SMARCA4 mutations drive metastatic progression and predict poor survival in breast cancer based on a real-world cohort study Authors: Yanfang Yang, Yaoyang Guo, Mingyu Zhu, Ziyi Dong, Siqi Jiang, Xinyi Wen, Zhansheng Jiang Published: 2026 ABSTRACT To investigate the incidence, clinicopathological characteristics, and prognostic impact of SMARCA4 mutations in breast cancer, we conducted a retrospective cohort study using the MSK-CHORD database, enrolling 4,032 patients categorized into SMARCA4 mutant (1.22%, 49/4032) and wild-type groups. We analyzed overall survival (OS) via the Kaplan-Meier method and log-rank test, performed subgroup analyses by HER2/hormone receptor (HR) status, and explored associations with tumor mutational burden (TMB), microsatellite instability (MSI), and co-mutations. Results showed the mutant group had significantly higher rates of bone (75.5% vs. 55.3%, p = 0.005) and lymph node metastasis (81.6% vs. 60.9%, p = 0.003), shorter OS (p = 0.0087), and attenuated prognostic advantages of HER2-positive (p = 0.25) and HR-positive (p = 0.72) status. Additionally, the mutant group exhibited higher TMB (p < 0.05), poorer prognosis in high-MSI patients (p = 0.04), and frequent co-mutations in TP53 and PIK3CA. SMARCA4 mutation is an independent adverse prognostic factor in breast cancer (HR = 1.6, p = 0.02), linked to enhanced metastasis, altered HER2/HR prognostic value, and genomic instability, with potential as a prognostic biomarker.Supplementary InformationThe online version contains supplementary material available at 10.1007/s12672-026-05110-0. Introduction Breast cancer remains the most common malignancy among women worldwide. While mortality rates have gradually declined with advances in therapeutic strategies, significant challenges persist [1]. Traditional classification based on molecular subtypes—Luminal A, Luminal B, HER2-positive, and triple-negative—forms the cornerstone of current treatment decision-making [2]. The advent of high-throughput sequencing technologies has identified numerous driver genes and resistance loci, such as ESR1, PIK3CA, and NTRK, offering new avenues for precision therapy. Recently, the role of epigenetic regulators in tumorigenesis has garnered increasing attention, with mutations in the SMARCA4 gene being of particular interest [3–5]. The SMARCA4 gene, located on chromosome 19, encodes the BRG1 protein, a core subunit of the SWI/SNF chromatin remodeling complex [6]. The multi-subunit (10–15 subunits) SWI/SNF complex utilizes ATP to alter chromatin structure, thereby regulating gene transcription [7, 8] and influencing critical processes such as the cell cycle [9] and cell differentiation [10]. Inactivating mutations in SMARCA4 impair BRG1 function, disrupting SWI/SNF complex activity, hindering normal DNA remodeling and transcription, and have been linked to the pathogenesis and poor prognosis of various malignancies, including thoracic SMARCA4-deficient undifferentiated tumor (SMARCA4-UT) [11], ovarian cancer [12], uterine sarcoma [13], and lymphoma [14]. Within breast cancer, although the incidence of SMARCA4 alterations is relatively low (approximately < 1%−2%), its biological functions are increasingly being elucidated. Importantly, prior studies have investigated different layers of SMARCA4 biology—genomic alterations of SMARCA4 versus BRG1 (the SMARCA4-encoded ATPase) protein expression—which are not necessarily concordant because BRG1 levels can also be influenced by epigenetic regulation, post-transcriptional control, or structural variants not captured by routine sequencing. In this context, Bai et al. reported that high BRG1 protein expression correlated with poorer 5-year overall survival in breast cancer patients, and its knockdown suppressed cell proliferation and invasion [15]. Conversely, Sun et al. found that SMARCA4 depletion promoted tumor metastasis in triple-negative breast cancer (TNBC) models by activating the RHOA signaling pathway [16]. These seemingly divergent observations likely reflect context dependence of SWI/SNF function, differences in tumor subtype and baseline lineage programs (e.g., TNBC versus non-TNBC), and the fact that protein overexpression and functional loss do not represent the same biological state. Furthermore, SMARCA4 has been identified as a key hub gene associated with tamoxifen resistance [17]and is implicated in mediating cisplatin resistance in TNBC [18]. Despite preclinical evidence suggesting a critical role for SMARCA4 in breast cancer progression and drug resistance, large-scale clinical cohort analyses evaluating its prognostic impact are still limited. As the ATPase subunit of the SWI/SNF chromatin-remodeling complex, SMARCA4/BRG1 can modulate chromatin accessibility at lineage-specific enhancers and thereby influence receptor-driven transcriptional programs. Prior studies suggest BRG1 participates in estrogen receptor–related transcriptional regulation, providing a biologically plausible basis for potential interaction between SMARCA4 alterations and HR signaling [19]. In addition, SMARCA4-associated transcriptional reprogramming may shift tumors toward greater reliance on downstream oncogenic pathways (e.g., PI3K/AKT), which could attenuate the clinical prognostic stratification typically conferred by HER2 or HR status—an association that remains to be clarified in large-scale cohorts [20]. This study aims to systematically analyze the clinicopathological characteristics, metastatic spectrum, and prognostic outcomes of breast cancer patients with SMARCA4 mutations using the Memorial Sloan Kettering-Clinical History of Real-World Data (MSK-CHORD) real-world cohort, thereby providing clinical evidence to inform the precise management of this rare molecular subtype. Methods and materials Samples Patients involved in this study were obtained from Memorial Sloan Kettering-Clinical History of Real-World Data (MSK-CHORD) cohort (Nature 2024), which included data on patients’ age, sex, smoking history, histologic type, tumor mutational burden (TMB) level, microsatellite instability (MSI), site of metastasis and HER2/HR expression status. We extracted patients diagnosed with breast cancer and excluded those with incomplete clinical information (including age, survival data, stage, metastatic sites, smoking history, etc.). Based on mutation information, ultimately, 4032 patients were divided into the SMARCA4-wild type group and SMARCA4-mutant type group for subsequent analyses. Study design This study adopted a retrospective cohort study design. We analyzed and compared the impact of SMARCA4 mutation on the prognosis of breast cancer. The SMARCA4 mutation group was defined as patients harboring missense mutations, frame shift mutations, in frame mutations, nonsense mutations, and splice site mutations in the SMARCA4 gene (Supplemental Table e1). In the SMARCA4 mutation group, further subgroup analysis was conducted stratified by age, metastatic sites, TMB, and HR and HER2. The efficacy endpoint was overall survival (OS), which was defined as the time from the MSK-IMPACT sequencing report date to the date of death. Statistical analysis Categorical variables were compared using the Chi-squared test or Fisher’s exact test, and continuous variables were compared using the independent samples t-test or Wilcoxon rank-sum test. Survival outcomes were estimated by the Kaplan-Meier method, and group differences were assessed using the log-rank test. All statistical analyses were performed using R software (version 4.3.3), and P < 0.05 was considered statistically significant. Results Genomic features In this study, 49 out of 4032 patients (1.22%) had SMARCA4 mutations, and the remaining 3983 patients (98.78%) were SMARCA4 wild-type. First, we conducted a preliminary analysis of the overall mutation status in all patients and the distribution of SMARCA4 alteration (Fig. 1A). The most common Variant Classification among breast cancer patients included in the study was missense mutation, followed by frame shift deletion; the most common Variant Type was SNP single nucleotide polymorphism (SNP), followed by deletion (DEL); the top 3 most frequently mutated genes were TP53, KRAS and APC, respectively (Supplemental Fig. 1). We analyzed the genetic data of the 49 patients in the mutation group and displayed the top 20 genes by mutation rate (Fig. 1B). Except for SMARCA4 mutations, TP53 mutation was the most common (49%), followed by PIK3CA mutation (39%). Subsequently, we analyzed the expression of above the 20 genes in all breast cancers. Comparison revealed that ATR and ATRX were relatively highly mutated in the SMARCA4-mutant group, whereas FOXA1 and NF1 ranked relatively high in the wild-type group (Fig. 1C). Fig. 1Breast cancer mutation landscape. (A) Mutation types of SMARCA4 shown in a lollipop plot. (B) SMARCA4 co-mutation landscape. (C) Mutation status of co-mutated genes in all breast cancer patients Breast cancer mutation landscape. (A) Mutation types of SMARCA4 shown in a lollipop plot. (B) SMARCA4 co-mutation landscape. (C) Mutation status of co-mutated genes in all breast cancer patients Clinical information Comparison of the two groups showed that they had a similar median age, were predominantly female, and were mostly at stage I-III. The main pathological type was infiltrating duct carcinoma, the proportion of smokers was roughly the equivalent, and the expression of HER2 and HR receptors was comparable. However, the proportions of patients with bone metastasis (75.5% vs. 55.3%, p = 0.005) and lymph node metastasis (81.6% vs. 60.9%, p = 0.003) were significantly higher in the SMARCA4-mutated group than in the SMARCA4-wild-type group (Table 1). Table 1Basic characteristics between the SMARCA4 wild and mutant groupCharacteristicsSMARCA4-mutant(n = 49)SMARCA4-wild(n = 3983)pvalueAge, median60 (54, 69)59 (50, 69)0.145 Sex 0.411Female48 (98%)3941 (98.9%)Male1 (2%)42 (1.1%) Pathological Stage 0.418Stage 411 (22.4%)716 (18%)Stage 1–338 (77.6%)3267 (82%) Histology 0.412Infiltrating Duct Carcinoma41 (83.7%)2999 (75.3%)Others5 (10.2%)522 (13.1%)Lobular Carcinoma2 (4.1%)431 (10.8%)Intraductal Carcinoma1 (2%)29 (0.7%)Ductal Carcinoma0 (0%)2 (0.1%) Smoking History 0.531Never31 (63.3%)2688 (67.5%)Former/Current Smoker18 (36.7%)1295 (32.5%) Adrenal Glands 0.168No43 (87.8%)3730 (93.6%)Yes6 (12.2%)253 (6.4%) Bone 0.005No12 (24.5%)1780 (44.7%)Yes37 (75.5%)2203 (55.3%) Brain 0.966No34 (69.4%)2775 (69.7%)Yes15 (30.6%)1208 (30.3%) Intra Abdominal 0.584No40 (81.6%)3365 (84.5%)Yes9 (18.4%)618 (15.5%) Liver 0.432No24 (49%)2175 (54.6%)Yes25 (51%)1808 (45.4%) Lung 0.232No23 (46.9%)2210 (55.5%)Yes26 (53.1%)1773 (44.5%) Lymph Node 0.003No9 (18.4%)1559 (39.1%)Yes40 (81.6%)2424 (60.9%) Pleura 0.085No36 (73.5%)3299 (82.8%)Yes13 (26.5%)684 (17.2%) Reproductive Organs 1.000No45 (91.8%)3628 (91.1%)Yes4 (8.2%)355 (8.9%) HER2 0.486No39 (79.6%)3319 (83.3%)Yes10 (20.4%)664 (16.7%) HR 0.684No9 (18.4%)826 (20.7%)Yes40 (81.6%)3157 (79.3%) Table 1Basic characteristics between the SMARCA4 wild and mutant groupCharacteristicsSMARCA4-mutant(n = 49)SMARCA4-wild(n = 3983)pvalueAge, median60 (54, 69)59 (50, 69)0.145 Sex 0.411Female48 (98%)3941 (98.9%)Male1 (2%)42 (1.1%) Pathological Stage 0.418Stage 411 (22.4%)716 (18%)Stage 1–338 (77.6%)3267 (82%) Histology 0.412Infiltrating Duct Carcinoma41 (83.7%)2999 (75.3%)Others5 (10.2%)522 (13.1%)Lobular Carcinoma2 (4.1%)431 (10.8%)Intraductal Carcinoma1 (2%)29 (0.7%)Ductal Carcinoma0 (0%)2 (0.1%) Smoking History 0.531Never31 (63.3%)2688 (67.5%)Former/Current Smoker18 (36.7%)1295 (32.5%) Adrenal Glands 0.168No43 (87.8%)3730 (93.6%)Yes6 (12.2%)253 (6.4%) Bone 0.005No12 (24.5%)1780 (44.7%)Yes37 (75.5%)2203 (55.3%) Brain 0.966No34 (69.4%)2775 (69.7%)Yes15 (30.6%)1208 (30.3%) Intra Abdominal 0.584No40 (81.6%)3365 (84.5%)Yes9 (18.4%)618 (15.5%) Liver 0.432No24 (49%)2175 (54.6%)Yes25 (51%)1808 (45.4%) Lung 0.232No23 (46.9%)2210 (55.5%)Yes26 (53.1%)1773 (44.5%) Lymph Node 0.003No9 (18.4%)1559 (39.1%)Yes40 (81.6%)2424 (60.9%) Pleura 0.085No36 (73.5%)3299 (82.8%)Yes13 (26.5%)684 (17.2%) Reproductive Organs 1.000No45 (91.8%)3628 (91.1%)Yes4 (8.2%)355 (8.9%) HER2 0.486No39 (79.6%)3319 (83.3%)Yes10 (20.4%)664 (16.7%) HR 0.684No9 (18.4%)826 (20.7%)Yes40 (81.6%)3157 (79.3%) Basic characteristics between the SMARCA4 wild and mutant group Analysis of the impact of SMARCA4 mutation on patient prognosis We collected the patients’ OS data and compared the impact of SMARCA4 mutation on the prognosis of breast cancer patients (Fig. 2A). Among them, 1461/3983 participants in SMARCA4-wild group and 25/49 participants in SMARCA4-mutant group experienced outcome events. The results showed that compared with the SMARCA4-wild-type group, the SMARCA4-mutant group had shorter overall survival time, higher risk of death within the same period, and poorer prognosis (p = 0.0087). The median survival time of the SMARCA4-wild-type group was significantly longer than that of the mutant group, approximately twice that of the mutant group. Fig. 2Prognostic analysis of SMARCA4 mutations and subtypes in breast cancer patients. (A) Comparison of OS between SMARCA4 mutant type and SMARCA4 wild type. (B) Survival analysis of SMARCA4 class I mutations and class II mutations. (C) The results of multivariate regression analysis Prognostic analysis of SMARCA4 mutations and subtypes in breast cancer patients. (A) Comparison of OS between SMARCA4 mutant type and SMARCA4 wild type. (B) Survival analysis of SMARCA4 class I mutations and class II mutations. (C) The results of multivariate regression analysis SMARCA4 mutations are classified into Class I and Class II, with different mutations showing differences in the efficacy of immunotherapy. Therefore, we summarized somatic mutation data and divided SMARCA4-mutant patients into Class II mutations (with missense mutations and small in-frame deletions of unknown significance) and Class I mutations (other mutations). We then compared the OS benefits (Fig. 2B) and 3 patients in ClassI and 22 patients in Class II experienced OS events. The results showed that the OS curves for SMARCA4 Class I and Class II mutations were highly overlapping, with no significant statistical difference between them (p = 0.626). The wild type showed a difference only with Class II SMARCA4 mutations (p = 0.008), while there was no statistically significant difference in OS between the wild type and Class I mutations (p = 0.611). Multivariate regression analysis To further validate the role of SMARCA4 in breast cancer, we included SMARCA4 mutation status, TP53 mutation status, age, pathological stage, and HER2/HR status in a multivariate regression analysis (Fig. 2C). The results demonstrated that SMARCA4 mutation was an independent prognostic factor (p = 0.02, HR = 1.6, 95%CI: 1.08–2.38). In addition, TP53 mutation and pathological stage were identified as adverse prognostic factors with HR > 1, whereas HER2 and HR status were favorable prognostic factors with HR < 1; all these associations were statistically significant (p < 0.001). After adjustment, age showed no significant prognostic impact. Exploration of the relationship between MSI, TMB and SMARCA4 mutations Microsatellites refer to genomic regions formed by the repetition of short sequences with 1–6 base pairs as units. MSI is defined as changes in the length of these regions, usually caused by defects in DNA mismatch repair function. It is commonly used in tumor precision therapy and influences the overall prognosis of cancer patients as well as the efficacy of immunotherapy. We therefore collated MSI scores and compared the differences in MSI scores between the SMARCA4-mutant group and the SMARCA4-wild-type group to clarify the significance of MSI in this specific mutation type (SMARCA4 mutation). Results showed no significant difference in MSI scores between the two groups (Fig. 3A). Subsequently, we divided patients in the SMARCA4-mutant group into High-MSI and Low-MSI subgroups based on the median MSI score. Survival analysis revealed that patients with high MSI scores in the mutant group had a poorer prognosis (P = 0.04) (Fig. 3B). 8 of 24 patients in Low-MSI and 17 of 25 in High-MSI had outcome events. Fig. 3Analysis of the significance of MSI and TMB in patients with SMARCA4 mutation. (A) Difference in MSI score between SMARCA4 mutant and wild-type. (B) Influence of MSI on prognosis in SMARCA4 mutant patients. (C) Difference in TMB between SMARCA4 mutant and wild-type. (D) Influence of TMB on prognosis in SMARCA4 mutant patients Analysis of the significance of MSI and TMB in patients with SMARCA4 mutation. (A) Difference in MSI score between SMARCA4 mutant and wild-type. (B) Influence of MSI on prognosis in SMARCA4 mutant patients. (C) Difference in TMB between SMARCA4 mutant and wild-type. (D) Influence of TMB on prognosis in SMARCA4 mutant patients TMB refers to the number of somatic mutations in the tumor genome, usually expressed as the number of mutations per megabase. TMB is an important indicator for evaluating the efficacy of tumor immunotherapy [21]. We conducted the analysis using the same method as described above. Results showed that the SMARCA4-mutant group had higher TMB scores (Fig. 3C), but there was no statistically significant difference in prognosis between the High-TMB group and the Low-TMB group (p = 0.25) (Fig. 3D). 13 of 24 patients in Low-TMB and 17 of 25 patients in High-TMB had outcome events. Analysis of the interaction between HR/HER2 and SMARCA4 In breast cancer, HER2 is often used to determine tumor biological behavior, guide the selection of targeted therapy regimens, and evaluate prognosis. First, we analyzed the prognostic significance of HER2 in the overall breast cancer patient population. Results showed that HER2-positive patients had a better prognosis (p < 0.0001) (Fig. 4A). 1297/3358 participants in HER2-negative group and 189/674 participants in HER2-positive group experienced outcome events. Subsequently, we incorporated SMARCA4 as a factor for re-analysis and found that the previously observed prognostic advantage of HER2-positive patients was lost when combined with SMARCA4 mutation (p = 0.25) (Fig. 4B). 6 of 10 participants in SMARCA4-mutant and HER2-positive group had OS events. In breast cancer, HR refers to estrogen receptor (ER) and progesterone receptor (PR), which play an important guiding role in breast cancer subtyping and the selection of endocrine therapy. Therefore, we performed the same analysis for HR and obtained similar results (p < 0.0001) (Fig. 4C), SMARCA4 mutation rendered the prognosis of HR-positive patients comparable to that of HR-negative patients (p = 0.72) (Fig. 4D). 362/835 participants in HR-negative group, 1142/3197 participants in HR-positive group and 20/40 participants in HR-positive with SMARCA4-mutant group experienced outcome events. Fig. 4Analysis of the prognostic impact of SMARCA4 mutation on HR + and HER2 + patients. (A) Prognostic difference between HER2 + and HER2- patients. (B) Prognostic changes in HER2 + patients with concurrent SMARCA4 mutation. (C) Prognostic difference between HR + and HR- patients. (D) Prognostic changes in HR+ patients with concurrent SMARCA4 mutation Analysis of the prognostic impact of SMARCA4 mutation on HR + and HER2 + patients. (A) Prognostic difference between HER2 + and HER2- patients. (B) Prognostic changes in HER2 + patients with concurrent SMARCA4 mutation. (C) Prognostic difference between HR + and HR- patients. (D) Prognostic changes in HR+ patients with concurrent SMARCA4 mutation To further understand the relationship between HR, HER2, and SMARCA4 mutation, we performed a more detailed subgroup analysis. Among them, 1278/3319 participants in SMARCA4-wild with HER2- group, 183/664 participants in SMARCA4-wild with HER2+ mutant group, 6/10 participants in SMARCA4-mutant with HER2 + group and 19/39 participants in SMARCA4-mutant with HER2- group experienced outcome events. In the SMARCA4-wild-type group, the performance of HER2 was consistent with the overall population—HER2 + patients had a better prognosis (p < 0.0001) (Fig. 5A). However, in the SMARCA4-mutant group, there was no significant difference in prognosis between HER2 + and HER2- subgroups (p = 0.95) (Fig. 5B). In the HER2 + group, patients with SMARCA4 mutation had a poorer prognosis compared to those with the wild-type, which was consistent with the overall results (p = 0.017) (Fig. 5C). In the HER2- group, SMARCA4 lost this prognostic grouping ability (p = 0.056) (Fig. 5D). Fig. 5Comprehensive analysis of the interaction between HER2 and SMARCA4. (A) Impact of HER2 on prognosis in SMARCA4 wild-type patients. (B) Impact of HER2 on prognosis in SMARCA4 mutant patients. (C) Impact of SMARCA4 on prognosis in HER2 + patients. (D) Impact of SMARCA4 on prognosis in HER2- patients Comprehensive analysis of the interaction between HER2 and SMARCA4. (A) Impact of HER2 on prognosis in SMARCA4 wild-type patients. (B) Impact of HER2 on prognosis in SMARCA4 mutant patients. (C) Impact of SMARCA4 on prognosis in HER2 + patients. (D) Impact of SMARCA4 on prognosis in HER2- patients We performed the same analysis in HR subgroups. Among them, 1104/3157 participants in SMARCA4-wild with HR+ group, 357/826 participants in SMARCA4-wild with HR- mutant group, 5/9 participants in SMARCA4-mutant with HR- group and 20/40 participants in SMARCA4-mutant with HR+ group experienced outcome events. HR+ patients exhibited a prognostic advantage in the SMARCA4-wild type group (p < 0.0001) (Fig. 6A), but this advantage was absent in the SMARCA4-mutant group (p = 0.85) (Fig. 6B). Furthermore, in the HR+ subgroup, SMARCA4 wild type conferred a prognostic advantage (p = 0.004) (Fig. 6C), while this advantage disappeared in the HR- subgroup (Fig. 6D). Fig. 6Comprehensive analysis of the interaction between HR and SMARCA4. (A) Impact of HR on prognosis in SMARCA4 wild-type patients. (B) Impact of HR on prognosis in SMARCA4 mutant patients. (C) Impact of SMARCA4 on prognosis in HR+ patients. (D) Impact of SMARCA4 on prognosis in HR- patients Comprehensive analysis of the interaction between HR and SMARCA4. (A) Impact of HR on prognosis in SMARCA4 wild-type patients. (B) Impact of HR on prognosis in SMARCA4 mutant patients. (C) Impact of SMARCA4 on prognosis in HR+ patients. (D) Impact of SMARCA4 on prognosis in HR- patients Analysis of patients with SMARCA4 co-mutation Gene co-mutations often affect or alter the role of single mutations in tumors. Based on the results of the aforementioned genetic data analysis, we analyzed the co-mutations between SMARCA4 and genes with high mutation rates. Firstly, we compared the relationship in all patients (Supplemental Fig. 2 A). The results showed that TP53 mutations showed a strong correlation with metastasis to multiple anatomical sites, including the adrenal glands, bone, brain, liver, lung, lymph nodes, and pleura. SMARCA4 mutations exhibited a moderate correlation with bone and lymph node metastases. PIK3CA mutations demonstrated a strong correlation with lymph node involvement, while ATR mutations showed a weak correlation with bone and liver metastases. No notable associations were observed between GATA3 or KMT2C mutations and any of the metastatic sites analyzed. We compared the relationship between gene mutation status and metastasis in patients with SMARCA4 mutation (Supplemental Fig. 2B), and found that TP53 was significantly positively correlated with adrenal metastasis, KMT2C with genital organ metastasis, and ATR with adrenal and liver metastasis. GATA3 mutation was negatively correlated with lymph node metastasis. We then quantified the number of metastatic sites in patients and found that the SMARCA4-mutant group exhibited a higher total metastatic burden (Supplemental Fig. 2 C). We first analyzed TP53. Firstly, patients with TP53 mutation among all breast cancer patients had a poor prognosis (p < 0.0001) (Fig. 7A). Among them, 729/1630 participants in TP53-mutant group and 757/2402 participants in TP53-wild group experienced outcome events. Subsequently, we analyzed the impact of concurrent TP53 mutation in patients with SMARCA4 mutation, and it was found that there was no significant statistical difference in prognosis between the two groups (p = 0.75) (Fig. 7B). Among them, 12/24 participants in SMARCA4-mutant and TP53-mutant group, 13/25 participants in SMARCA4-mutant and TP53-wild group experienced outcome events. Then, we screened out patients with TP53 mutation and compared the impact of SMARCA4 mutation status on these patients; the comparison results did not show any differences (p = 0.5) (Fig. 7C). Finally, we compared the prognosis of SMARCA4-mutant and SMARCA4-wild-type patients in the TP53 wild-type subgroup. The results showed that SMARCA4 mutation remained associated with poorer prognosis in the absence of TP53 co-mutation (p = 0.0077) (Fig. 7D). Among them, 717/1606 participants in TP53-mutant with SMARCA4-wild group and 744/2377 participants in TP53-wild with SMARCA4-wild group experienced outcome events. Later analysis targeting PIK3CA also failed to yield statistically significant results (Supplemental Fig. 2D-E). Fig. 7Analysis of SMARCA4 co-mutation. (A) Role of TP53 in breast cancer. (B) Role of TP53 in SMARCA4 mutant patients. (C) Impact of SMARCA4 on prognosis in TP53 mutant patients. (D) Impact of SMARCA4 on prognosis in TP53 wild patients Analysis of SMARCA4 co-mutation. (A) Role of TP53 in breast cancer. (B) Role of TP53 in SMARCA4 mutant patients. (C) Impact of SMARCA4 on prognosis in TP53 mutant patients. (D) Impact of SMARCA4 on prognosis in TP53 wild patients Discussion This large-scale real-world cohort study systematically evaluated the prognostic relevance of SMARCA4 mutation in breast cancer, with an observed incidence of 1.22%. We found that SMARCA4 mutation was associated with shorter OS and a higher metastatic burden. We also observed that the prognostic patterns typically linked to HER2-positive and HR-positive status appeared attenuated among patients harboring SMARCA4 mutation. The following discussion integrates our results with existing literature to explore the clinical significance and potential biological mechanisms. Our data supportSMARCA4 mutation as an adverse prognostic indicator in breast cancer, consistent with observations in other cancers [11–14]. Importantly, we found that SMARCA4 mutation correlated with a higher metastasis burden, suggesting that SMARCA4 inactivation may enhance tumor invasiveness and metastatic potential. The study by Sun et al. provides a mechanistic basis for this clinical observation: in TNBC models, SMARCA4 knockdown led to decreased ARHGAP29, hyperactivation of RHOA signaling, disrupted cell adhesion, and promoted loose tumor spheroid growth and in vivo metastasis [16]. Our clinical data strongly align with these experimental findings, indicating that SMARCA4 mutation-driven defects in cell adhesion and increased motility are likely key factors contributing to the higher propensity for distant metastasis in these patients. We classified SMARCA4 mutations based on the presence of missense mutations and small in-frame deletions of unknown significance. The OS curves for Class 1 and Class 2 mutations were very close, without a statistically significant difference. However, a significant difference in OS was observed between the wild-type and Class 2 SMARCA4 mutations (p = 0.008), but not between wild-type and Class 1 mutations (p = 0.611). This pattern should be interpreted cautiously, as the Class 1 group had limited sample size and few events, which may reduce statistical power. Missense mutations are the most common SMARCA4 mutation type (~ 50%), constituting the majority of Schoenfeld-defined Class 2 mutations [22]. Although less frequently leading to loss of protein expression, studies show that missense mutations can cause SWI/SNF remodeling dysfunction and reduced chromatin accessibility [23]. Nonsense and frameshift mutations, constituting Class 1 mutations and often associated with protein loss, account for 22% and 12% of cases, respectively [22]. SMARCA4 mutation and protein loss are not perfectly correlated; 40%−80% of SMARCA4-mutant NSCLCs show loss of BRG1 protein [24, 25], while 35% of BRG1-deficient patients lack detectable SMARCA4 mutations, potentially due to undetected epigenetic changes or chromosomal rearrangements. An additional finding was that, within the SMARCA4-mutant subgroup, the survival advantage commonly associated with HER2-positive and HR-positive status was not clearly observed. This implies that for patients carrying a SMARCA4 mutation, the tumor’s biological behavior may resemble more aggressive subtypes, irrespective of conventional Luminal or HER2-positive classification. But because this small subgroup (n = 49) may reflect limited power rather than true absence of effect, we interpret this as a signal of potential effect modification rather than evidence that SMARCA4 mutation eliminates HER2/HR prognostic benefit. Biologically, one plausible explanation is that SMARCA4 deficiency may be associated with altered epigenetic programs and cell-state plasticity. Epigenomic disruption is one of the core characteristics of breast cancer [26]. Zhou et al. demonstrated in TNBC that the long non-coding RNA TGFB2-AS1 interacts with SMARCA4 protein, suppressing TGFβ signaling and stemness-related genes like SOX2, thereby inhibiting cancer stem cell properties [27]. Loss of SMARCA4 function may lift this suppression, potentially leading to dedifferentiation and enhanced stemness. Similarly, single-cell RNA sequencing by Bhat-Nakshatri, P. et al. revealed that BRCA1/2 mutations (whose proteins interact with the SWI/SNF complex) perturb signaling pathways in mammary epithelial cells, with BRCA1 mutation involving SMARCA4-related pathways. This suggests that SMARCA4 deficiency, as a core chromatin remodeler, might render cells more susceptible to oncogenic signals like PIK3CA (a frequent co-mutation in our study, 39%), driving a more aggressive phenotype [28, 29]. Thus, SMARCA4 mutation maybe promotes cell plasticity, stemness, and hypersensitivity to oncogenic signals, overriding the relatively “indolent” tumor behavior dictated by dependence on HER2 or HR signaling pathways. The higher TMB observed in the SMARCA4-mutant group, together with the association of high MSI with poorer outcomes within this subgroup, suggests a relationship between SMARCA4 alteration and genomic instability. This is consistent with mechanistic work by Zhu et al. in ovarian cancer showing that SMARCA4 loss promotes Pol II pausing, R-loop accumulation, and functional impairment of BRCA1-mediated repair, creating a “BRCA-deficient-like” state [30]. While this mechanism has not been demonstrated in breast cancer, it provides a testable hypothesis for future studies aimed at explaining the genomic features of SMARCA4-mutant tumors. Future directions. Our study did not evaluate treatment response; therefore, therapeutic implications are speculative. Nevertheless, prior literature suggests potentially actionable vulnerabilities in SMARCA4-deficient contexts, including synthetic-lethal approaches targeting SMARCA2, EZH2 inhibition, and DNA-repair–oriented strategies such as PARP inhibition [31–34]. These hypotheses require prospective clinical validation or dedicated translational studies in breast cancer before clinical application can be recommended. Our co-mutation analysis identified TP53 and PIK3CA as the most frequently co-mutated genes in SMARCA4-mutant breast cancer. Although TP53 co-mutation did not show an additional prognostic impact in this cohort, this may reflect the dominant adverse association of the SMARCA4 mutation itself. Schwartz et al. noted that SMARC alterations in breast cancer are associated with complex genomic landscapes and may correlate with specific histologies (e.g., rhabdoid features) [35]. In addition, the study by Mehta et al. demonstrated that SOX4 and SMARCA4 cooperatively regulate the PI3K/Akt signaling pathway and may play a crucial role in breast cancer [20]. Future studies with larger samples are needed to dissect how different co-mutation combinations influence the clinicopathological features and treatment responses of SMARCA4-mutant breast cancer. This study has limitations: its retrospective design introduces selection bias; missing treatment details prevent assessment of predictive value; and the low mutation frequency limits power for subgroup and interaction analyses. Future work should prioritize: prospective validation of prognostic associations; mechanistic studies elucidating links to metastasis and subtype-dependent outcomes; and clinical/translational efforts to test whether SMARCA4 status can guide therapy selection. In conclusion, SMARCA4 mutation defines a clinically meaningful subset of breast cancer characterized by association with higher metastatic burden, genomic instability features, and poorer survival. We additionally observed that conventional HER2/HR prognostic patterns were less distinct within the SMARCA4-mutant subgroup, a finding that should be interpreted cautiously given sample size and requires confirmation. Collectively, these results support further investigation of SMARCA4 alteration as a prognostic marker and as a potential stratifier for future translational studies. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplemental Fig. S1 Gene mutation status in all breast cancer patients. Figure S2 A) Relation between gene mutation and metastasis in all patients. B) Relation between gene mutation and metastasis in SMARCA4 mutant patients C) The number of metastasis sites between two groups (*p< 0.05). D) Role of PIK3CA in SMARCA4 wild group. E) Role of PIK3CA in SMARCA4 mutant patients. 533/1402 participants in PIK3CA mutant group, 953/2630 participants in PIK3CA-wild group, 16/30 in PIK3CA-wild with SMARCA4-mutant group and 9/19 in PIK3CA-mutant with SMARCA4-mutant group experienced outcome events. Supplementary Material 1. Supplementary Material 2. Supplemental Fig. S1 Gene mutation status in all breast cancer patients. Figure S2 A) Relation between gene mutation and metastasis in all patients. B) Relation between gene mutation and metastasis in SMARCA4 mutant patients C) The number of metastasis sites between two groups (*p< 0.05). D) Role of PIK3CA in SMARCA4 wild group. E) Role of PIK3CA in SMARCA4 mutant patients. 533/1402 participants in PIK3CA mutant group, 953/2630 participants in PIK3CA-wild group, 16/30 in PIK3CA-wild with SMARCA4-mutant group and 9/19 in PIK3CA-mutant with SMARCA4-mutant group experienced outcome events.