zwimpfer reports personal consulting fees from AbbVie that are outside the submitted work.
Christian Kurzeder reports honoraria from Tesaro, GSK, Astra Zeneca, Novartis, PharmaMar, Genomic Health, Roche, Eli Lilly S.A., Pfizer, Daichi; consulting or advisory role for Tesaro, GSK, Astra Zeneca, Novartis, PharmaMar, Genomic Health, Roche, Eli Lilly S.A., Merck MSD, Pfizer, and travel, accommodations and expenses from GSK, Astra Zeneca, and Roche.
Walter P. Weber received research support from Agendia paid to the 10.13039/100016015University Hospital Basel (last tranche in January 2022) for the TAXIS study (OPBC-03, SAKK 23/16, IBCSG 57-18, ABCSG-53, GBG 101). Payment or honoraria for lectures: MSD (for Best of SGBCC in April 2023).
All other authors report no conflicts of interest.