Postmastectomy radiation therapy in patients with human epidermal growth factor receptor 2 (HER2)-positive breast cancer: a retrospective exploratory cohort analysis of the adjuvant lapatinib and/or trastuzumab treatment optimization (ALTTO) trial [BIG 2-06/NCCTG N063D (Alliance)] Authors: Maria Carmen De Santis, Christine Campbell, Faye Samy, Fiorenza De Rose, Bruno Meduri, Matteo Lambertini, Christos Sotiriou, Maria Grazia Carnevale, Antonella Ferro, Cristina Saura, Anup Choudhury, Carlotta Giandini, Paolo Verderio, Saranya Chumsri, Valentina Appierto, Martine Piccart, Giancarlo Pruneri, Riccardo Ray Colciago, Evandro de Azambuja, Serena Di Cosimo Published: 2026 ABSTRACT SummaryBackgroundPostmastectomy radiation therapy (PMRT) improves locoregional control and survival in patients with early breast cancer, but it is unclear whether these benefits extend to human epidermal growth factor receptor 2 (HER2)-positive patients in the anti-HER2 era.MethodsThis was a retrospective cohort analysis of prospectively collected data from the phase III ALTTO trial and includes all ALTTO participants who underwent mastectomy regardless of randomisation arm. The phase III ALTTO trial enrolled patients between 2007 and 2011 from 946 centers from 44 countries in four different continents. Eligible patients had histologically confirmed HER2-positive breast cancer with either node-positive disease or node-negative disease with pathologic tumour size ≥1 cm. PMRT was non-randomised and delivered at the investigator's discretion. Primary efficacy analyses were conducted in the intention-to-treat population; safety analyses included all patients who received at least 1 dose of anti-HER2 therapy. Endpoints were locoregional recurrence (LRR), distant recurrence (TTDR), disease-free (DFS), and overall survival (OS). DFS and OS were analysed using Cox models; LRR and TTDR competing risk models. All models included the interaction between number of metastatic nodes, 0 (pN0), 1-3 (pN1) or ≥4 (pN2) and PMRT. Multivariable models accounted for patient and tumour characteristics. Hazard ratios (HR) and their 95% confidence interval (CI) are reported.FindingsOverall, 4154 patients treated with mastectomy were analysed: 1987 (47.8%) with and 2167 (52.2%) without PMRT. Baseline characteristics were worse in the PMRT group, whereas adjuvant endocrine and chemotherapy use/type was similar between groups. At a median follow-up of 9.2 years (IQR, 6.4-10), 129 LRR events occurred: 43 (2.2%) in the PMRT group and 86 (4.0%) in the no-PMRT group. In patients with pN2, PMRT was associated with improved LRR control (HR 0.31; 95% CI 0.16–0.59), TTDR (HR 0.59; 95% CI 0.44–0.79), DFS (HR 0.58; 95% CI 0.45–0.75), and OS (HR 0.61; 95% CI 0.44–0.86). In patients with pN1, PMRT showed marginal benefit in DFS (HR 0.82; 95% CI 0.64–1.05) but not in LRR (HR not calculated), TTDR (HR 0.89; 95% CI 0.65–1.23), nor OS (HR 0.88; 95% CI 0.62–1.24). Only 229 (15%) patients with pN0 received PMRT, with no evidence of benefit in any endpoints.InterpretationIn this large HER2-positive, predominantly node-positive, adjuvant cohort, PMRT improved locoregional control and conferred a clinically meaningful survival benefit in pN2. In the landscape of personalized treatment and growing interest in axillary de-escalation, future prospective studies are warranted to better define which HER2-positive patients with low nodal burden (particularly pN1 disease) derive meaningful benefit from PMRT in the context of modern systemic therapy.FundingThe publication of this analysis was funded by the 10.13039/501100003196Italian Ministry of Health through Ricerca Corrente funds. Research in contextEvidence before this studyWe conducted a comprehensive search of PubMed from the 1st of January 2005 up to the 25th of July 2025 to identify studies evaluating the role of radiation therapy (RT) in patients with Human Epidermal Growth Factor Receptor 2 (HER2)-positive breast cancer undergoing mastectomy. Search terms included combinations of “breast cancer”, “HER2-positive”, “mastectomy”, “radiotherapy”, “postmastectomy radiation therapy”, and “axillary lymph nodes”. We included clinical trials, prospective and retrospective cohort studies, and meta-analyses in English and this research yielded 160 studies.Studies were included if they specifically addressed HER2-positive breast cancer, patients treated with mastectomy, and the use or omission of RT, particularly in the context of limited axillary nodal involvement. We excluded studies not reporting HER2 status, those including mixed populations without subgroup analysis, studies focused exclusively on breast-conserving surgery, and articles published before the introduction of anti-HER2 therapies.The available evidence shows that while RT improves locoregional control, its benefit in patients with low nodal burden in the HER2-positive subgroup remains uncertain, particularly in the era of effective systemic therapies.Added value of this studyThis study provides focused evidence on the role of RT in a clinically relevant and underexplored subgroup of patients with HER2-positive breast cancer undergoing mastectomy with limited axillary nodal involvement. By specifically analyzing this population in the context of contemporary systemic treatments, our findings help address a current gap in the literature, where most available data derive from heterogeneous cohorts or predate the widespread use of anti-HER2 therapies.Our results contribute to clarifying the potential benefit of RT in this setting, offering data that may support more tailored treatment decisions and help refine current de-escalation strategies.Implications of all the available evidenceTaken together, current evidence suggests that the role of RT after mastectomy should be individualised, particularly in patients with HER2-positive disease and low nodal burden. While advances in systemic therapies may reduce the absolute benefit of RT in some patients, its contribution to locoregional control remains clinically relevant in selected cases.Our findings support a more nuanced approach to postmastectomy RT in HER2-positive breast cancer, highlighting the need for careful risk stratification and multidisciplinary decision-making. Further prospective studies are warranted to better define which patients may safely omit RT and which are most likely to benefit from its use. Research in context Evidence before this study We conducted a comprehensive search of PubMed from the 1st of January 2005 up to the 25th of July 2025 to identify studies evaluating the role of radiation therapy (RT) in patients with Human Epidermal Growth Factor Receptor 2 (HER2)-positive breast cancer undergoing mastectomy. Search terms included combinations of “breast cancer”, “HER2-positive”, “mastectomy”, “radiotherapy”, “postmastectomy radiation therapy”, and “axillary lymph nodes”. We included clinical trials, prospective and retrospective cohort studies, and meta-analyses in English and this research yielded 160 studies. Studies were included if they specifically addressed HER2-positive breast cancer, patients treated with mastectomy, and the use or omission of RT, particularly in the context of limited axillary nodal involvement. We excluded studies not reporting HER2 status, those including mixed populations without subgroup analysis, studies focused exclusively on breast-conserving surgery, and articles published before the introduction of anti-HER2 therapies. The available evidence shows that while RT improves locoregional control, its benefit in patients with low nodal burden in the HER2-positive subgroup remains uncertain, particularly in the era of effective systemic therapies. Added value of this study This study provides focused evidence on the role of RT in a clinically relevant and underexplored subgroup of patients with HER2-positive breast cancer undergoing mastectomy with limited axillary nodal involvement. By specifically analyzing this population in the context of contemporary systemic treatments, our findings help address a current gap in the literature, where most available data derive from heterogeneous cohorts or predate the widespread use of anti-HER2 therapies. Our results contribute to clarifying the potential benefit of RT in this setting, offering data that may support more tailored treatment decisions and help refine current de-escalation strategies. Implications of all the available evidence Taken together, current evidence suggests that the role of RT after mastectomy should be individualised, particularly in patients with HER2-positive disease and low nodal burden. While advances in systemic therapies may reduce the absolute benefit of RT in some patients, its contribution to locoregional control remains clinically relevant in selected cases. Our findings support a more nuanced approach to postmastectomy RT in HER2-positive breast cancer, highlighting the need for careful risk stratification and multidisciplinary decision-making. Further prospective studies are warranted to better define which patients may safely omit RT and which are most likely to benefit from its use. Introduction Several large randomised trials have demonstrated that postmastectomy radiation therapy (PMRT) improves locoregional control and survival in breast cancer. The Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) meta-analysis of 22 randomised trials comparing mastectomy with or without PMRT reported that PMRT reduced locoregional recurrence (LRR) by 19%, translating into an 8% reduction in breast cancer mortality at 20 years in patients with 4 or more positive nodes (pN2). Although this meta-analysis showed a consistent survival advantage for all node-positive patients, its applicability to patients with 1–3 positive nodes (pN1) remains debated.1 The main criticisms of the trials reviewed in the meta-analysis focus on their systemic therapies, which lacked taxanes, dose-dense chemotherapy, HER2-targeted treatments, extended endocrine therapy, and optimal axillary lymph node dissection. Moreover, several other studies have shown low LRR in pN1 patients.2 In addition, recent series show smaller absolute benefits in LRR and overall survival (OS) than historical cohorts,3, 4, 5, 6 likely reflecting advances in the understanding of breast cancer biology and improvements in systemic therapies, diagnostic tools, and surgical techniques, that have enhanced both local and systemic disease control.7 In parallel, the past five decades have seen a huge improvement in radiotherapy technique, reducing unwanted incidental irradiation dose of nearby organs, thus limiting side effects.8 A new meta-analysis of the EBCTCG on individual data, from 14,324 patients enrolled in 16 trials during 1961–2008, showed that radiotherapy after breast-conserving surgery, or after mastectomy in node-positive disease, reduced recurrence (relative risk, RR 0.88; p = 0.0008), breast cancer mortality (RR 0.87; p = 0.0010), and all cause of mortality (RR 0.90; p = 0.0022) with no significant effect of non-breast cancer mortality.9 Analysis of the Danish 82 b&c trials uncovered heterogeneity in PMRT outcomes according to tumour subtype.10, 11, 12 However, randomised trials specifically assessing the role of PMRT by molecular subtype and expression of the human epidermal growth factor receptor 2 (HER2), in the anti-HER2 era, are lacking. The ALTTO BIG 2-06 trial was a large, international, randomised phase III study evaluating contemporary adjuvant chemotherapy combined with anti-HER2 therapy (trastuzumab or lapatinib, administered alone, in sequence, or in combination).13 Conducted between 2007 and 2011, after the time window of the trials included in the latest EBCTCG meta-analysis, ALTTO BIG 2-06 provides a large prospectively collected dataset of patients with centrally confirmed HER2-positive early breast cancer, which we used to perform a post-hoc analysis assessing the impact of PMRT on locoregional control and survival outcomes in the anti-HER2 era. The results of the present analysis of the ALTTO trial were partially reported at ESMO Breast Cancer 2025, Berlin, May 14–17, 2025. Methods Patients This is an unplanned exploratory analysis of the ALTTO BIG 2-06 trial, which evaluated trastuzumab and/or lapatinib as adjuvant treatment in 8381 patients with HER2-positive early breast cancer.13 Patients were randomised to one of four anti-HER2 treatment arms of 52-week duration: trastuzumab alone (n = 2097), lapatinib alone (n = 2100), trastuzumab for 12 weeks followed by lapatinib for 34 weeks (n = 2091), or combination trastuzumab + lapatinib (n = 2093).13 For the purpose of the present study, eligible patients included all ALTTO participants who underwent mastectomy regardless of randomization arm. To evaluate the effect of PMRT, patients were divided into 2 groups: PMRT (1987 patients), and no-PMRT (2167 patients) (Fig. 1). Total ALTTO population, patients undergoing mastectomy, exclusion criteria applied; and final analysed cohort are shown in Fig. 1 (Study Profile).Fig. 1Study profile. Study profile. PMRT Delivery of PMRT was site-level and non-randomised. Investigators followed protocol recommendations (Supplementary Table S1) for candidate selection, yet ultimate treatment decisions rested with the treating physician. At the time of patient enrollment in the ALTTO trial, there was no robust evidence supporting the benefit of PMRT in patients with 1–3 positive lymph nodes. Radiation targets included the ipsilateral chest wall, the supraclavicular region, and the axilla; internal mammary nodes were included according to the local protocol (Supplementary Table S2). All patients were treated with conventional fractionation 50 Gy in 25 fractions. Outcomes Study endpoints were defined according to the ALTTO protocol as follows: LRR, defined as any local or regional recurrence of invasive breast cancer, time to distant recurrence (TTDR), defined as time to randomisation to first distant breast cancer recurrence, ignoring loco regional recurrence and second primary cancers (including contra lateral breast cancers and non-breast second malignancies), disease-free survival (DFS), time from randomization to the first occurrence of any of the following: recurrence of invasive breast cancer at local, regional, or distant sites; contralateral invasive breast cancer; second non-breast malignancy; or death as a result of any cause; overall survival (OS). DFS, TTDR and OS were computed using the original definitions of the main ALTTO analysis. Time to LRR (TTLRR) was not a main ALTTO study endpoint and was defined as the interval between the date of randomization and the first occurrence of LRR, death without LRR was treated as a competing risk. First LRR was defined using the main ALTTO study derivation of local and regional recurrences. Statistical analysis Survival function for OS and DFS were estimated by the Kaplan–Meier method, the hazard ratio and 95 CI% for PMRT vs. no PMRT were estimated using Cox’s proportional hazards models. The cumulative incidence function was generated for TTLRR and TTDR. Hazard ratios and 95 CI% for PMRT vs. no PMRT were generated by Fine–Gray competing risk regression analyses, death was considered competing risk for both endpoints. Main ALTTO study stratification factors (hormone receptor status, timing of chemotherapy, nodal status) were included in the models as covariates and only included as stratification factors where assumption of non-proportional hazards is violated for that covariate. In addition, pre-specified baseline covariates including age, tumour size, chemotherapy timing, anti-HER2 regimen, surgical management, and race (not pre-specified), were fitted individually, then penalised regression methods by Elastic Net regression were performed to select, from the baseline characteristics, the set of variables for inclusion into the multivariable models. To explore whether the effect of radiotherapy is consistent within subgroups, within each of the nodal strata, tests for heterogeneity were performed. The present analysis was not pre-planned in the study protocol, and the power of the statistical analyses for these endpoints was not pre-specified. Ethics The ALTTO (BIG 2-06/NCCTG N063D [Alliance]) trial was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. The study protocol and all amendments were approved by the institutional review board or independent ethics committee at each participating center. All patients provided written informed consent before study enrollment. The present analysis was pre-specified and approved by the ALTTO Executive Committees and Steering Committees prior to its initiation. Role of the funding source The publication of this analysis was funded by the Italian Ministry of Health through Ricerca Corrente funds. Conversely, the funders had no involvement in study design, data collection, data analyses, data interpretation, or the writing of the report. Results Of the 8381 patients enrolled in the ALTTO BIG 2-06 trial, 4154 (49.6%) underwent mastectomy (Fig. 1). Among them, 1987 (47.8%) received PMRT, and 2167 (52.2%) did not. Baseline characteristics are summarized in Table 1. Compared with no-PMRT, patients treated with PMRT were slightly younger (median 50 versus [vs.] 52 years), had larger nodal involvement (pN2 55.2% vs. 8.3%) and primary tumours (pT3, 16.7% vs. 3.8%). Most (55.8%) had poorly differentiated grade 3 invasive ductal carcinoma. Estrogen/progesterone receptor status was similar (53.7% vs. 53.3%). Anthracycline-taxane chemotherapy was more frequent with PMRT (79.1% vs. 61.2%), whereas endocrine therapy use was comparable. Median follow-up was 9.2 (IQR 6.4-10) years. Overall, 129 LRR occurred (2.2% with PMRT vs. 4.0% without).Table 1Demographic and disease characteristics of the patients at baseline.No PMRT (N = 2167)PMRT (N = 1987)Age at diagnosis median (IQR)52 (44, 58)50 (43, 57)Age at diagnosis ≤40 years338 (15.6%)382 (19.2%) 41–64 years1598 (73.7%)1428 (71.9%) ≥65 years231 (10.7%)177 (8.9%)Race White1143 (52.7%)1360 (68.4%) Asian880 (40.6%)541 (27.2%) Black or African American43 (2.0%)26 (1.3%) American Indian or Alaska Native75 (3.5%)43 (2.2%) Native Hawaiian or other Pacific Islander5 (0.2%)4 (0.2%) Not reported21 (1.0%)13 (0.7%)Hormone receptor status ER and/or PR positive1155 (53.3%)1068 (53.7%) ER and PR negative1012 (46.7%)919 (46.3%)Metastatic lymph nodes Node negative1266 (58.4%)229 (11.5%) 1–3 positive nodes722 (33.3%)662 (33.3%) ≥4 positive nodes179 (8.3%)1096 (55.2%)Histology Ductal carcinoma1154 (53.3%)1098 (55.3%) Other40 (1.8%)46 (2.3) No carcinoma in Situ973 (44.9%)843 (42.4%)Tumor grade G1: Well differentiated57 (2.6%)39 (2.0%) G2: Moderately differentiated920 (42.5%)675 (34.0%) G3: Poorly differentiated/undifferentiated1111 (51.3%)1207 (60.7%) Missing/GX: Differentiation cannot be assessed79 (3.6%)66 (3.3%)Tumor size ≤2 cm916 (42.3%)529 (26.6%) >2 cm1240 (57.2%)1446 (72.8%) ≤5 cm2073 (95.7%)1643 (82.7%) >5 cm83 (3.8%)332 (16.7%)Anti-HER2 Trastuzumab550 (25.4%)502 (25.3%) Lapatinib538 (24.8%)500 (25.2%) Trastuzumab → Lapatinib535 (24.7%)476 (24.0%) Trastuzumab + Lapatinib544 (25.1%)509 (25.6%)Timing of chemotherapy Concurrent1054 (48.6%)959 (48.3%) Sequential1113 (51.4%)1028 (51.7%)Type of chemotherapy Anthracycline-based692 (31.9%)321 (16.2%) Anthracycline and taxane-based1326 (61.2%)1571 (79.1%) Anthracycline free140 (6.5%)95 (4.8%)Adjuvant endocrine therapy AI305 (26.4%)327 (30.6%) AI and SERM166 (14.4%)174 (16.3%) SERM512 (44.3%)477 (44.7%) Other4 (0.3%)7 (0.7%)Type of surgery Mastectomy + axillary dissection1282 (59.2%)1441 (72.5%) Mastectomy + sentinel node biopsy883 (40.7%)545 (27.4%) Mastectomy2 (0.1%)1 (0.1%)PMRT, Post-mastectomy radiation therapy; IQR, Interquartile range; ER, Estrogen receptor; PR, Progesterone receptor; G1, Grade 1; G2, Grade 2; G3, Grade 3; GX, Grade cannot be assessed; AI, Aromatase inhibitor; SERM, Selective Estrogen Receptor Modulators. Table 1Demographic and disease characteristics of the patients at baseline.No PMRT (N = 2167)PMRT (N = 1987)Age at diagnosis median (IQR)52 (44, 58)50 (43, 57)Age at diagnosis ≤40 years338 (15.6%)382 (19.2%) 41–64 years1598 (73.7%)1428 (71.9%) ≥65 years231 (10.7%)177 (8.9%)Race White1143 (52.7%)1360 (68.4%) Asian880 (40.6%)541 (27.2%) Black or African American43 (2.0%)26 (1.3%) American Indian or Alaska Native75 (3.5%)43 (2.2%) Native Hawaiian or other Pacific Islander5 (0.2%)4 (0.2%) Not reported21 (1.0%)13 (0.7%)Hormone receptor status ER and/or PR positive1155 (53.3%)1068 (53.7%) ER and PR negative1012 (46.7%)919 (46.3%)Metastatic lymph nodes Node negative1266 (58.4%)229 (11.5%) 1–3 positive nodes722 (33.3%)662 (33.3%) ≥4 positive nodes179 (8.3%)1096 (55.2%)Histology Ductal carcinoma1154 (53.3%)1098 (55.3%) Other40 (1.8%)46 (2.3) No carcinoma in Situ973 (44.9%)843 (42.4%)Tumor grade G1: Well differentiated57 (2.6%)39 (2.0%) G2: Moderately differentiated920 (42.5%)675 (34.0%) G3: Poorly differentiated/undifferentiated1111 (51.3%)1207 (60.7%) Missing/GX: Differentiation cannot be assessed79 (3.6%)66 (3.3%)Tumor size ≤2 cm916 (42.3%)529 (26.6%) >2 cm1240 (57.2%)1446 (72.8%) ≤5 cm2073 (95.7%)1643 (82.7%) >5 cm83 (3.8%)332 (16.7%)Anti-HER2 Trastuzumab550 (25.4%)502 (25.3%) Lapatinib538 (24.8%)500 (25.2%) Trastuzumab → Lapatinib535 (24.7%)476 (24.0%) Trastuzumab + Lapatinib544 (25.1%)509 (25.6%)Timing of chemotherapy Concurrent1054 (48.6%)959 (48.3%) Sequential1113 (51.4%)1028 (51.7%)Type of chemotherapy Anthracycline-based692 (31.9%)321 (16.2%) Anthracycline and taxane-based1326 (61.2%)1571 (79.1%) Anthracycline free140 (6.5%)95 (4.8%)Adjuvant endocrine therapy AI305 (26.4%)327 (30.6%) AI and SERM166 (14.4%)174 (16.3%) SERM512 (44.3%)477 (44.7%) Other4 (0.3%)7 (0.7%)Type of surgery Mastectomy + axillary dissection1282 (59.2%)1441 (72.5%) Mastectomy + sentinel node biopsy883 (40.7%)545 (27.4%) Mastectomy2 (0.1%)1 (0.1%)PMRT, Post-mastectomy radiation therapy; IQR, Interquartile range; ER, Estrogen receptor; PR, Progesterone receptor; G1, Grade 1; G2, Grade 2; G3, Grade 3; GX, Grade cannot be assessed; AI, Aromatase inhibitor; SERM, Selective Estrogen Receptor Modulators. Demographic and disease characteristics of the patients at baseline. PMRT, Post-mastectomy radiation therapy; IQR, Interquartile range; ER, Estrogen receptor; PR, Progesterone receptor; G1, Grade 1; G2, Grade 2; G3, Grade 3; GX, Grade cannot be assessed; AI, Aromatase inhibitor; SERM, Selective Estrogen Receptor Modulators. It was observed that PMRT indication and prognosis depended on nodal involvement, patients with pN0 were more likely to not receive PMRT (No PMRT, 1266 (58.4%) vs. PMRT, 229 (11.5%)), while those with pN2 were more likely to be treated with PMRT (No PMRT, 179 (8.3%) vs. PMRT, 1096 (55.2%)). Formal testing of the interaction between PMRT and the number of metastatic lymph nodes (pN0, pN1, pN2) found the interaction to be significant, and to mitigate confounding by indication, the interaction between PMRT and the number of metastatic lymph nodes was fitted in a single model, including all the data, for each efficacy endpoint. With 33 (3.0%) and 13 (7.3%) recurrences in those with and without PMRT respectively, PMRT was associated with a reduction in LRR in patients with pN2 (adjusted hazard ratio (aHR) 0.31, 95% CI 0.16–0.59), with early and sustained separation of cumulative–incidence curves (Fig. 2a). Precision was limited in patients with pN1 (4 (<1%) vs. 38 (5.3%) events with vs. without PMRT), and pN0 (6 (2.6%) vs. 35 (2.8%) events with vs. without PMRT).Fig. 2a: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by hormone receptor status. Adjusted for chemotherapy timing and anti-HER2 treatment. HR, Hazard ratio; LRR, Loco-regional recurrence; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. b: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by hormone receptor status. Adjusted for chemotherapy timing, tumour size, anti-HER2 treatment, and loco regional recurrence. aHR, Adjusted hazard ratio; DR, distant recurrence; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. a: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by hormone receptor status. Adjusted for chemotherapy timing and anti-HER2 treatment. HR, Hazard ratio; LRR, Loco-regional recurrence; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. b: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by hormone receptor status. Adjusted for chemotherapy timing, tumour size, anti-HER2 treatment, and loco regional recurrence. aHR, Adjusted hazard ratio; DR, distant recurrence; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. PMRT reduced distant recurrence in patients with pN2 (recurrences, 287 (26%) PMRT vs. 58 (32%) no PMRT, aHR, 0.56, 0.42–0.75), but not in pN1 (recurrences, 69 (10%) PMRT vs. 80 (11%) no PMRT, aHR 1.09, 0.79–1.52), or pN0 (recurrences, 24 (10%) PMRT vs. 76 (6%) no PMRT, aHR 1.57, 0.99–2.49) (Fig. 2b). PMRT improved DFS in patients with pN2 (PMRT, 355 (32%) vs. no PMRT 69 (39%) aHR 0.58, 0.45–0.75). No evidence of benefit was observed in those patients with pN1, 114 (17%) and 141 (20%) events in those with and without PMRT, respectively, aHR (0.85, 0.66–1.09). In those patients with pN0, there were 36 (16%) and 167 (13%) events (1.03, 0.71–1.47) (Fig. 3a). PMRT improved OS in patients with pN2 (PMRT, 226 (21%) vs. no PMRT 42 (23%), aHR 0.62, 0.45–0.88); with no evidence of significant effects in patients with pN1 (deaths, 61 (9%) PMRT vs. 71 (10%) no PMRT, aHR 0.90, 0.64–1.27), or pN0 (deaths, 17 (7%) PMRT vs. 73 (6%) no PMRT, aHR 1.06, 0.60–1.76) (Fig. 3b).Fig. 3a: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by chemotherapy timing and hormone receptor status. Adjusted for tumour size, age, anti-HER2 treatment and surgery. aHR, Adjusted hazard ratio; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. b: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by hormone receptor status. Adjusted for chemotherapy timing, tumour size, age, anti-HER2 treatment, surgery and systemic therapy. aHR, Adjusted hazard ratio; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. a: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by chemotherapy timing and hormone receptor status. Adjusted for tumour size, age, anti-HER2 treatment and surgery. aHR, Adjusted hazard ratio; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. b: Model includes interaction between PMRT and number of metastatic lymph nodes. Stratified by hormone receptor status. Adjusted for chemotherapy timing, tumour size, age, anti-HER2 treatment, surgery and systemic therapy. aHR, Adjusted hazard ratio; PMRT, Post-mastectomy radiation therapy; HER2, human epidermal growth factor receptor 2. Formal subgroup analyses were limited by low event counts. Results within subgroups for age, tumour size, multifocal/multicentric tumour, axillary treatment, chemotherapy timing, hormonal receptor status, and anti-HER2 treatment were consistent with the overall effect for DFS, OS and TTDR and no evidence of statistical heterogeneity in the effect of PMRT across the subgroups (Supplementary Tables S3–S5). The extent of irradiation to chest wall alone vs. chest wall plus regional nodes had no detectable impact on DFS within nodal strata (Supplementary Table S6). Discussion In the large ALTTO cohort, PMRT was associated with lower LRR (2.2% vs. 4.0%) and, most notably, with significantly fewer distant recurrences and improved DFS and OS among patients with pN2 disease; effects in patients with pN1 were directionally favorable but not statistically significant, whereas no benefit was detected in patients with pN0, where the low numbers limited inference. Our results align with and extend the existing evidence base. The EBCTCG meta-analysis and the Danish DBCG 82b/82c trials established that PMRT reduces LRR and breast cancer mortality in node-positive disease.1,14 More recent randomised studies,15, 16, 17 confirmed that comprehensive regional nodal irradiation improves disease control and reduces breast cancer mortality in the modern era, notably without excess non-cancer mortality observed in earlier radiotherapy epochs, an evolution plausibly attributable to heart- and lung-sparing techniques (e.g., 3D/IMRT planning, DIBH/respiratory gating). Within this context, the ALTTO cohort with a median follow-up of 9.2 years and treatment delivered using contemporary planning offers external validity to current practice. The stratified analyses clarify where PMRT is most impactful. In pN2 disease, PMRT was associated with a marked reduction in LRR (HR 0.31, 95% CI 0.16–0.59), fewer distant recurrences (HR 0.59, 0.44–0.79), and improved DFS (HR 0.58, 0.45–0.75) and OS (HR 0.61, 0.44–0.86), consistent with a biologically coherent mechanism in which treatment of residual microscopic foci in the chest wall and regional basins reduces the locoregional reservoir and, downstream, systemic dissemination. In pN1 disease, effects trended in the same direction but did not reach statistical significance overall; signals were stronger for tumours >5 cm, supporting a risk-adapted interpretation. Importantly, the pN1 findings should be interpreted with caution, given the observational nature of the analysis, the limited number of events and the risk of residual confounding; therefore, they should be considered hypothesis-generating rather than sufficient to support definitive clinical recommendations and should not be interpreted as evidence of a definitive interaction between PMRT and HER2-targeted therapy. In pN0, where baseline LRR risk is low under multimodality therapy, no benefit was detected, and absolute gains are expected to be small. The HER2-positive setting deserves special attention. Before trastuzumab, HER2 positivity conferred high recurrence risk, amplifying the absolute benefit of locoregional control. In the trastuzumab era, absolute risks are lower, and contemporary data on PMRT remain limited. Recently, SUPREMO trial provides a valid external comparator for PMRT in intermediate-risk postmastectomy disease; however, because HER2 status was not an enrollment criterion and HER2-positive patients represented a minority of the trial, it could not specifically address whether the effect of PMRT is modified in the setting of HER2-targeted therapy.18 In the HERA cohort, PMRT reduced LRR substantially in patients with 1–3 positive nodes (HR ∼0.28) without a statistically significant OS advantage.19 Earlier work from Kyndi et al.12 suggested higher LRR in HER2-positive and hormone-receptor-negative tumours and an OS benefit primarily in hormonal receptor-positive disease; however, this pre-anti-HER2 era likely overestimates modern benefits. Our findings fit this evolving risk landscape, showing that LRR events are overall low. Tseng et al.20 showed that the molecular subtype of breast cancer was related to a different LRR risk; Basal-like and HER2-like tumours are associated with higher and Luminal A tumours with lower LRR rates.21 Moreover, HER2-positive disease receiving anti-HER2 therapy had a substantially lower LRR risk.22 Our data should also be interpreted against shifting surgical paradigms. Contemporary management increasingly avoids axillary lymph-node dissection (ALND) in favor of conservative approaches for macro- and micrometastatic disease.22, 23, 24 Together with AMAROS trial,25 demonstrating axillary radiotherapy provides disease control comparable to ALND with lower morbidity, these changes are expanding the pool of patients considered for PMRT and elective nodal irradiation. This evolution, while patient-centered, further limits the applicability of older trials to current practice and amplifies the need for biologically informed selection. Biology driven selection of patients for RT holds great potential. Few gene expression profiles have been published that are either prognostic for LRR, or predictive of radio-sensitivity or radio-resistance.26,27 One of these was developed from tumour tissues from patients included in the DBCG 82 b&c cohorts.10,11 The genes in the DBCG-RT profile were related to the immune system. A high level of tumour infiltrating lymphocytes (TILs) is found in only 10% of breast cancers and especially Basal like and HER2-enriched subtypes respect to luminal subtypes.28 The relationship between TILs, RT and clinical outcome is unclear. High TILs are associated with increased benefit from RT in terms of OS, with a trend towards lower DM risk, but without effect on LRR.29 Our study has limitations Radiotherapy was not randomised, introducing potential confounding by indication (PMRT patients were younger and had higher stage and more intensive chemotherapy), and residual confounding may persist despite multivariable adjustment and competing-risk modelling. It is clinically plausible that the no-PMRT group may have included some patients who, according to current standards, would now be considered candidates for PMRT. However, in ALTTO, PMRT was mandatory for patients with pN2 and T4 breast cancer, which likely limited the inclusion of the highest-risk categories in the no-PMRT group. Event counts for LRR, particularly in pN0 and pN1 disease, were low, limiting precision and power to detect modest effects and to interrogate treatment-field nuances (chest wall vs. chest wall and nodes). Details of systemic therapy and RT field design/dose, while contemporary, may vary across centers. Regarding late cardiac events, with a median follow-up of 6.9 years, in the ALTTO trial most events occurred during anti-HER2 treatment rather than during follow-up. However, a residual risk of late cardiac events cannot be completely ruled out, as patient-level data beyond the published follow-up were unavailable. Finally, subgroup analyses (e.g., by tumour size within pN1) should be viewed as exploratory. In summary, despite the limitations discussed above, within a modern HER2-positive population, PMRT was associated with clinically meaningful reductions in recurrence and improvements in survival concentrated in patients with pN2 disease, with no evidence of benefit in those patients with pN1 or pN0. The convergence of biological plausibility, statistical interaction across endpoints, and contemporary RT delivery supports strong consideration of PMRT for pN2 disease and selective, risk-adapted use for those with pN1, while routine PMRT for unselected pN0 appears unwarranted. As surgical de-escalation and nodal irradiation expand, the integration of validated biomarkers of radio-sensitivity and immune contexture will be essential to target PMRT to those most likely to benefit. Contributors SDC and MCDS conceptualized and designed the study. SDC, CSa, AF, EdA, and MP provided the study materials and patients. SDC, MCDS and CC accessed and verified the data. All authors contributed to collection, assembling and analysis and interpretation of the data. SDC and MCDS wrote the original draft and all authors reviewed and approved of its submission. Data sharing statement The datasets used and analysed during the current study are available from the corresponding author on reasonable request. Declaration of interests AC reports stock options from Novartis (as part of ESOP) and role as full-time employee of Novartis. SC reports Institutional grants from Merck Sharp & Dohme, Pfizer, Salix Pharmaceuticals, Rebiotix Inc, Novartis, BriaCell Therapeutics; consulting fees from AstraZeneca, Daiichi Sankyo, Immunomedics, Biotheranostics, Novartis, Athenex, Syndax, Puma Biotechnology, Eisai, Seagen, Genentech, Pfizer, Menarini Stemline, Axiom, Cardinal Health, Cepton Strategies, Eradigm, Equinox, Dedham group, Clarivate, MDOutlook; honoraria from Roche Thailand, Merck Sharp & Dohme Thailand, MDOutlook. EdA reports Institutional research grants from F. Hoffman-Roche, Genentech, AstraZeneca, GlaxoSmithKline, Novartis, Gilead Sciences, Pfizer, Seagen; consulting fees from F. Hoffman-Roche, Genentech, Novartis, Pierre Fabre, Eli Lilly, AstraZeneca, Merck Sharp & Dohme, Gilead Sciences, Bayer, Boehringer Ingelheim, Menarini; honoraria for educational events from F. Hoffman-Roche, Genentech, Novartis, Pierre Fabre, Eli Lilly, AstraZeneca, Merck Sharp & Dohme, Gilead Sciences, Bayer, Boehringer Ingelheim, Menarini; travel support from Merck Sharp & Dohme, Gilead Sciences; honoraria for advisory board participation for F. Hoffman-Roche/Genentech, Novartis, Pierre Fabre, Eli Lilly, AstraZeneca, Merck Sharp & Dohme, Gilead Sciences, Bayer, Boehringer Ingelheim, Menarini; and lastly unpaid leadership roles in BSMO (president 2023–2026) and BIG (Executive Board member 2025–2026). SDC reports institutional grant from the Ministry of Health for the project PNRR-POC-2023-12378113; “ad hoc” medical monitor for Medica Scientia Innovation Research (MEDSIR), Barcelona (Spain).” AF reports honoraria from AstraZeneca, Novartis, Daiichi Sankyo; other financial support (e.g., for attending meetings or travel) from Menarini Stemline, AstraZeneca, Eli Lilly. ML reports an Institutional grant from Gilead; consulting fees from Hoffman–La Roche, Lilly, Novartis, AstraZeneca, Pfizer, Gilead Sciences, Merck Sharp & Dohme, Pierre Fabre, Menarini, Nordic Pharma, Bayer, Ipsen, Daiichi Sankyo; honoraria from Roche, Eli Lilly, Novartis, Astra Zeneca, Pfizer, Gilead Sciences, Merck Sharp & Dohme, Pierre Fabre, Menarini, Nordic Pharma, Bayer, Ipsen, Daiichi Sankyo; other financial support (e.g., for attending meetings or travel) from Gilead, Daiichi Sankyo, Hoffman–La Roche; and non-financial interests as JCO Consultant Editor, Chair of the ESO Certificate of Competence in Breast Cancer. MP reports grants from AstraZeneca, Eli Lilly, Menarini, Merck Sharp & Dohme, Novartis, Pfizer, F. Hoffman- Roche/Genentech, Servier, Gilead Sciences; honoraria from AstraZeneca, Gilead Sciences, Eli Lilly, Menarini, Mersana, Merck Sharp & Dohme, Novartis, Pfizer, F. Hoffman- Roche/Genentech, Summit Therapeutics, Ellipses Pharma, Jazz Pharmaceuticals. CSa reports grants from AstraZeneca, Boehringer Ingelheim, Bristol Myers Squibb, Byondis BV, Daiichi Sankyo, Eisai, F. Hoffman–La Roche, Genentech, Gilead Sciences, GlaxoSmithKline, Menarini, Merck Sharp & Dohme, Merus, Novartis, Pfizer, Philips, Pierre Fabre, Puma, Sanofi Aventis, Seagen, Synthon, Zymeworks; consulting fees from AstraZeneca, Daiichi Sankyo, Eisai, MediTech, Novartis, Pfizer, Philips Healthcare, Pierre Fabre, Puma Biotechnology, F. Hoffman–La Roche, Seagen, Gilead Sciences, Eli Lilly, Synthon, Zymeworks, Pharmalex; honoraria from AstraZeneca, Daiichi Sankyo, Pfizer, Pierre Fabre, Puma Biotechnology, Seagen, Exeter Pharmaceuticals, Eli Lilly; payment for expert testimony from AX's Consulting SARL, Boehringer Ingelheim, Bristol Myers Squibb, Genentech, Innoup, Merck Sharp & Dohme Espana; Novartis, Sanofi; support for attending meetings and/or travels from Pfizer, Novartis, F. Hoffman–La Roche, AstraZeneca, Puma Biotechnology, Daiichi Sankyo, Eisai Europe, Gilead Sciences, Eli Lilly, Pierre Fabre, Seagen; participation on a Data Safety Monitoring Board or Advisory Board for AstraZeneca, Eisai, Daiichi Sankyo, Gilead Sciences, Lilly, Novartis, Pfizer, Pierre Fabre, Philips, F. Hoffman- Roche, Seagen, Menarini; unpaid role as SOLTI board of directors and also reports being member of MSD Advisory Board. CS reports honoraria from Exact Sciences, Prime Oncology; other financial support (e.g., for attending meetings or travel) from Exact Sciences, Prime Oncology; stock or stock options from Signature Biosciences. All other authors declare no competing interests.