Pathological complete response and postmastectomy radiation therapy after primary systemic therapy: results from the ReSTORE study Authors: Liat Mizrahi, Marianna Valzano, Rinat Bernstein Molho, Carlotta Becherini, Luca Visani, Viola Salvestrini, Chiara Mattioli, Andrea Rampini, Lorenzo Livi, Ben Boursi, Icro Meattini, Orit Kaidar-Person Published: 2026 ABSTRACT AbstractBackgroundThe ReSTORE study aimed to evaluate oncological outcomes and explore the association between pathological complete response (pCR), postmastectomy radiotherapy and nodal irradiation (PMRT), and recurrence.MethodsThis multicentre retrospective study included patients with breast cancer treated with chemotherapy-based primary systemic therapy (PST) followed by skin-or nipple-sparing mastectomy and immediate reconstruction. Locoregional recurrence-free survival (LRRFS) and disease-free survival (DFS) were analysed.ResultsA total of 256 patients were analysed, with a median follow-up of 82 months (range 16-166). In multivariable analysis, node-positive disease at diagnosis was associated with worse LRRFS (HR 8.28, 95% CI 1.06–64.34; p = 0.043). PMRT was associated with a non-significant reduction in locoregional recurrence (HR 0.74, 95% CI 0.29–1.92; p = 0.54). pCR was significantly associated with improved DFS (HR 0.48, 95% CI 0.28–0.82; p = 0.008). Among patients receiving PMRT, those achieving pCR had superior DFS compared with those without pCR (HR 0.23, 95% CI 0.09–0.60; p = 0.002). No significant associations with DFS were observed for age, BRCA status, tumour stage, or receptor status.ConclusionsIn this cohort, pCR and baseline nodal status were key prognostic factors, whereas the true impact of PMRT on outcomes remained uncertain. Nonetheless, our findings do not support omission of PMRT based on pCR alone. Prospective trials and long follow up are required to define response-adapted locoregional treatment strategies.Keywords: Breast cancer, Neoadjuvant, Primary systemic therapy, Pathological complete response, Mastectomy, Radiation, Radiotherapy Introduction Pathological complete response (pCR) after primary systemic therapy (PST) is a well-established prognostic factor associated with improved survival outcomes, particularly in aggressive breast cancer subtypes [1]. The CTNeoBC pooled analysis demonstrated that eradication of tumour in both breast and lymph nodes is strongly associated with improved survival, while conversely a residual nodal disease is associated with a worse survival regardless the response in the breast after PST [2]. For these reasons, locoregional management is increasingly tailored according to response to therapy [[3], [4], [5], [6]]. While de-escalation strategies have been successfully implemented in surgery [3], the role of tailoring postmastectomy radiation therapy (PMRT) including regional node irradiation (RNI) remains uncertain, particularly in patients achieving pCR [[4], [5], [6]]. The recently published National Surgical Adjuvant Breast and Bowel Project (NSABP) B-51-Radiation Therapy Oncology Group (RTOG) 1304 trial, recruiting cN1 breast cancer patients who turned ypN0, reported that RNI did not significantly improve the invasive breast cancer recurrence-free interval, distant recurrence-free interval, disease-free survival, or overall survival [7]. This question is especially relevant in patients undergoing skin- or nipple-sparing mastectomy (SSM/NSM) with immediate reconstruction, where PMRT may increase complication rates and affect reconstructive outcomes [[8], [9], [10]]. The ReSTORE (Radiotherapy after Systemic Therapy, Outcomes, and Response Evaluation) study aims to describe oncological outcomes in this specific population and explore the association between pCR, PMRT, and recurrence outcomes. This analysis is intended as exploratory and hypothesis-generating. Methods This retrospective study was conducted at two university teaching hospitals and approved by the ethics committee of each centre. Electronic medical records were reviewed for the period 2017–2023. The analysis included patients aged ≥18 years with a diagnosis of breast cancer who received chemotherapy-based primary systemic therapy (PST) and subsequently underwent skin-sparing mastectomy (SSM) or nipple-sparing mastectomy (NSM) with immediate reconstruction, with a minimum of 6 months of clinical follow-up. Patients with reconstruction failure were not excluded. All histological and biological subtypes of invasive breast cancer were eligible. Exclusion criteria included patients who underwent non–skin-sparing mastectomy, bilateral breast cancer, mastectomy and reconstruction following breast cancer recurrence, or mastectomy in the presence of metastatic disease, or with metastatic disease diagnosed within 3 months of mastectomy or completion of systemic therapy; patients for whom limited information regarding PST was available or for whom sufficient follow-up data was not available. We also excluded patients who had undergone neoadjuvant endocrine therapy. Clinical, demographic, and treatment-related variables were collected. Based on the 8th edition of the American Joint Committee on Cancer (AJCC) staging manual, a pCR was defined as the absence of residual invasive carcinoma upon histopathological review following PST and surgery. The presence of in situ carcinoma alone, without any invasive component, qualified as an in-breast pCR [11]. Radiation therapy was according to institutional protocol, during the study period some patients were treated with 50 Gy in 25 fractions which was considered as the standard for PMRT and regional nodes, more recent cases were treated with moderate hypofractionation (40.05 Gy or 42.4 Gy in 15 or 16 fractions, respectively). All cases were planned for chest wall and RNI. Both institutions used the European Society for Radiotherapy and Oncology (ESTRO) delineation atlas for planning [12,13]. However, for patients treated before 2020 at Sheba Medical Center, the Radiation Therapy Oncology Group (RTOG) atlas was primarily used. Notably, the ESTRO guidelines for delineation following implant reconstruction were not applied in any of the cases [14]. Statistical analysis Local recurrence (LR) was defined as any ipsilateral event documented in the medical records as a local recurrence, without reclassification based on histopathological features or anatomical subsite. Chest wall recurrence was defined as any ipsilateral recurrence involving the chest wall distinct from LR. Regional recurrence, in the absence of LR or chest wall involvement, was classified as regional-only recurrence. Cases with synchronous LR or chest wall recurrence and regional recurrence were classified as locoregional recurrence (LRR). Locoregional events were considered as such only in the absence of documented synchronous distant failure. For survival analyses, all locoregional events were grouped as LRR due to the limited number of events. Locoregional recurrence-free survival (LRRFS) was defined as the time from the end of therapy to the occurrence of any LR or regional recurrence. Disease-free survival (DFS) was defined as the time from the end of therapy to any breast cancer event, including local, regional, or distant recurrence. Descriptive statistics are presented as mean, median, range, and standard deviation (SD), as appropriate. Hazard ratios (HRs) were estimated using Cox proportional hazards models in both univariable and multivariable analyses. Survival outcomes were analysed using the Kaplan–Meier method, and differences between groups were assessed using appropriate statistical tests. All statistical tests were two-sided, and a p value < 0.05 was considered statistically significant. Statistical analyses were performed using Stata/IC version 15 (StataCorp, College Station, TX, USA). Results Of the 383 cases initially identified, 118 were excluded (Fig. 1 supplement). This left a total of 265 breast cancer patients who underwent PST and immediate reconstruction for the final analysis, with a median age of 46 years (range: 25–75 years). Patient characteristics and treatment details are summarised in Table 1. Overall, 34% of patients achieved a pCR and 144 underwent PMRT (56%). Clinical characteristics according to pCR status and radiotherapy are summarised in Table 2. At a median follow-up of 82 months (range 16–166), the cumulative incidence of recurrence was 4.7% (7/256) for LR, 1.2% for regional recurrence (without local events), and 16% for distant recurrence. The cumulative incidence of contralateral new primary breast cancer was 1.2%, and 4% for other (non-breast) malignancies. Sites of recurrence are detailed in Table 3. Three patients (1.2%) had a contralateral new primary breast cancer at the time of follow up. Locoregional free survival (LRRFS) in patients with cN0 versus cN1-3. Table 1Clinicopathological characteristics of the study population according to pathological complete response.Table 1CharacteristicOverall cohort (n = 256)pCR (n = 87)Non-pCR (n = 169)Age, years46 (25–75)45.6 (25–73)46.3 (26–75)Menopausal statusPremenopausal169 (66.0)53 (60.9)116 (68.6)Postmenopausal79 (30.9)29 (33.3)50 (29.6)Unknown8 (3.1)5 (5.7)3 (1.8)BRCA statusBRCA146 (18.0)30 (34.5)16 (9.5)BRCA223 (9.0)6 (6.9)17 (10.1)Unknown115 (45.0)29 (33.3)86 (50.9)Multicentric/multifocal disease154 (60.2)40 (46.0)114 (67.5)HistologyDCIS with microinvasion4 (1.6)4 (4.6)0IDC133 (52.0)47 (54.0)86 (50.9)ILC12 (4.7)1 (1.1)11 (6.5)DCIS + IDC61 (23.8)15 (17.2)46 (27.2)IDC + ILC11 (4.3)011 (6.5)Other10 (3.9)1 (1.1)9 (5.3)Missing25 (9.8)19 (21.8)6 (3.6)Tumour grade13 (1.2)03 (1.8)276 (29.7)12 (13.8)64 (37.9)3115 (44.9)41 (47.1)74 (43.8)Missing62 (24.2)34 (39.1)28 (16.6)Clinical T stageT132 (12.5)14 (16.1)18 (10.7)T2139 (54.3)54 (62.2)85 (50.3)T349 (19.1)13 (14.9)36 (21.3)T435 (13.7)6 (6.9)29 (17.2)Clinical N stageN064 (25.0)33 (37.9)31 (18.3)N1112 (43.8)36 (41.4)76 (45.0)N263 (24.6)15 (17.2)48 (28.4)N317 (6.6)3 (3.4)14 (8.3)SubtypeTNBC60 (23.4)37 (42.5)23 (13.6)ER-positive/PR-negative/HER2-negative21 (8.2)6 (6.9)15 (8.9)ER-positive/PR-positive/HER2-negative80 (31.3)6 (6.9)74 (43.8)ER-positive/HER2-positive61 (23.8)19 (21.8)42 (24.9)ER-negative/HER2-positive34 (13.3)19 (21.8)15 (8.9)Axillary surgerySLNB69 (27.0)38 (43.7)31 (18.3)ALND159 (62.1)37 (42.5)122 (72.2)Targeted axillary surgery26 (10.2)10 (11.5)16 (9.5)Missing2 (0.8)2 (2.3)0Number of involved lymph nodes0133 (52.0)82 (94.3)51 (30.2)1–369 (27.0)5 (5.7)64 (37.9)>454 (21.1)054 (32.0)Extranodal extension present55 (21.5)055 (32.5)Data are presented as n (%) unless otherwise indicated. Age is presented as mean (range).Abbreviations: pCR, pathological complete response; DCIS, ductal carcinoma in situ; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; ER, oestrogen receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2; TNBC, triple-negative breast cancer; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection. Clinicopathological characteristics of the study population according to pathological complete response. Data are presented as n (%) unless otherwise indicated. Age is presented as mean (range). Abbreviations: pCR, pathological complete response; DCIS, ductal carcinoma in situ; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; ER, oestrogen receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2; TNBC, triple-negative breast cancer; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection. Table 2Clinicopathological characteristics according to pathological complete response and postmastectomy radiation therapy.Table 2CharacteristicpCR + PMRT (n = 42)pCR, no RT (n = 45)Non-pCR + PMRT (n = 102)Non-pCR, no RT (n = 67)Age, years44.9 (26–69)46.2 (25–73)46.0 (26–75)46.8 (29–67)Menopausal statusPremenopausal28 (66.7)25 (55.6)71 (69.6)45 (67.2)Postmenopausal12 (28.6)17 (37.8)29 (28.4)21 (31.3)Unknown2 (4.8)3 (6.7)2 (2.0)1 (1.5)BRCA statusBRCA117 (40.5)13 (28.9)11 (10.8)5 (7.5)BRCA24 (9.5)2 (4.4)14 (13.7)3 (4.5)Unknown8 (19.0)21 (46.7)43 (42.2)43 (64.2)Multicentric/multifocal disease18 (42.9)22 (48.9)78 (76.5)36 (53.7)HistologyDCIS with microinvasion2 (4.8)2 (4.4)00IDC27 (64.3)20 (44.4)56 (54.9)30 (44.8)ILC01 (2.2)6 (5.9)5 (7.5)DCIS + IDC6 (15.6)9 (20.0)26 (25.5)20 (29.9)IDC + ILC006 (5.9)5 (7.5)Other01 (2.2)4 (3.9)5 (7.5)Missing7 (16.7)12 (26.7)4 (3.9)2 (3.0)Tumour grade1003 (2.9)024 (9.5)8 (17.8)43 (42.2)21 (31.3)326 (61.9)15 (33.3)42 (41.2)32 (47.8)Missing12 (28.6)22 (48.9)14 (13.7)14 (20.9)Clinical T stageT110 (23.8)4 (8.9)14 (13.7)4 (6.0)T217 (40.5)37 (82.2)37 (36.3)48 (71.6)T310 (23.8)3 (6.7)25 (24.5)11 (16.4)T45 (11.9)1 (2.2)25 (24.5)4 (6.0)Clinical N stageN010 (23.8)23 (51.1)5 (4.9)26 (38.8)N119 (45.2)17 (37.8)46 (45.1)30 (44.8)N210 (23.8)5 (11.1)37 (36.3)11 (16.4)N33 (7.1)014 (13.7)0SubtypeTNBC20 (47.6)17 (37.8)15 (14.7)8 (11.9)ER-positive/PR-negative/HER2-negative3 (7.1)3 (6.7)12 (11.8)3 (4.5)ER-positive/PR-positive/HER2-negative2 (4.8)4 (8.9)54 (52.9)20 (29.9)ER-positive/HER2-positive7 (16.7)12 (26.7)21 (20.6)21 (31.3)ER-negative/HER2-positive10 (23.8)9 (20.0)015 (22.4)Axillary surgerySLNB15 (35.7)23 (51.1)9 (8.8)22 (32.8)ALND18 (42.9)19 (42.2)83 (81.4)39 (58.2)Targeted axillary surgery7 (16.7)3 (6.7)10 (9.8)6 (9.0)Missing2 (4.8)000Number of involved lymph nodes038 (90.5)44 (97.8)14 (13.7)37 (55.2)1–34 (9.5)1 (2.2)35 (34.3)29 (43.3)>40053 (52.0)1 (1.5)Data are presented as n (%) unless otherwise indicated. Age is presented as mean (range).Abbreviations: PST, primary systemic therapy; pCR, pathological complete response; RT, radiation therapy; PMRT, postmastectomy radiotherapy (chest wall with or without regional nodal irradiation); DCIS, ductal carcinoma in situ; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; ER, oestrogen receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2; TNBC, triple-negative breast cancer; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection. Clinicopathological characteristics according to pathological complete response and postmastectomy radiation therapy. Data are presented as n (%) unless otherwise indicated. Age is presented as mean (range). Abbreviations: PST, primary systemic therapy; pCR, pathological complete response; RT, radiation therapy; PMRT, postmastectomy radiotherapy (chest wall with or without regional nodal irradiation); DCIS, ductal carcinoma in situ; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; ER, oestrogen receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2; TNBC, triple-negative breast cancer; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection. Table 3Sites of breast cancer events.Table 3Site of recurrencen (% of total population)Number of patients who received PMRTLocal recurrencea7 (2.7)4Chest wall recurrence3 (1.2)2Regional (lymph nodes only)5 (2)2Chest wall and regional nodes3 (1.2)2Distant recurrence41 (16)27aLocal recurrence defined as an ipsilateral event recorded in the medical records as occurring at the primary tumour site. Sites of breast cancer events. Local recurrence defined as an ipsilateral event recorded in the medical records as occurring at the primary tumour site. In multivariable analysis, node-positive disease at diagnosis (regardless of nodal stage) was associated with significantly worse locoregional recurrence-free survival (LRRFS) (p = 0.043; HR 8.28, 95% CI 1.06–64.34) (Fig. 1). PMRT was associated with a reduction in the risk of LRR, although this did not reach statistical significance (p = 0.54; HR 0.74, 95% CI 0.29–1.92). Distant recurrence was associated with higher clinical nodal stage, although this did not reach statistical significance (p = 0.08; HR 2.09, 95% CI 0.92–4.72), and a similar non-significant trend was observed for higher clinical T stage (T1-2 vs T3-4) (p = 0.11; HR 1.65, 95% CI 0.89–3.07). Patients achieving pCR had significantly improved DFS compared with those without pCR (p = 0.008; HR 0.48, 95% CI 0.28–0.82) (Fig. 2). Across the overall cohort, patients receiving PMRT had shorter DFS (60 vs 69 months), although this difference was not statistically significant (p = 0.237; HR 1.33, 95% CI 0.83–2.11) (Fig. 3). Among patients who received PMRT, those achieving pCR had significantly better DFS than those without pCR (p = 0.002; HR 0.23, 95% CI 0.09–0.60). As illustrated in Fig. 4, patients who achieved a pCR and PMRT appeared to have the most favourable DFS. However, due to the limited sample sizes within these subgroups, a formal statistical comparison was not performed. No statistically significant associations with DFS were observed for age, BRCA status, clinical T stage, oestrogen receptor (ER), or HER2 status. Disease free survival (DFS) in patients with pathological complete response (pCR) versus not. Disease free survival (DFS) according to postmastectomy radiotherapy (PMRT). Disease free survival (DFS) according to pathological complete response (pCR) and postmastectomy radiotherapy (PMRT). Discussion This multicentre retrospective analysis evaluates oncological outcomes in a clinically relevant and increasingly frequent scenario: patients treated with PST followed by SSM/NSM with immediate reconstruction, in whom the role of PMRT remains debated if pCR is achived. The present study was designed to explore the association between pCR, PMRT, and recurrence outcomes, and should be interpreted as hypothesis-generating due to the limited number of patients. Despite relatively high pCR rates in this cohort, reflecting the predominance of biologically aggressive subtypes such as triple-negative and HER2-positive disease, the incidence of distant events remained substantial over a median follow-up of approximately seven years, consistent with previous observations in similar populations [3]. First, pCR was associated with improved DFS, in line with its well-established prognostic role [6,15]. The CTNeoBC pooled analysis demonstrated that eradication of tumour in both breast and lymph nodes (ypT0/is ypN0) is strongly associated with improved event-free and overall survival, particularly in aggressive subtypes [2]. Conversely, residual nodal disease (ypN+) is consistently associated with poorer outcomes [6,15]. Our findings reinforce pCR as a robust prognostic marker, but not as a standalone criterion for treatment de-escalation [16]. Second, clinical nodal status at diagnosis remained associated with LRRFS, consistent with prior evidence [15,17]. The combined analysis of NSABP B-18 and B-27 trials identified clinical T and N stage, together with post-treatment pathological stage, as key predictors of locoregional recurrence following PST [17]. In mastectomy-treated patients, baseline nodal burden and residual disease remained central determinants of recurrence risk [17]. Our results confirm that baseline disease extent continues to influence outcomes despite response to systemic therapy. The interpretation of locoregional recurrence patterns in the neoadjuvant setting remains complex. In NSABP B-18 and B-27, local recurrences were approximately threefold more frequent than regional recurrences [17], suggesting that the primary site remains a dominant source of failure. Differences between breast-conserving therapy and mastectomy cohorts may partly reflect the impact of radiation delivery [15], including incidental irradiation of the lower axilla with tangential breast fields [17,18]. In our cohort, the relatively low number of regional events compared with distant recurrences may reflect both surgical management of the axilla and the selective use of PMRT in higher-risk patients [19]. Like the outcomes that were described in the upfront surgery, unplanned analysis from the prospective EORTC 22922/10925 showing that different components of therapy are related to the individual risk of recurrence but also influenced the site of first event [19]. The question of tailoring radiation therapy according to response to PST remains unresolved. The EBCTCG meta-analysis raised concerns regarding higher LRR rates when locoregional treatment is inadequately adapted after PST [6]. More recently, prospective data have begun to inform this area. The NSABP B-51/RTOG 1304 trial suggests that, in patients with cN1 disease converting to ypN0, omission of RNI may not compromise invasive recurrence-free interval or survival outcomes [7]. Similarly, the RAPCHEM study explored response-adapted radiation therapy strategies, suggesting that de-escalation may be feasible in selected patients [4]. However, these approaches remain investigational and require longer follow-up before routine adoption. Our study focuses on patients undergoing SSM/NSM with immediate reconstruction, a subgroup in whom PMRT carries specific clinical implications. Implant-based reconstruction is associated with increased risks of complications, including capsular contracture and reconstructive failure, when combined with radiation therapy [[20], [21], [22], [23]]. Therefore, identifying patients in whom PMRT could potentially be omitted without compromising oncological safety is of particular importance. However, our findings do not support the omission of PMRT for all patients. Currently, clinical trial evidence remains limited, particularly because mastectomy outcomes were not reported as a separate subgroup with statistical power [7]. Several limitations must be acknowledged, given the limited number of LRR events, multivariable analyses should be interpreted cautiously and considered exploratory. A few other limitations for generalizing the results include: first, the retrospective design, which introduces inherent biases, particularly confounding by indication in the comparison between PMRT and no-PMRT groups, as patients receiving PMRT had more advanced baseline disease. Consequently, observed differences in DFS reflect underlying risk rather than treatment effect. Second, the study population is heterogeneous, including variability or missing data in tumour stage, nodal stage in the no-PMRT group, molecular subtype, systemic therapy, and surgical approach, with some subgroup analyses limited by small numbers. Third, locoregional recurrence definitions were not standardised in the literature, and differences in surgical techniques, including the extent of mastectomy and residual glandular tissue, may influence recurrence patterns and comparability with other cohorts [[24], [25], [26]]. Variations in reporting of locoregional events, including differences in classification of nodal recurrences (e.g., axillary versus internal mammary), may partly explain discrepancies with other studies such as the MARI trial [3,27]. These issues highlight the importance of standardised reporting of locoregional treatments and outcomes, as advocated in recent consensus efforts [28,29]. In addition, response to PST was assessed as a binary pCR variable rather than using residual cancer burden (RCB), which provides a more refined prognostic stratification and may better capture heterogeneity among non-pCR patients [30]. Finally, the evolving landscape of systemic therapy must be considered: NSABP B-18 and B-27 did not incorporate molecular tumour subtypes [17] and modern treatments such as anti-HER2 agents, which significantly impact pCR rates and outcomes [31], limiting the direct applicability of historical data to current practice. This study confirms the prognostic relevance of pCR and baseline nodal status in a reconstructed mastectomy population but does not provide conclusive evidence to support omission of PMRT. Pathological complete response has been shown to predict long-term outcomes in breast cancer patients undergoing PST in aggressive early breast cancer. However, our study showed that patients who achieved a pCR and treated with PMRT had the most favourable DFS. Importantly, the caveats of determining pCR, and should be taken into consideration prior to fast adoption of omission strategies [16]. Our findings should be considered exploratory and may inform the design of future prospective trials aimed at safely individualising locoregional therapy. In conclusions, pCR is associated with improved outcomes, while baseline disease characteristics remain critical in determining recurrence risk. Prospective trials are required to guide treatment de-escalation strategies in this setting. CRediT authorship contribution statement Liat Mizrahi: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Marianna Valzano: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Rinat Bernstein Molho: Conceptualization, Data curation, Investigation, Methodology, Supervision, Writing – review & editing. Carlotta Becherini: Conceptualization, Supervision, Writing – review & editing. Luca Visani: Writing – review & editing. Viola Salvestrini: Writing – review & editing. Chiara Mattioli: Writing – review & editing. Andrea Rampini: Writing – review & editing. Lorenzo Livi: Supervision, Writing – review & editing. Ben Boursi: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – review & editing. Icro Meattini: Conceptualization, Methodology, Project administration, Supervision, Writing – review & editing. Orit Kaidar-Person: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Supervision, Writing – review & editing. Declaration of competing interest All authors confirm no COI. Footnotes The work was accepted as a poster discussion at ESTRO 2026. During the preparation of this work the author(s) used Gemini in order to labels on the images provided by that statistician. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article. Supplementary data to this article can be found online at https://doi.org/10.1016/j.breast.2026.104832. Supplementary data The following is the Supplementary data to this article: References