Changing Practices in Axillary Surgery After Neoadjuvant Breast Cancer Therapy: Insights From the EUSOMA European Database.

1
Introduction
Over the past few years, the increasing use of neoadjuvant treatment (NAT) and the introduction of targeted therapy have had a substantial impact on increasing pathological complete response (pCR) rates in breast cancer, ranging from 21% to 75%, depending on tumor subtype [1, 2]. This has enabled less invasive axillary staging techniques such as sentinel lymph node biopsy (SLNB) in clinically node‐negative (cN0) patients and also in selected clinically node‐positive (cN1) patients after NAT, with low false negative rates (FNRs) between 7% and 10%, respectively [3, 4, 5, 6, 7]. Importantly, a substantial proportion of initially node‐positive patients, approximately 13%–60%, depending on tumor subtype, achieve axillary pCR after NAT, further supporting the feasibility of de‐escalated axillary surgery in this setting [8].
Early clinical trials, including SENTINA, ACOSOG Z1071, and FNAC [9, 10, 11] established the feasibility of SLNB after NAT; however, they also highlighted persistently elevated FNRs as a significant limitation. To overcome this challenge, the technique of targeted axillary dissection (TAD) was introduced, which entails the excision of a previously biopsied and marked metastatic lymph node in addition to the sentinel lymph nodes. This approach has demonstrated high identification rates (96.7%–100%) and substantially reduced FNRs (4.2%–7%), thereby improving the accuracy of post‐NAT axillary staging [12, 13, 14, 15].
The oncologic outcomes associated with SLNB and TAD in patients who convert to node‐negative status after NAT have been consistently favorable, with axillary recurrence rates between 0% and 1.8% and disease free survival rates (DFS) of 92.6%–94.2%, which are comparable to those achieved with axillary lymph node dissection (ALND) [16, 17, 18, 19, 20]. These data have provided evidence supporting the de‐escalation of axillary surgery in appropriately selected patients.
Recent population‐level data from Europe and the United States further corroborate this shift, demonstrating a significant decrease in ALND rates and an increase adoption of SLNB (48%–62% in the United States and 46% in Europe), particularly among patients with favorable tumor subtype or excellent responses to NAT [21, 22].
Accordingly, this study aims to examine temporal trends in axillary surgery de‐escalation among breast cancer patients treated with NAT across a large European cohort, thereby providing contemporary insights into evolving surgical practice patterns.

2
Materials and Methods
2.1
Data Sources
The EUSOMA database is a central data warehouse of prospectively collected information, including pseudonymized individual patient records on primary BC cases diagnosed and treated in European breast centres that undergo the voluntary certification process and wish to participate in benchmarking and research activities. The EUSOMA database includes a set of quality indicators incorporated by the breast centres to allow standardized auditing and quality assurance. The database was started in 2006 and includes over 200,000 patients. For the present study data were analysed for the period 2010–2021 from 68 European breast centres (Belgium = 8; The Netherlands = 2; Sweden = 1; Austria = 2; Switzerland = 5; Germany = 16; France = 1; Portugal = 3; Italy = 26; Poland =1; Spain = 1; Croatia = 1; Cyprus = 1).

2.2
Inclusion Criteria
We have included patients with newly diagnosed breast cancer cT1‐T4 cN0‐N1 receiving neoadjuvant treatment, either chemotherapy or endocrine therapy. Those regimens varied according to each institution protocols. The indication for type of surgery was based on the discretion of the treating surgeon. The clinical and imaging response to chemotherapy, tumor biology, extent of disease at presentation and after chemotherapy, and patient age were all considered and a decision was made on a case‐by‐case basis. Tumor biologic subtype was categorized based on estrogen/progesterone receptor (HR) and human epidermal growth factor receptor (HER2) status.

2.3
Statistical Analysis
We have considered for the statistical study patients who received SLNB as the first treatment, independently if they received a final ALND due to nodal involvement. The comparisons of proportions in the two periods were performed with chi‐square test for categorical variables, t‐Student's for age. Multivariate analysis was performed to find determinants of pCR. Data were analyzed using R 4.3.2. All tests were two‐sided with an alpha‐level of 0.05.

3
Results
A total of 17,321 patients met the inclusion criteria and were included in the analysis. The mean age at diagnosis was 52.8 years (range: 24–86). Most patients presented with T2 tumors (54.7%), were clinically node‐negative (cN0, 53.3%), and received chemotherapy NAT (91%). Baseline characteristics are described in Table 1.
The overall pCR rate (for both breast and axilla in the cN+ setting) was 1,174 (14.5%). pCR rates varied significantly by tumor subtype: 36.8% in HR − /HER2+, 17.7% in HR+/HER2+, 22.0% in triple‐negative breast cancer (TNBC), and 4.3% in HR+/HER2− tumors (p < 0.001).
A total of 3,554 patients (20.7%) achieved a breast pCR, distributed as follows: 54.6% in HR − /HER2+, 27.9% in HR + /HER2+, 33.4% in TNBC, and 6.5% in HR + /HER2− tumors (p < 0.001).
Axillary pCR was observed in 3,487 patients (43.3%) among initially cN1 patients, with subtype‐specific rates of 77.6% in HR − /HER2+, 61.3% in HR+/HER2+, 57.8% in TNBC, and 24.8% in HR+/HER2− (p < 0.001).
3.1
Clinically Node‐Negative (cN0) Patients
Among cN0 patients, 7,086 (76.8%) underwent SLNB, 1,339 (14.6%) underwent SLNB + ALND, and 801 (8.6%) underwent ALND. During 2010–2015, 1,927 (65%) patients underwent SLNB, 614 (21%) underwent SLNB + ALND, and 412 (14%) underwent ALND. During 2016–2021, the corresponding numbers were 5,159 (82%), 725 (12%), and 389 (6%), respectively (Table 2).
When analyzed by the first axillary procedure, during the first period, 2541 (86%) patients underwent SLNB as the initial axillary surgery, compared to 412 (14%) who had a primary ALND. In the second period, 5884 (94%) underwent SLNB as the initial procedure, compared to 389 (6%) with primary ALND. This represents a significant increase in the use of SLNB over time (p < 0.001, Figure 1).

3.2
Clinically Node‐Positive (cN + ) Patients
In the cN+ group, 1,534 (18.9%) underwent SLNB, 1,324 (16.5%) underwent SLNB + ALND, and 5,237 (64.7%) underwent ALND. During 2010–2015, 239 (10%) patients underwent SLNB, 383 (15%) underwent SLNB + ALND, and 1,862 (75%) underwent ALND. During 2016–2021, these proportions shifted to 1,295 (23%), 941 (17%), and 3,375 (60%), respectively (Table 2).
Regarding the initial axillary procedure, 622 (25%) patients in the first period underwent SLNB as the first axillary surgery, compared to 1862 (75%) who had a direct ALND. In the second period, 2,236 (40%) underwent SLNB first, compared to 3375 (60%) with primary ALND, indicating a significant increase in SLNB utilization among cN+ patients (p < 0.001, Figure 1).

3.3
Subtypes and Predictive Factors
Regarding tumor subtypes, there is a statistically significant decrease in the rates of ALND in all molecular subtypes (p < 0.001), nevertheless, the increase in women having a SLNB as the first treatment was greater in the HER2‐enriched (OR 3.53, p < 0.001) and triple negative tumors (OR 2.84, p < 0.001).
A statistically significant decrease in ALND rates was observed across all molecular subtypes (p < 0.001). The increase in SLNB as the first axillary procedure was most pronounced in HER2‐enriched (OR = 3.53, p < 0.001) and triple‐negative subtypes (OR = 2.84, p < 0.001).
In the multivariate analysis of the cN1 subgroup, factors independently associated with increased pCR (and consequently higher SLNB use) were treatment period (2016–2021 vs. 2010–2015) OR = 1.37 (95% CI 1.12–1.67, p = 0.02), neoadjuvant chemotherapy OR = 3.44 (95% CI 2.08–5.88, p < 0.001), breast conserving surgery OR = 1.63 (95% CI 1.4–2.15, p < 0.001) and tumor subtype: HR + /HER2 + OR = 4.5 (95% CI 3.54–5.72, p < 0.001), HR‐/HER2 + OR = 12.61 (95% CI 9.84–16.15, p < 0.001) and TNBC OR = 5.49 (95% CI 4.3–7, p < 0.001).
When analyzed yearly rather than in two time periods, a consistent and statistically significant increase in SLNB rates was observed throughout the study period (p < 0.001).

4
Discussion
This study is the first at the European level to include a large and representative population from various countries. It shows a significant increase in the use of SLNB after NAT in Europe, in both cN0 and cN+ breast cancer patients, together with a corresponding significant decrease in ALND rates. The increase in less extensive axillary surgery has been more pronounced in the later years of the study period, probably reflecting changes in clinical practice after the publication of randomized trials on SLN after neoadjuvant treatment in cN+ breast cancer [9, 10, 11] and the inclusion into the 2015 National Comprehensive Cancer Network (NCCN) guidelines, which permitted SLN in selected cN1 patients [23].
Data show an overall axillary pCR rate of 46%, rising to 71% in non‐luminal HER2+ tumors, consistent with other reports. For example, Nguyen et al. [21] reported an overall axillary pCR of 46% and 76% in non‐luminal HER2+ tumors, while the ACOSG Z1071 trial reported an overall axillary pCR rates of 41.1% and 64.7%, respectively, in HER2+ tumors [24]. Recent trial data indicate that pCR rates following NAT have improved across all breast cancer subtypes, driven by the addition of immunotherapy and platinum‐based agents, as well as better patient stratification. In the KEYNOTE‐756 trial, adding pembrolizumab to chemotherapy in hormone receptor‐positive early breast cancer increased pCR from 15.6% with chemotherapy alone to 24.3% (treatment difference 8.5%, 95% CI 4.2–12.8; p = 0.00005) [25]. Additional improvements are seen in HER2‐positive breast cancer: a recent retrospective study reported pCR rates of 58.9% in HR‐/HER2+ tumours, highlighting the efficacy of dual anti‐HER2 blockade with trastuzumab and pertuzumab combined with chemotherapy [26]. In triple negative breast cancer (TNBC), the phase II BCT1902/IBCSG 61–20 Neo‐N study which combine nivolumab with non‐anthracycline carboplatin plus paclitaxel achieved an overall pCR rate of 53%–55% (90% CI, 44%–61%) [27]. Moreover, adding immunotherapy to chemotherapy in TNBC approximately doubles the odds of achieving pCR compared with chemotherapy alone (meta‐analysis OR 1.95; 95% CI 1.27–2.99) [28].
4.1
cN0 Patients
SLN is considered the standard of care for cN0 patients undergoing either upfront surgery or following NAT [9, 29]. The increasing adoption of SLNB as the initial axillary treatment in our study cohort aligns with findings from the Netherlands Cancer Registry, which reported SLNB use in up to 94% of 12,461 cN0 patients [22]. Similarly, Xiang et al. analyzed 18,548 cN0 patients from the SEER database and observed an increase in SLNB utilization from 76.4% in 2010 to 93.8% in 2021 [30]. Wong et al. reported favorable outcomes in 244 patients with cN0/cN1‐2 tumors. Among those with cN0 disease and ypN0 pathology, the 5‐year local and regional recurrence rates were 5.7% and 1%, respectively [31]. A national Italian survey of 21,949 invasive pN0 cases also showed a significant rise (p  <  0.001) in SLNB as the sole axillary staging procedure, with over 90% of patients receiving SLNB only by 2015 [32].
Ongoing trials are assessing the safety of omitting sentinel lymph node biopsy (SLNB) in selected patients achieving breast pathological complete response (pCR). The EUBREAST‐01 trial (NCT04101851) evaluates axillary surgery omission in cN0 triple‐negative or HER2‐positive patients undergoing breast‐conserving surgery (BCS) with pCR in the lumpectomy specimen, with axillary recurrence‐free survival as the primary endpoint [33]. The Dutch ASICS trial (NCT04225858) similarly investigates SLNB omission in clinically node‐negative HER2‐positive or triple‐negative patients showing complete radiologic response on MRI after neoadjuvant therapy, aiming to confirm safety based on 5‐year axillary recurrence rates [34].

4.2
cN1 Patients
Our study demonstrates a significant increase in the use of SLNB after NAT in cN+ patients, with a corresponding decrease in ALND. This trend has been particularly notable after the publication of feasibility studies of performing SLNB in selected cN+ cases with no residual axillary disease after NAT [9, 10, 11, 12, 13] and randomized trials supporting SLNB after NAT in this population, as well as in single‐center and national registries. Consistent to our findings, Xiang et al. reported on 7772 cN1 patients from the SEER database nd observed that the use of SLNB increased from 25.2% in 2010% to 55.0% in 2021 [30]. Also, in the Netherlands Cancer Registry SLNB was performed as first axillary treatment in 46% of cN+ patients [22], while the single‐center series by Nguyen et al. including 430 patients with cN1 tumors reports a greater increase in the SLNB (as the first treatment) rates from 28% in 2009 to 86% in 2017 [21]. Results from a national Italian survey including 4115 cN+ patients receiving BCS from the year 2000 to 2015 showed that SLNB was the only staging procedure in an increasing proportion of patients since 2008 (p < 0.001) and more steeply since 2011 [32].
The SenTa prospective registry, including 199 patients with cN+ tumors from 50 centers across Germany reported that TAD alone was not associated with an increased risk of recurrence (HR 0.83; 95% CI, 0.34−2.05; p = 0.69) or death (HR, 1.07; 95% CI, 0.31−3.70; p = 0.91). Similarly, the study by Montagna et al. [35] including 1144 patients, reported 5‐year rates of any axillary, locoregional, or invasive recurrence in the entire cohort of 1.0% (95% CI 0.49%–2.0%), 2.7% (95% CI 1.6%–4.1%), and 10% (95% CI 8.3%–13%) respectively. There were no differences regarding the technique. At 3 years, SLNB and TAD did not differ significantly in the rate of axillary recurrence (0.8% vs. 0.5%; p = 0.55) or locoregional recurrence (1.9% vs. 0.8%; p = 0.19) or in the rate of any invasive recurrence (7.8% vs. 7.3%; p = 0.60). Regarding the surgical technique, the AXSANA study is an European prospective multicenter registry on axillary surgical techniques after neoadjuvant chemotherapy, whose primary endpoints are 5‐years invasive disease‐free survival, 3 years axillary recurrence rate, quality of life and arm morbidity [36]
Although the current study describes a large cohort of patients from different centers across Europe, there are some limitations. One is the retrospective nature of the study, even though the EUSOMA database is regularly updated by the certified centers. There is also, mainly in the first period a high rate of missing information, that could be explained by the fact that some items related to neoadjuvant treatments were not mandatory. Finally, the EUSOMA database does not capture detailed information on the types of chemotherapy regimens administered, nor does it include data on the use of anti‐HER2 targeted therapies. Nevertheless, this lack of treatment‐specific information is expected to be addressed with the upcoming version of the database and the publication of new quality indicators [29]. Similarly, it lacks documentation regarding the specific clinical decision‐making that led to the performance of ALND in cN0 patients. The absence of these variables limits our ability to fully understand why certain treatments were chosen and how they may have influenced patient outcomes.

5
Conclusions
Our findings reflect an ongoing shift in clinical practice across Europe toward de‐escalation of axillary surgery after neoadjuvant therapy in both cN0 and cN+ breast cancer, consistent with international trends. This approach has the potential to improve patients’ quality of life by reducing the morbidity associated with axillary lymph node dissection, without compromising oncological safety. However, careful patient selection remains essential, particularly for those with initially node‐positive disease. Long‐term prospective data are needed to confirm the oncological safety of omitting extensive axillary surgery and to define evidence‐based criteria for optimal patient selection.

Eusoma Working Group
Chiara Annunziata Pasqualina Anghelone8, Antonella Baldissera9a–b, Elisabetta Benozzi10, Martine Berlière11, Marina Bortul12, Barbara Bussels13, Francesco Caruso14, Carla Cedolini15, Fabio Corsi16a–b, Evelyn Despierre17, Boudewijn Dullens18, Gianluca Fogazzi19, Lucio Fortunato20, José Luis Fougo21, Gianluca Frisoni22, Daniele Generali23 a–b, Alessandra Gennari24, Matteo Ghilli25, Marco Golinelli26, Pedro Gouveia27 a–b, Simona Grossi28, Ruth Helfgott29, Alessandra Huscher30, Karolina F. Larsson31, Sophie Marquette32, Carmela Mocerino33, Ida Negreiros34, Francesca Pellini35, Federico Piacentini36, Romano Polato37, Annemie Prové38, Karol Rogowski39, Giovanna Romanucci40, Lorenzo Rossi41, Claudia Francesca Rossi42, Margherita Serra43, Gracienne Staelens44, Maud Vassilieff45, Paolo Veronesi46 a–b, Daniele Zanoni47

8Struttura Complessa di Chirurgia 3 ‐ Senologia ‐ Breast Cancer Centre, Fondazione
IRCCS Policlinico San Matteo, Pavia, Italy

9a.Radiation Oncology Bellaria Hospital Bologna Italy

b.Clical Reference Breast Unit Oncology Department AUSL Bologna Italy

10Breast Surgery Unit, Centro di Riferimento Oncologico CRO Aviano IRCCS, National
Cancer Institute, Aviano (PN), Italy

11Cliniques Universitaires StLuc, Bruxelles, Belgium

12Breast Unit Trieste ASUGI ‐ University of Trieste, Italy

13Breast Cinic Unit, AZ Delta, Roeselare, Belgium

14Department of Oncology HICC and Humanitas Breast Center
Humanitas Istituto Clinico Catanese (HICC) Misterbianco (Catania), Italy

15Breast Unit Azienda Ospedaliero Univrsitaria Friuli Centrale (ASUFC) Udine ‐ Italy

16a.Department of Biomedical and Clinical Sciences, University of Milan, Milan, Italy

b.Breast Unit, Istituti Clinici Scientifici Maugeri IRCCS, Pavia, Italy

17Breast Clinic – Department of Gynaecology and Obstetrics, AZORG Aalst, Belgium

18Breast Clinic Voorkempen, AZ Klina, Brasschaat, Belgium

19Medical Oncology, Istituto Clinico S. Anna–Brescia, Italy

20Breast Center, San Giovanni‐Addolorata Hospital, Rome, Italy

21Breast Center, ULS São João, Porto, Portugal

22Breast Surgery–Breast Unit Rimini‐Santarcangelo ‐ Rimini, Italy

23aMultidisciplinary Breast Cancer Unit, Cremona Hospital, Cremona, Italy

bDepartment of Medicine, Surgery and Health Sciences, University of Trieste, Italy

24Università del Piemonte Orientale, SCDU Oncologia, AOU Maggiore della Carità
Novara, Italy

25Breast Centre, Pisa University Hospital, Pisa, Italy

26Breast Unit AUSL di Modena, Italy

27a.Breast Unit, Champalimaud Foundation/Champalimaud Clinical Centre, Portugal

b.Lisbon School of Medicine, Lisbon, Portugal

28Breast Centre Ortona–ASL02 Abruzzo, Italy

29Brustgesundheitszentrum des Tumorzentrums Oberösterreich, Linz, Austria

30Fondazione Poliambulanza Brescia, Italy

31Breast Centre at Sahlgrenska University Hospital and Institute of Clinical Sciences,
Sahlgrenska Academy, University of Gothenburg, Sweden

32Breast Surgery, Hasselt Jessa Ziekenhuis, Belgium

33Breast Unit, Azienda Ospedaliera di Rilievo Nazionale Antonio Cardarelli di Napoli,
Italy

34Hospital CUF Descobertas, Lisboa, Portugal

35Breast Surgery, Breast Unit University Hospital of Verona, Italy

36Division of Medical Oncology ‐ Department of Medical and Surgical Sciences for
Children and Adults, University Hospital of Modena, Italy

37Breast Unit Bolzano, Italy

38Medical Oncology, Breast Cancer Clinic, ZAS Augustinus, Wilrijk/Antwerpen, Belgium

39M. Skłodowska‐Curie Bialystok Oncology Center, Bialystok, Poland

40UOSD Breast Unit ULSS9, Ospedale Di Marzana, Piazzale Lambranzi, Verona, Italy

41Istituto Oncologico della Svizzera Italiana (IOSI), Ente Ospedaliero Cantonale (EOC), Switzerland

42Fondazione IRCCS Ca’ Granda Ospedale Policlinico Milano, Breast Surgery Unit, Italy

43Breast Surgery Unit, IRCCS Azienda Ospedaliero‐Universitaria di Bologna, Italy

44Breast Surgery, Azgroeninge Kortrijk, Belgium

45Clinique du Sein ISALA, Bruxelles, Belgium

46a.Breast Surgery Division, Istituto Europeo di Oncologia, IRCCS ‐ Milan, Italy

b.Dipartimento di Emato‐Oncologia, Università degli Studi di Milan, Italy

47Azienda Ospedaliero‐ Universitaria of Parma, Italy

Author Contributions
Conceptualization, Methodology, Writing‐original draft: I. Rubio, A. J. Esgueva Methodology, Formal Analysis, Investigation, Writing – review and editing: M. Tomatis, A. Ponti Writing – review and editing, data curation: L. Marotti Writing – review and editing: MJ. Cardoso, KL. Cheung; J. De Vries; P. van Dam.

Funding
The study received no specific funding for this work.

Conflicts of Interest

DG: Speaker grant from Lilly, Pfizer, Roche, Novartis, Astrazeneca

KFL: Advisory funding from Pfizer.

Synopsis
This study evidences a substantial shift towards SLNB as the primary axillary surgery following NAT during the study period. This trend emphasizes a preference for less invasive procedures, likely due to the efficacy of neoadjuvant therapy in reducing axillary lymph node involvement.