Survival outcomes after breast-conserving surgery with radiotherapy versus mastectomy in apocrine carcinoma of the breast: a Surveillance, Epidemiology, and End Results-based retrospective cohort study with time-dependent analysis

Introduction

Apocrine carcinoma (AC) of the breast is a rare subtype of breast cancer defined as an epithelial tumor with characteristic morphology and immunophenotype in the World Health Organization (WHO) classification (1,2). Its typical presentation includes eosinophilic cytoplasm, distinct nucleoli, and immunological features of estrogen receptor (ER) and progesterone receptor (PR) negativity and androgen receptor (AR) positivity (1-3). Although AC is frequently grouped clinically with hormone receptor-negative or triple-negative breast cancer, its AR-positive “molecular apocrine” phenotype suggests that its biological behavior, prognosis, and treatment response may differ from those of classical triple-negative breast cancer. Recent genomic and clinicopathological studies further indicate that triple-negative AC (TNAC) may represent a biologically distinct subgroup with different chemosensitivity and survival patterns compared with non-apocrine triple-negative breast cancer (4,5). Because of its incidence of less than 1%, most available studies are small-sample or single-center reports and high-quality evidence remains limited (6-8). Therefore, the clinicopathological features, prognostic factors, and optimal treatment strategies for this rare subtype remain incompletely defined (9).

Although the prevailing clinical approach to invasive ductal carcinoma is well-established, the comparative survival outcomes of different surgical approaches and radiotherapy (RT) remain to be elucidated in AC of the breast. Several earlier studies have demonstrated that breast-conserving surgery combined with RT (BCS + RT) is non-inferior or even superior to mastectomy (MAST) with regard to overall survival (OS) in the general breast cancer population (10-13). More recent population-based studies and meta-analyses have also suggested that BCS + RT may be associated with comparable or even better survival than MAST in selected patients with early-stage breast cancer (14,15). However, the evidence in AC of the breast remains limited (1,6-8,16). Whether survival findings from the general breast cancer population can be directly extrapolated to AC remains uncertain because of its rarity and distinct biological profile. For AC specifically, available literature suggests that surgery remains the cornerstone of treatment, but the optimal extent of surgery and the role of adjuvant local therapy have not been clearly established (17). Furthermore, there is a paucity of systematic assessment of the independent role of RT in this subtype. Therefore, several clinically relevant questions remain unresolved: whether RT provides an independent survival benefit after BCS, whether BCS + RT achieves survival outcomes comparable to MAST in early-stage patients eligible for breast conservation, and whether treatment effects vary over time.

A Surveillance, Epidemiology, and End Results (SEER)-based analysis can partially address this evidence gap by providing a relatively large cohort for survival-based comparisons in this rare subtype. However, SEER lacks information on surgical margins, local-regional recurrence, disease-free survival, RT details, systemic treatment regimens, and treatment compliance. Therefore, the present study cannot directly establish equivalence in local control or recurrence-free survival, but can provide real-world evidence on OS and breast cancer-specific survival (BCSS) to inform clinical decision-making and future studies.

Consequently, the present study constructed two distinct analytical cohorts based on the SEER database. The primary objective was to evaluate the survival benefit of RT in patients who underwent BCS. In addition, the study compared survival outcomes between BCS + RT and MAST in patients with early-stage disease who were eligible for breast-conserving therapy. By combining propensity score-based methods and time-dependent analyses, supplemented by a single-center descriptive clinicopathological case series, this study aimed to provide survival-based evidence to inform individualized treatment of AC of the breast. The findings should be interpreted as survival-based evidence rather than definitive evidence of equivalent local-regional control. We present this article in accordance with the STROBE reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-2026-0220/rc).

Methods

Data source and study design

This study was a population-based retrospective cohort study using data from the SEER database, supplemented by a single-center descriptive clinicopathological case series. The SEER cohort was derived from Incidence SEER Research Data, 17 registries, Nov 2024 Sub [2000–2022]. Female patients aged ≥20 years who were diagnosed between 2010 and 2022 with histologically confirmed primary AC of the breast were identified using the International Classification of Diseases for Oncology, Third Edition (ICD-O-3) histology code 8401/3 and breast primary sites. Cases identified solely through death certificates or autopsy reports were excluded. The SEER database is a population-based cancer registry covering multiple geographic regions and diverse populations in the United States, making it suitable for real-world survival analyses of rare tumor subtypes.

Patients were included in the final surgical comparison cohorts if they had clear documentation of surgery type and available survival follow-up information. Patients without surgery of the primary site, those with unknown or other surgery types, and those with missing follow-up information were excluded. Two analytical modules were designed. First, a BCS cohort was established to compare survival outcomes between patients treated with BCS + RT and those treated with BCS alone. Second, an early-stage cohort was defined as patients with Stage I–II disease, T1–2 tumors, N0–1 nodal status, and tumor size ≤50 mm, representing patients considered eligible for BCS, and was used to compare the survival outcomes of BCS + RT versus MAST. In addition, patients with AC of the breast treated at The First Affiliated Hospital of Kunming Medical University were included as a single-center descriptive case series for supplementary analysis of clinicopathological and treatment characteristics. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by The First Affiliated Hospital of Kunming Medical University (approval No. 202506-3847), and individual consent for this retrospective analysis was waived.

For the SEER cohort, follow-up was obtained from SEER survival variables, including Survival months, Vital status, and SEER cause-specific death classification. Follow-up was calculated from diagnosis to death or the administrative censoring date of December 31, 2022. Patients alive at last follow-up were censored, and those with missing follow-up information were excluded. The final SEER cohort had a mean follow-up of 73.8 months [standard deviation (SD) 42.7] and a median follow-up of 75.0 (range, 0–155). For the single-center cohort, follow-up was conducted through medical records, outpatient records, and/or telephone follow-up from diagnosis to death or last contact, with an administrative censoring date of December 10, 2025.

Covariates and propensity score adjustment

Covariates included demographic and clinicopathological characteristics, age, race/ethnicity, marital status, tumor grade, stage, chemotherapy status, RT status, surgery type, and axillary level I–II nodal burden. All variables were categorized prior to analysis.

To minimize confounding bias while preserving the sample size of this rare-disease cohort, stabilized inverse probability of treatment weighting (IPTW) based on the full eligible cohort was used as the primary propensity score-based adjustment method. Propensity scores were estimated using a logistic regression model. Propensity score matching (PSM) was performed only as a sensitivity analysis using 1:1 nearest-neighbor matching without replacement with a standardized caliper width of 0.2. Covariate balance was assessed using the standardized mean difference (SMD), with SMD <0.1 considered indicative of adequate balance.

For SEER-coded variables with unknown values, such as tumor grade, stage, tumor size, and breast subtype, “Unknown” categories were retained as independent categories in the analyses. Multiple imputation was not performed. Patients were excluded only when key eligibility variables, including surgery type or follow-up information, were unavailable.

Survival, competing risk, and time-dependent analyses

The primary outcome was OS, defined as the time from diagnosis to death from any cause, while the secondary outcome was BCSS, defined as the time from diagnosis to death attributable to breast cancer according to the SEER cause-specific death classification. Survival curves were estimated using the Kaplan-Meier method, and differences between groups were evaluated using the log-rank test. Hazard ratios (HRs) and their 95% confidence intervals (CIs) were estimated using a Cox proportional hazards model.

For BCSS, the Fine-Gray competing risks model was employed to estimate subdistribution HRs (sHRs), with non-breast cancer deaths treated as competing events. Given the potential for treatment effects to vary over time, time-dependent analytical approaches were additionally applied, including landmark analysis, continuous landmark analysis, and restricted mean survival time (RMST) analysis, to assess survival differences across different follow-up periods.

Statistical analysis

All statistical analyses were performed using R software (version 4.5.1), primarily utilizing the packages survival, survminer, cmprsk, MatchIt, and cobalt. All statistical tests were two-sided, and a P<0.05 was considered statistically significant.

Results

Baseline characteristics of the study population

A total of 1,199 patients with AC of the breast were enrolled in this study, including 202, 469, 400 and 128 patients in the BCS, BCS + RT, MAST and MAST + RT groups (Figure 1, Table 1). The overall population was predominantly composed of patients aged ≥50 years (88.2%), and the predominant ethnicity was NH White (65.2%). With regard to pathological characteristics, 58.5% of cases were classified as grade I–II, while 84.2% of cases were classified as stage I–II, suggesting that the overall population was dominated by early to mid-stage cases. A notable imbalance in baseline characteristics was observed between the treatment groups: the BCS + RT group had a higher prevalence of early-stage cases, whereas the MAST + RT group comprised more advanced and high-risk patients, with the highest proportions of stage III disease, T3–4 disease, N1–3 disease, and tumors >50 mm. The highest percentage of patients receiving chemotherapy was observed in the MAST + RT group (88.3%). The predominant molecular subtype was hormone receptor-negative (HR−)/human epidermal growth factor receptor 2-negative (HER2−) (52.3%), while HER2-negative cases accounted for 71.2%. Axillary level I–II nodal burden differed across treatment groups, with the MAST + RT group showing a higher proportion of patients with positive nodes, especially ≥4 positive nodes, suggesting greater baseline disease burden in this group.

Figure 1 Flowchart of patient selection and study design. BCS, breast-conserving surgery; ICD-O-3, International Classification of Diseases for Oncology, Third Edition; MAST, mastectomy; RT, radiotherapy; SEER, Surveillance, Epidemiology, and End Results.

Prognostic factors for OS and BCSS

The Kaplan-Meier analysis of the entire cohort demonstrated significant disparities in OS and BCSS curves for each of the four local treatment modalities (all P<0.001) (Figure 2A,2B). The BCS + RT group demonstrated the most favorable survival outcomes, while the MAST + RT group exhibited the least favorable outcomes. Multivariate Cox regression analysis further demonstrated (Table 2) that advanced age and increasing tumor stage were independent poor prognostic factors for OS, with a significantly increased risk of death in patients aged 65–79 years and ≥80 years; increasing tumor stage also significantly increased the risk of BCSS. The administration of chemotherapy was found to be associated with improved OS, with the hormone receptor-positive (HR+)/human epidermal growth factor receptor 2-positive (HER2+) subtype demonstrating a protective effect in terms of OS. With regard to treatment patterns, BCS + RT was found to be independently protective in both OS (HR =0.53, 95% CI: 0.38–0.75, P<0.001) and BCSS (HR =0.42, 95% CI: 0.23–0.74, P=0.003), with BCS serving as the reference group. Conversely, neither MAST nor MAST + RT exhibited significant differences following adjustment.

Figure 2 Unadjusted survival curves according to local treatment patterns in the overall cohort. (A) Kaplan-Meier curves for OS. (B) Kaplan-Meier curves for BCSS. BCS, breast-conserving surgery; BCSS, breast cancer-specific survival; MAST, mastectomy; OS, overall survival; RT, radiotherapy.

Survival benefit of RT in patients undergoing BCS

In view of the imbalance in baseline characteristics between treatment groups, stratified analyses of patients undergoing BCS were performed to assess the impact of RT on survival outcomes while minimizing confounding. The primary adjusted analysis was based on stabilized IPTW, which retained the full eligible BCS cohort. PSM was performed only as a sensitivity analysis to assess the consistency of the IPTW-based findings. Before matching, 670 patients were included in this cohort. After 1:1 PSM, 374 patients remained, with 187 patients in each group, and 296 unmatched patients were excluded.

RT was associated with significantly improved OS. Kaplan-Meier analysis demonstrated significantly superior OS in the BCS + RT group compared with the BCS group (P<0.001). The observed difference remained consistent after IPTW weighting (Figure 3A, P=0.01). Cox regression yielded consistent results. Compared with BCS alone, BCS + RT was associated with a significant reduction in the risk of all-cause mortality in the unadjusted model (HR =0.395, 95% CI: 0.283–0.551, P<0.001). This association remained consistent in IPTW-weighted analyses (HR =0.608, 95% CI: 0.415–0.889, P=0.01), which were interpreted as the primary adjusted results, and in PSM sensitivity analyses (HR =0.585, 95% CI: 0.391–0.876, P=0.009) (Table 3). In terms of BCSS, unadjusted analysis indicated that BCS + RT was associated with a reduced risk of BCSS (HR =0.351, 95% CI: 0.197–0.626, P<0.001). However, this difference did not reach statistical significance in the IPTW-weighted analysis (HR =0.701, P=0.32) or PSM sensitivity analysis (HR =0.553, P=0.11) (Figure 3B, Table 3). Covariate balance improved after IPTW adjustment, and PSM further improved balance in the matched sensitivity cohort (Table S1, Figure S1). The OS survival curves after PSM are displayed in Figure S2. In conclusion, RT was consistently associated with superior OS in patients undergoing BCS, while its potential benefit in BCSS did not reach statistical significance after adjustment.

Figure 3 Survival analyses comparing BCS with and without RT. (A) IPTW-weighted Kaplan-Meier curves for OS (BCS + RT vs. BCS). (B) IPTW-adjusted CIF curves for BCSS. BCS, breast-conserving surgery; BCSS, breast cancer-specific survival; CIF, cumulative incidence function; IPTW, inverse probability of treatment weighting; OS, overall survival; RT, radiotherapy.

Comparative survival outcomes of BCS + RT versus MAST in early-stage, breast-conservation-eligible patients

Further comparison of survival outcomes between BCS + RT and MAST in early-stage (I–II) patients eligible for breast-conserving therapy has important clinical implications for assessing different local treatment strategies. The primary adjusted comparison was based on stabilized IPTW using the full eligible early-stage cohort, whereas PSM was used only as a sensitivity analysis. Before matching, 696 patients were included in this cohort, comprising 297 patients treated with MAST and 399 treated with BCS + RT. After 1:1 PSM, 522 patients remained, with 261 patients in each group, and 174 unmatched patients were excluded. Covariate balance improved after IPTW and PSM adjustment (Table S2 and Figure S3). IPTW-weighted Kaplan-Meier curves showed that the difference in OS between the two groups was substantially attenuated compared with the unadjusted analysis (Figure 4A). Unadjusted Cox analysis suggested that BCS + RT was associated with a lower risk of all-cause mortality (HR =0.670, 95% CI: 0.472–0.951, P=0.03); however, this association was no longer statistically significant in the IPTW-weighted analysis (HR =0.785, 95% CI: 0.542–1.138, P=0.20), which was considered the primary adjusted result, or in the PSM sensitivity analysis (HR =0.783, 95% CI: 0.530–1.158, P=0.22) (Table 4). In terms of BCSS, no significant differences were observed between the two groups in unadjusted analysis, the IPTW-weighted Fine-Gray model, or the PSM sensitivity analysis (Figure 4B, Table 4). Tests of the proportional hazards assumption suggested time dependency (P=0.01), and landmark and time-dependent analyses were further conducted. Landmark analysis with a cutoff at 36 months showed that BCS + RT was associated with a lower risk of death during the first 36 months of follow-up (HR =0.377, 95% CI: 0.189–0.754, P=0.006), whereas this difference was no longer significant after 36 months (HR =0.858, 95% CI: 0.563–1.307, P=0.48) (Figure 4C). Serial landmark analysis further showed that the HR gradually approached or even exceeded 1 with prolonged follow-up, suggesting that the survival advantage was mainly concentrated during early follow-up (Figure 4D, Table S3). RMST analysis showed consistent results, with BCS + RT exhibiting longer mean survival times at multiple time points (Table S4). Overall, long-term OS and BCSS outcomes were generally comparable between BCS + RT and MAST in early-stage patients eligible for breast-conserving therapy, but the potential survival benefit of BCS + RT was mainly reflected in early follow-up.

Figure 4 Survival comparison and time-dependent analyses between breast-conserving therapy and MAST in early-stage, breast-conservation-eligible patients. (A) IPTW-weighted Kaplan-Meier curves for OS comparing BCS + RT and MAST. (B) IPTW-adjusted CIF curves for BCSS. (C) Landmark analysis at 3 years showing OS before and after the landmark time point. (D) Continuous landmark analysis demonstrating HRs for OS across multiple time points. BCS, breast-conserving surgery; BCSS, breast cancer-specific survival; CIF, cumulative incidence function; HR, hazard ratio; IPTW, inverse probability of treatment weighting; MAST, mastectomy; OS, overall survival; RT, radiotherapy.

Clinicopathological characteristics of AC in our cohort

To provide additional clinicopathological context, the SEER-based analysis was supplemented by the collection and analysis of clinicopathological data from 24 patients with AC of the breast (Table 5). The median age at diagnosis was 51.5 years (range, 36–71 years), with comparable proportions of patients <50 and ≥50 years (50.0% each). Tumor size was predominantly ≤5 cm (62.5%). With regard to stage, T1, N0, and American Joint Committee on Cancer (AJCC) stage I accounted for 41.7%, 66.7%, and 50.0% of cases, respectively. The overall cohort was primarily composed of early- and mid-stage cases, although a proportion of locally advanced cases was also represented. Histologic grades were grade 1–2 and grade 3 in 54.2% and 45.8% of cases, respectively, and lymphovascular invasion was present in 33.3% of patients. The proportion of patients with a Ki-67 index >14% was 58.3%, suggesting high proliferative activity in some cases. Immunophenotyping showed that all cases were estrogen receptor (ER) negative and progesterone receptor (PR) negative, androgen receptor (AR) expression was consistently positive, the GCDFP-15 positivity rate was 83.3%, and the HER2 positivity rate was 25.0%. These findings suggest that this cohort had typical immunophenotypic characteristics of AC of the breast, namely hormone receptor negativity, high AR expression, and a high proportion of GCDFP-15 positivity. In terms of treatment, 62.5% of patients underwent BCS and 37.5% underwent MAST; 87.5% received RT and 50.0% received chemotherapy. Axillary management consisted mainly of sentinel lymph node biopsy (SLNB) alone (75.0%), while 16.7% underwent axillary lymph node dissection (ALND) and 8.3% underwent SLNB followed by ALND. RT was administered to 21 patients (87.5%), most commonly using volumetric-modulated arc therapy (VMAT) or intensity-modulated RT (IMRT), and the main RT fields were whole-breast irradiation and chest-wall irradiation, with regional nodal irradiation used in 19.0% of irradiated patients. The median prescribed RT dose was 40.05 Gy (range, 40–66 Gy). Chemotherapy was administered to 12 patients (50.0%), predominantly using anthracycline plus taxane-based regimens, mostly in the adjuvant setting. Median follow-up was 40.0 months (range, 15–60 months). Given the limited sample size at a single center, these results were used mainly for descriptive analysis, and no further stratified or inferential statistical analysis was performed.

Discussion

In recent years, the optimal local treatment strategy for AC of the breast has remained controversial because this tumor is a rare and biologically distinct subtype of breast cancer. Previous studies have mostly been based on small-sample or single-center data, and consistent conclusions regarding the long-term survival difference between BCS + RT and MAST in this population are lacking. Therefore, it is important to systematically evaluate different local treatment strategies using a large population-based database supplemented by real-world clinicopathological data. Clinically, the key question is whether patients with AC who are suitable for breast conservation require more radical local surgery, or whether BCS + RT can achieve survival outcomes comparable to MAST. In this study, two clinical comparison scenarios were constructed based on the SEER database and supplemented by a single-center descriptive case series to provide additional clinicopathological context. RT was associated with significantly improved OS in patients undergoing BCS, whereas BCS + RT was generally comparable to MAST in long-term OS and BCSS among early-stage patients eligible for breast conservation. Notably, this study further observed a clear time-dependent effect: the survival benefit of BCS + RT was mainly concentrated in the early follow-up period and gradually diminished over time. This finding suggests that analyses based solely on the proportional hazards assumption may have difficulty adequately capturing dynamic differences between treatment strategies.

Several previous randomized controlled trials and population studies in the general breast cancer population have demonstrated that BCS + RT is non-inferior or even superior to MAST in terms of OS (11,18,19). Recent meta-analytic evidence has also suggested that BCS + RT may provide survival outcomes comparable to or better than MAST in selected patients with early-stage breast cancer (17). The results of this study in breast-conserving patients are consistent with previous evidence and further support the important role of RT in local control and long-term survival (18). However, after further focusing on early-stage patients eligible for breast-conserving therapy, we observed that BCS + RT and MAST were generally comparable in long-term survival outcomes, which differs somewhat from reports in some SEER or national database studies suggesting that breast-conserving therapy is superior to MAST (11). This difference may be related to differences in study population, tumor biological characteristics, and the degree of confounding control (17).

Importantly, AC should not be interpreted simply as conventional TNBC. Although many AC cases, particularly TNAC, are ER/PR-negative and may therefore be clinically grouped with TNBC, AC is commonly characterized by AR positivity and a molecular apocrine phenotype (1,4,5,11). Recent genomic and clinicopathological studies further suggest that TNAC represents a biologically distinct subgroup with molecular features, chemosensitivity, and survival patterns that differ from non-apocrine TNBC (4,20). This biological distinction may partly explain why a more radical surgical approach did not translate into superior survival in our early-stage AC cohort. Therefore, the comparable survival between BCS + RT and MAST observed in this study supports the concept that breast-conserving should not be dismissed solely because AC is frequently hormone receptor-negative.

The comparison with classical TNBC is also clinically informative. TNBC has traditionally been regarded as an aggressive subtype with a higher risk of early recurrence, which may lead some clinicians and patients to favor MAST. However, studies in TNBC have shown that breast-conserving therapy with RT can achieve acceptable local-regional and survival outcomes in appropriately selected patients (21). Our findings extend this discussion to AC, a rare AR-positive subtype that overlaps immunophenotypically with TNBC but may differ biologically. Thus, local treatment decisions in AC should be individualized rather than automatically escalated to MAST based only on receptor negativity.

From a practical surgical perspective, the survival curves in this study may be used during preoperative counselling to support shared decision-making for early-stage AC patients who meet standard criteria for breast conservation. For patients with Stage I–II disease, T1–2 tumors, N0–1 nodal status, tumor size ≤50 mm, and technically feasible negative margins, the observed comparable long-term OS and BCSS between BCS + RT and MAST suggest that MAST should not be automatically recommended solely because AC is rare or frequently hormone receptor-negative. Instead, BCS + RT can be discussed as a reasonable survival-based local treatment option when breast preservation is clinically desired and adjuvant RT is feasible.

These findings may be particularly relevant for patients with favorable clinicopathological features, such as small tumor size, low nodal burden, lower histologic grade, HER2-negative disease, and the typical AR-positive apocrine phenotype. Such patients may be particularly suitable for considering breast-conserving management. However, this interpretation should remain cautious. SEER does not include AR status, detailed margin information, RT parameters, recurrence outcomes, or comorbidity profiles; therefore, the present study cannot define a definitive de-escalation subgroup based on AR positivity, HER2 status, or tumor grade. Rather, our findings generate a clinically relevant hypothesis that favorable-biological-risk AC patients may be suitable candidates for breast-conserving management, which should be further evaluated in prospective or multicenter cohorts with detailed molecular and recurrence data.

Further analysis showed that the potential survival advantage of BCS + RT was mainly reflected in the early follow-up period and gradually diminished in the long-term follow-up. This phenomenon has been less systematically explored in previous studies, suggesting that the benefit of different local treatment strategies may be time-dependent. Several explanations may account for this observation. First, RT after BCS may reduce early local-regional failure, which could indirectly translate into an early OS benefit. However, this hypothesis cannot be directly tested in SEER because recurrence data are unavailable. Second, the attenuation of survival differences over time may reflect increasing competing mortality in this relatively older population. In our cohort, 88.2% of patients were aged ≥50 years, and a substantial proportion were aged ≥65 years. As follow-up lengthens, deaths related to age, comorbidities, or non-cancer causes may increasingly dilute the relative contribution of local therapy to OS. Third, residual confounding related to patient fitness, comorbidity burden, surgical margin status, RT technique, endocrine therapy adherence, anti-HER2 therapy, and systemic treatment details cannot be excluded. Although the effect of RT did not reach statistical significance after adjustment in the BCSS analysis, it still showed a protective direction, suggesting that the clinical value of RT may be more evident in OS than in BCSS. In addition, the relatively limited number of BCSS events in SEER data may also reduce statistical power and make potential differences insufficiently apparent. Therefore, the early survival advantage should be interpreted as a survival association rather than definitive evidence of a causal local-control effect.

From a treatment strategy perspective, BCS + RT was associated with survival outcomes comparable to MAST in selected early-stage patients; however, because recurrence data were unavailable, equivalence in local control could not be established. Its early benefit may be related to lower tumor burden, more reasonable patient selection, and the advantages of early local control brought about by RT; while with prolonged follow-up, the impact of factors such as increasing age, comorbidities, and non-tumor-related death gradually increases, making the two treatment strategies tend to approach in long-term OS. Therefore, the clinical use of these survival curves should be framed as reassurance regarding survival outcomes rather than as proof that local recurrence risks are identical between BCS + RT and MAST. Surgical decision-making should still incorporate tumor-to-breast ratio, multifocality or multicentricity, feasibility of negative margins, patient age and comorbidity, expected adherence to RT, patient preference, and multidisciplinary assessment.

In addition, the single-center cohort results of this study showed that AC of the breast had typical immunophenotypic features, ER/PR negativity, AR high expression, and concomitant high proportion of GCDFP-15 positivity. This result is consistent with previous reports and provides additional clinicopathological context for interpreting the SEER findings (1,3-5,11). These findings reinforce the biological distinctiveness of AC and support the rationale for analyzing it separately from conventional invasive ductal carcinoma or classical TNBC. However, because the institutional series included only 24 patients, it was used solely for descriptive clinicopathological supplementation and was not intended to validate the SEER-derived survival findings.

Several limitations should be acknowledged. First, the retrospective study design and the limitations of the database itself make it difficult to completely eliminate residual confounding, and the database lacks key information such as local recurrence, systemic therapy, and comorbidities. Second, because AC of the breast is rare, the number of events in some analyses was limited, especially in the later phase of the time-dependent analysis, and the estimates may have been unstable and should be interpreted with caution. Another limitation is that the SEER database does not capture disease-free survival or local-regional recurrence, preventing direct assessment of local control. Therefore, although comparable OS and BCSS were observed, these findings should not be interpreted as evidence of equivalent local control between BCS + RT and MAST. Future prospective studies incorporating recurrence-related endpoints are warranted. In addition, SEER does not provide detailed information on surgical margin status, RT dose and fields, chemotherapy regimens, anti-androgen therapy, anti-HER2 therapy, endocrine therapy adherence, or comorbidities. Although PSM was performed as a sensitivity analysis, it excluded unmatched patients and may therefore reduce statistical power or introduce selection bias in this rare-disease cohort. Accordingly, the primary interpretation of adjusted treatment effects was based on IPTW analyses using the full eligible cohort. Finally, the single-center cohort had a small sample size and was used only for descriptive supplementation; it was not sufficient to support further inferential analyses.

Despite these limitations, this study provides a comprehensive assessment of the survival outcomes of different local treatment strategies for AC of the breast based on a large population-based cohort. It also reveals potential dynamic effects over time while providing additional clinicopathological context through institutional data. From a clinical perspective, this study supports BCS + RT as a reasonable survival-based treatment option for appropriately selected early-stage patients, while recurrence-related endpoints remain necessary to confirm local-control safety. For surgeons, the main clinical implication is that BCS + RT may be presented as a legitimate option for eligible early-stage AC patients rather than treating MAST as the default approach because of the rarity or hormone receptor-negative phenotype of this tumor. Future multicenter cohorts and prospective studies incorporating recurrence-related endpoints, margin status, detailed RT parameters, and molecular features such as AR and HER2 are needed to further evaluate the local-control safety and individualized treatment strategy of BCS + RT in AC.

Conclusions

RT was associated with improved OS in patients with AC of the breast undergoing BCS. In early-stage patients eligible for breast-conserving therapy, BCS + RT achieved survival outcomes comparable to MAST, supporting its role as a reasonable survival-based local treatment option when breast preservation is clinically desired. The potential survival benefit of BCS + RT was mainly observed during early follow-up and diminished over time. Because SEER lacks recurrence, margin, and detailed treatment information, these findings should not be interpreted as evidence of equivalent local control.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://gs.amegroups.com/article/view/10.21037/gs-2026-0220/rc

Data Sharing Statement: Available at https://gs.amegroups.com/article/view/10.21037/gs-2026-0220/dss

Peer Review File: Available at https://gs.amegroups.com/article/view/10.21037/gs-2026-0220/prf

Funding: This work was supported by the Graduate Research Innovation Project of the Yunnan Provincial Department of Education (Nos. 2025Y0408, K13219855 and 2026J0299).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-2026-0220/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by The First Affiliated Hospital of Kunming Medical University (approval No. 202506-3847), and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. D'Arcy C, Quinn CM. Apocrine lesions of the breast: part 2 of a two-part review. Invasive apocrine carcinoma, the molecular apocrine signature and utility of immunohistochemistry in the diagnosis of apocrine lesions of the breast. J Clin Pathol 2019;72:7-11.
2. Horimoto Y, Kumamaru H, Orimoto K, et al. Chemotherapy is associated with differential survival outcomes in apocrine carcinoma: A nationwide cohort study from the Japanese National Clinical Database-Breast Cancer Registry. Breast 2026;87:104779. [Crossref] [PubMed]
3. Ilhan B, Emiroğlu S, Türkay R, et al. The role of histopathologic testing on apocrine carcinoma of the breast. Curr Probl Cancer 2020;44:100501. [Crossref] [PubMed]
4. Kim JY, Park S, Cho EY, et al. Genomic characteristics of triple negative apocrine carcinoma: a comparison to triple negative breast cancer. Exp Mol Med 2023;55:1451-61. [Crossref] [PubMed]
5. Basoglu T, Ozkerim U, Oksuz S, et al. Triple-Negative Apocrine Carcinoma: Largest Cohort Highlights Unique Biology and Survival Advantage. J Clin Med 2025;14:7103. [Crossref] [PubMed]
6. Sun X, Zuo K, Yao Q, et al. Invasive apocrine carcinoma of the breast: clinicopathologic features and comprehensive genomic profiling of 18 pure triple-negative apocrine carcinomas. Mod Pathol 2020;33:2473-82. [Crossref] [PubMed]
7. Mills AM, E, Gottlieb C, M, Wendroth S, et al. Pure Apocrine Carcinomas Represent a Clinicopathologically Distinct Androgen Receptor-Positive Subset of Triple-Negative Breast Cancers. Am J Surg Pathol 2016;40:1109-16. [Crossref] [PubMed]
8. Choe AI, Maley OR, Kim PM. Case of the Season: Invasive Apocrine Carcinoma of the Breast. Semin Roentgenol 2022;57:121-5. [Crossref] [PubMed]
9. Yoshino R, Hatanaka N, Takahashi K, et al. Clinicopathological features and prognostic factors of apocrine carcinoma of the breast in a cohort of 63 patients. Discov Oncol 2025;16:1743. [Crossref] [PubMed]
10. Ratosa I, Plavc G, Pislar N, et al. Improved Survival after Breast-Conserving Therapy Compared with Mastectomy in Stage I-IIA Breast Cancer. Cancers (Basel) 2021;13:4044. [Crossref] [PubMed]
11. Ke S, Wang W, Li B, et al. Superior survival for breast-conserving therapy over mastectomy in patients with breast cancer: A population-based SEER database analysis across 30 years. Front Oncol 2022;12:1032063. [Crossref] [PubMed]
12. Yu T, Cheng W, Wang T, et al. Survival Outcomes of Breast-Conserving Therapy versus Mastectomy in Early-Stage Breast Cancer, Including Centrally Located Breast Cancer: A SEER-Based Study. Breast J 2022;2022:5325556. [Crossref] [PubMed]
13. Litière S, Werutsky G, Fentiman IS, et al. Breast conserving therapy versus mastectomy for stage I-II breast cancer: 20 year follow-up of the EORTC 10801 phase 3 randomised trial. Lancet Oncol 2012;13:412-9. [Crossref] [PubMed]
14. Abdelwahed A, Ahmed M. Rare epithelial breast cancer: surgery and adjuvant therapy. Transl Cancer Res 2019;8:S479-92. [Crossref] [PubMed]
15. Rajan KK, Fairhurst K, Birkbeck B, et al. Overall survival after mastectomy versus breast-conserving surgery with adjuvant radiotherapy for early-stage breast cancer: meta-analysis. BJS Open 2024;8:zrae040. [Crossref] [PubMed]
16. Vranic S, Gatalica Z. An Update on the Molecular and Clinical Characteristics of Apocrine Carcinoma of the Breast. Clin Breast Cancer 2022;22:e576-85. [Crossref] [PubMed]
17. Li H, Tian S, Wu J, et al. Survival comparison of breast conserving therapy and mastectomy with breast reconstruction for breast cancer using propensity score matched cohort. Sci Rep 2025;15:30310. [Crossref] [PubMed]
18. Vasilyeva E, Hamm J, Nichol A, et al. Breast-Conserving Therapy is Associated with Improved Survival Without an Increased Risk of Locoregional Recurrence Compared with Mastectomy in Both Clinically Node-Positive and Node-Negative Breast Cancer Patients. Ann Surg Oncol 2023;30:6413-24. [Crossref] [PubMed]
19. Ji J, Yuan S, He J, et al. Breast-conserving therapy is associated with better survival than mastectomy in Early-stage breast cancer: A propensity score analysis. Cancer Med 2022;11:1646-58. [Crossref] [PubMed]
20. Hu T, Liu Y, Wu J, et al. Triple-Negative Apocrine Breast Carcinoma Has Better Prognosis despite Poor Response to Neoadjuvant Chemotherapy. J Clin Med 2022;11:1607. [Crossref] [PubMed]
21. Kim K, Park HJ, Shin KH, et al. Breast Conservation Therapy Versus Mastectomy in Patients with T1-2N1 Triple-Negative Breast Cancer: Pooled Analysis of KROG 14-18 and 14-23. Cancer Res Treat 2018;50:1316-23. [Crossref] [PubMed]