TP53 mutation is associated with shorter treatment duration with CDK4/6 inhibitors, but not with immune checkpoint inhibitors, in ER-positive/HER2-negative breast cancer: real-world perspective Authors: Masanori Oshi, Makoto Sugimori, Kei Kawashima, Shipra Gandhi, Akimitsu Yamada, Kazutaka Narui, Takashi Ishikawa, Itaru Endo, Kazuaki Takabe Published: 2026 ABSTRACT AbstractBackgroundTP53 mutation is a critical driver of breast cancer, yet its relationship with treatment outcomes for breast cancer remains unclear. This study investigates the association between TP53 mutation and response to CDK4/6 inhibitors (CDK4/6i) and immune checkpoint inhibitors (ICI) in breast cancer using real-world data from the Center for Cancer Genomics and Advanced Therapeutics (C-CAT) cohort, focusing on age-related differences.MethodsMETABRIC (n = 1355) was used for initial survival analysis, and gene set enrichment analysis (GSEA) was performed to identify pathways enriched in TP53-mutated ER-positive/HER2-negative breast cancers. Clinical and genomic data of metastatic breast cancer (mBC) patients in Japan were analyzed in the C-CAT (n = 1668) cohort. Treatment duration was used as a real-world surrogate endpoint reflecting treatment exposure and potential clinical benefit. Patients were stratified by age into four groups: Adolescents and Young Adults (AYA; 15–39 years), perimenopausal (40–54 years), menopausal (55–64 years), and older (≥ 65 years).ResultsTP53 mutations were associated with poor prognosis in the METABRIC cohort. GSEA showed TP53-mutated tumors were enriched with gene sets related to cell proliferation and immune response, while TP53 wild-type tumors enriched for estrogen response pathways. In the C-CAT cohort, TP53 mutations were linked to shorter CDK4/6i treatment duration with similar trends observed for both abemaciclib and palbociclib. However, TP53 mutations were not associated with ICI treatment duration. Patients in the AYA group had the shortest treatment duration, with TP53 mutation prevalence decreasing with age. In age-stratified analyses, TP53 mutation was associated with shorter treatment duration in perimenopausal, menopausal, and older groups but not in the AYA group. However, formal interaction testing did not detect a significant interaction between TP53 mutation and age group.ConclusionTP53 mutation is associated with shorter CDK4/6i treatment duration, highlighting the importance of considering both genomic and age-related clinical context when interpreting treatment duration in ER-positive/HER2-negative breast cancer.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13058-026-02305-9.Keywords: Breast cancer, TP53, CDK4/6 inhibitor, Immune checkpoint inhibitor, Adolescents and young adults, Treatment duration Introduction Estrogen receptor (ER)-positive/HER2-negative breast cancer represents the majority of breast cancer cases and is generally associated with a relatively favorable prognosis with current therapies [1]. While CDK4/6 inhibitors (CDK4/6i), including palbociclib, ribociclib, and abemaciclib, have shown substantial improvements in progression-free survival by targeting CDK4 and CDK6 [2], key regulators of the cell cycle, long-term outcomes remain suboptimal. The majority of patients eventually develop resistance to treatment, resulting in recurrence and disease progression [3]. Despite their effectiveness in the first-line setting, resistance to these agents is inevitable, and the mechanisms underlying this resistance are not fully understood [4]. Emerging evidence suggests that TP53 mutations, one of the most common genomic alterations in ER-positive/HER2-negative breast cancer, have been implicated in reduced sensitivity to treatment, including resistance to CDK4/6 inhibitors. TP53, a tumor suppressor gene, regulates the cell cycle, DNA repair, and apoptosis [5, 6]. The loss of TP53 function is associated with uncontrolled cell division and genomic instability, which are key features of cancer progression and resistance to multiple therapeutic agents [7, 8]. TP53 mutations have been widely reported to confer resistance to conventional chemotherapies [9]. Recently, several targeted therapies have been FDA approved, including CDK4/6i for patients with HR positive breast cancer and ICI for patients with MSI high or TMB high breast cancer. However, the association of TP53 mutation with clinical outcomes of CDK4/6i and ICI, remains unclear. Recent studies suggest that TP53 loss leads to CDK4/6i resistance [10]. Although much of the current understanding of TP53 mutations in CDK4/6i resistance is based on preclinical studies, clinical data validating these findings in real-world settings are limited. In addition to genetic mutations, age-related factors play a significant role in treatment response in ER-positive/HER2-negative breast cancer. The peak incidence of ER-positive/HER2-negative breast cancer occurs in patients aged 40–60 years in Asia, with an increasing proportion of older patients [11]. The treatment response in ER-positive/HER2-negative breast cancer is significantly influenced by the host’s hormonal status, which can vary with age, as we have previously reported [12]. Immune signatures have shown significantly higher enrichment in TP53-mutated breast cancers than in TP53 wild-type breast cancers. Moreover, in vitro experiments have confirmed that mutant p53 could increase breast cancer immunogenicity [13]. However, data also exists suggesting that TP53 mutations may influence response to immune-based therapies, such as, immune checkpoint inhibitors, by driving immune evasion of tumor cells. p53 mutant tumors could evade immune surveillance from T cells, natural killer (NK cells), and macrophages [14]. Given these findings, it is critical to examine the association of TP53 mutations with outcomes to immune checkpoint inhibitors in breast cancer. Importantly, TP53 mutation frequency and spectrum vary across ethnic populations. Japanese breast cancers, despite the lower overall incidence of the disease, exhibit a high prevalence of TP53 mutations and distinct mutation patterns compared with Western populations [15]. Such inter-ethnic differences have also been reported to affect treatment outcomes and toxicity profiles, emphasizing the need for population-specific analyses to inform precision oncology. To address these gaps, we investigated the association between TP53 mutations and treatment outcomes in ER-positive/HER2-negative breast cancer using two complementary approaches. First, we evaluated the clinicopathological features and transcriptomic signatures of TP53 mutations in primary tumors using the METABRIC cohort. Next, we examined the association of TP53 mutations with real-world treatment outcomes using the Japanese C-CAT cohort, a unique real-world data of metastatic breast cancer (mBC) patients in Japan, leveraging treatment duration as a surrogate for endpoint for systemic therapies, including CDK4/6i and immune checkpoint inhibitors (ICI) treatment duration. This multi-step approach, combining biological insights with real-world clinical data, provides an opportunity to better understand the clinical and biological implications of TP53 mutations in ER-positive/HER2-negative breast cancer. Our study uniquely focuses on the relationship between TP53 mutations and CDK4/6i treatment in the several age groups, especially Adolescents and Young Adults (AYA) population, an area that has not been previously explored. This novel contribution to the field adds valuable insights into how TP53 mutations influence treatment outcomes across different age patient groups with ER-positive/HER2-negative breast cancer. Materials and methods Study design and patient selection The METABRIC (Molecular Taxonomy of Breast Cancer International Consortium) dataset [16], along with additional data from cBioPortal [17], was used to perform a biological analysis of TP53 mutation in primary breast cancer. ER and HER2 status were defined based on the clinical annotations provided in the dataset. TP53 mutation status was determined using the available mutation data in METABRIC. The METABRIC cohort includes a large collection of clinical, genomic, and transcriptomic data from breast cancer patients, providing a comprehensive resource for investigating the molecular characteristics of the disease [18]. This retrospective observational study analyzed data from patients with mBC who underwent comprehensive genomic profiling (CGP) test using tissue samples by FoundationOne CDx (Foundation Medicine, Boston, MA, USA) between June 2019 and November 2024, during which the database was established. Data regarding patient age, administered drugs, treatment courses, and CGP test results were extracted from the C-CAT database [19, 20]. Only ER-positive/HER2-negative mBC patients with a history of hormone therapy who were treated with CDK4/6i (abemaciclib or palbociclib), or ICI (pembrolizumab) were included in the current study. ER and HER2 status were obtained from the C-CAT dataset and used as recorded. The line of therapy was defined based on the sequence of systemic treatments as recorded in the C-CAT database. Because radiographic progression data and reasons for treatment discontinuation were not uniformly available in the C-CAT dataset, we used treatment duration, defined as the interval from treatment initiation to the final recorded administration, as a real-world surrogate endpoint. Genomic profiling Mutation variants were cross-referenced using ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/) and OncoKB (https://www.oncokb.org) databases to assess their pathogenic potential. Genetic alterations were classified into two categories; truncating changes (including loss-of-function, splice-site mutations, and nonsense mutations), and single-nucleotide variants (SNVs), which include missense mutations. Functional analysis Gene Set Enrichment Analysis (GSEA) was performed using the Molecular Signatures Database Hallmark gene set [21, 22] to investigate the differentially enriched gene sets in TP53-mutated tumors and wild-type tumors in the METABRIC cohort. The statistical significance of the gene sets was determined based on False Discovery Rate (FDR) < 0.25, as recommended [21]. Statistical analysis All statistical analyses were conducted using R software (version 4.2.3). Kaplan–Meier survival curves were generated to visualize survival and treatment duration differences of TP53 mutation status. Log-rank tests were used to compare the survival curves. To assess the impact of hormonal status on treatment duration, patients were stratified into four age groups to reflect the influence of hormonal status: adolescent and young adult (AYA, age ≤ 39), perimenopausal (age 40–54), menopausal (age 55–64), and older (age ≥ 65). For comparisons between two groups, Mann-Whitney U test or Fisher’s exact test were used. Ethical considerations Both METABRIC and the C-CAT cohorts consist of de-identified data. The METABRIC dataset is publicly available and therefore exempt from ethical approval. The use of C-CAT data in this study was reviewed and approved by the Ethical Advisory Committee of the YCU (IRB approval number: F250900014), in accordance with Japanese ethical guidelines. Results Prognostic impact and biological characterization of TP53 mutations in ER-positive/HER2-negative breast cancer To investigate the clinical implications of TP53 mutation in ER-positive/HER2-negative breast cancer, we first compared the prognosis of TP53-mutated versus wild-type primary tumors in the METABRIC cohort (n = 1,355). Disease-free survival (DFS), disease-specific survival (DSS), and overall survival (OS) were all significantly worse in the TP53 mutation group compared to the wild-type group (Fig. 1A, DFS: p < 0.0001 [median survival: 20.5 years vs. NR], DSS: p < 0.0001 [18.8 vs. 24.7 years], OS: p = 0.0005 [10.4 vs. 13.9 years]). Next, gene set enrichment analysis (GSEA) was performed to identify biological pathways enriched in TP53-mutated tumors. The analysis revealed significant enrichment of cell proliferation-related gene sets and immune-related gene sets in the TP53 mutation group (Fig. 1B, FDR < 0.25). In contrast, the TP53 wild-type group showed enrichment of early estrogen response pathway alone (Fig. 1C). These findings suggest that TP53 mutations are associated with tumor progression and enrichment of cell proliferation and immune microenvironment pathways, while TP53 wild-type tumors are characterized by estrogen-related signaling, highlighting the potential for therapeutic strategies targeting these distinct molecular alterations. Survival outcomes and pathway enrichment analysis in TP53-mutated and wild-type ER-positive/HER2-negative breast cancer. A Kaplan–Meier survival curves comparing disease-free survival (DFS), disease-specific survival (DSS), and overall survival (OS) between TP53-mutated and wild-type tumors in the METABRIC cohort (n = 1355). B Bar plots showing the gene sets significantly enriched in the TP53-mutated group. (C) Gene set enrichment analysis (GSEA) enrichment plot showing enrichment of the Estrogen response early signaling pathway in the TP53 wild-type group Association of TP53 mutation with treatment duration of CDK4/6i and ICI in ER-positive/HER2-negative breast cancer Given the results that TP53-mutation tumors showed significant enrichment of multiple immune-related and cell proliferation-related signaling pathways in ER-positive/HER2-negative breast cancer, we examined whether TP53 mutation status correlated with treatment duration for immune checkpoint inhibitors (ICIs) or CDK4/6 inhibitors (CDK4/6i). To assess this hypothesis, we analyzed the C-CAT cohort, which includes clinical and genomic data of metastatic breast cancer (mBC) patients in Japan. Contrary to our expectations, we found that no association was detected between TP53-mutated (n = 185) and wild-type (n = 30) tumors in terms of ICI treatment duration (Fig. 2A, p = 0.77). In contrast, TP53 mutation was significantly associated with shorter treatment duration for CDK4/6i (Fig. 2A, p < 0.0001 [median: 6.97 vs. 9.53 months]). Given these findings, we sought to further investigate the relationship between TP53 mutation and CDK4/6i treatment duration. Considering the potential impact of treatment line on CDK4/6i efficacy, we stratified patients by the line of CDK4/6i initiation and performed a sub-analysis. Patients were classified into three subgroups: 1–3 lines, 4–7 lines, and 8 or more lines of treatment. In the 1–3-line subgroup, TP53 mutation was significantly associated with shorter treatment durations (Fig. 2B, p = 0.002 [median: 7.4 vs. 9.0 months]). In the 4–7-line subgroup, the association was even stronger, with TP53 mutation linked to a significantly shorter treatment duration (p < 0.0001 [median: 7.0 vs. 10.7 months]). However, in the 8 or more lines subgroup, a trend toward shorter treatment duration was observed, but this difference did not reach statistical significance (Fig. 2B, p = 0.068 [median: 3.6 vs. 7.2 months]). These results suggest that TP53 mutations are associated with shorter treatment durations in CDK4/6i therapy, particularly during the earlier treatment lines, but not in those with heavily pretreated tumors. However, no association was detected between TP53 mutation and ICI treatment duration in this cohort. Given the small size of the TP53 wild-type group and the real-world use of ICI in biomarker-selected cohort in later line settings, this finding should be interpreted with caution. To further account for potential confounding, we performed a multivariable Cox proportional hazards analysis. TP53 mutation remained significantly associated with shorter CDK4/6i treatment duration after adjustment for age, treatment line, and CDK4/6 inhibitor type (Figure S1, HR, 1.40; 95% CI 1.25–1.56; p < 0.001). Relationship between TP53 mutation and treatment duration of CDK4/6i and ICI in ER-positive/HER2-negative breast cancer. A Kaplan–Meier survival curves comparing treatment durations between TP53 mutation and wild-type groups for ICI (left; n = 185 and n = 30, respectively) and CDK4/6i (right; n = 519 and n = 765, respectively) in ER-positive/HER2-negative breast cancer in C-CAT cohort. B Kaplan-Meier survival curves of CDK4/6i treatment duration, stratified by treatment line: 1–3 lines (n = 253 and n = 357, respectively), 4–7 lines (n = 243 and n = 351, respectively), and 8 or more lines (n = 28 and n = 61, respectively). Y-axis represents the proportion of patients remaining on treatment, and X-axis represents treatment duration (months) Association of TP53 mutation with treatment duration of abemaciclib and palbociclib in ER-positive/HER2-negative breast cancer To further investigate the impact of TP53 mutation on treatment duration, we conducted separate analyses for abemaciclib and palbociclib, and the fact that ribociclib is not approved for use in Japan yet. Both abemaciclib and palbociclib treatments showed significantly shorter treatment durations in the TP53 mutation group compared to wild-type (Fig. 3, abemaciclib, p = 0.0002 [median: 6.8 vs. 7.9 months]; palbociclib, p = 0.002 [median; 6.6 vs. 10.1 months]). Next, we conducted line-based subgroup analyses for each drug, similar to the analysis in Fig. 2. In the abemaciclib group, we found that TP53 mutation was significantly associated with shorter treatment duration in both the 1–3 lines (p = 0.044 [median: 6.5 vs. 7.5 months]) and 4–7 lines (p = 0.00054 [median: 7.0 vs. 9.1 months]) subgroups, but not in the 8 or more lines subgroup (p = 0.39 [median: 5.9 vs. 6.0 months]), showing consistent results with the overall analysis. In contrast, palbociclib treatment showed a significant difference in treatment duration only in the 4–7 lines subgroup (p = 0.0061 [median: 6.5 vs. 10.3 months]), while no significant differences were found in the 1–3 lines (p = 0.14 [median: 7.4 vs. 10.5 months]) and 8 or more lines (p = 0.28 [median: 1.9 vs. 6.1 months]) subgroups. These results suggest that TP53 mutation was associated with shorter treatment durations with both abemaciclib and palbociclib. While the pattern of association varied across treatment lines, these subgroup analyses were exploratory and should be interpreted with caution. Treatment duration of abemaciclib and palbociclib stratified by TP53 mutation status and treatment line in ER-positive/HER2-negative breast cancer. Kaplan–Meier survival curves for treatment duration of abemaciclib (top; n = 822, TP53-mutated, n = 352; and wild-type, n = 470) and palbociclib (bottom; n = 415, TP53-mutated n = 154 and wild-type n = 261) in ER-positive/HER2-negative breast cancer patients, stratified by TP53 mutation status. Left panels show the overall population, while the subsequent panels represent the subgroups based on treatment line: 1–3 lines, 4–7 lines, and 8 or more lines. Y-axis represents the proportion of patients remaining on treatment, and X-axis represents treatment duration (months) No significant association observed between TP53 mutation type or pathogenicity and treatment duration, but co-occurring mutations in PI3K, GATA3, and ESR1 altered the genetic landscape Given the potential biological differences among different TP53 mutations and the clinical relevance of pathogenicity in mutation-driven treatment outcomes, we investigated the outcomes by TP53 mutation type on treatment duration. We compared truncation and single nucleotide variant (SNV) mutations, as well as classified mutations into pathogenic and non-pathogenic based on ClinVar annotations. We found that no significant difference in treatment duration between TP53 truncating and SNV mutations (Fig. 4A, p = 0.12) was observed. Similarly, no significant difference was observed when comparing pathogenic versus non-pathogenic mutations (Fig. 4B, p = 0.50). Additionally, we assessed the co-occurrence of TP53 mutation with other commonly observed genetic alterations in ER-positive/HER2-negative breast cancer. PI3K mutation was significantly more frequent in the TP53 mutation group, whereas GATA3 and ESR1 mutation were more frequently observed in the wild-type group. These findings suggest that TP53 mutation is associated with specific co-mutation patterns, reflecting differences in the broader genomic landscape of the disease. Association of TP53 mutation type with treatment duration. A Kaplan–Meier curve comparing treatment duration between TP53 truncating mutation (n = 308) and single nucleotide variant (SNV) (n = 71) mutation groups. B Kaplan–Meier curve comparing treatment duration based on ClinVar classification of pathogenic (n = 320) and non-pathogenic (n = 204) TP53 mutation. C Bar plot showing the frequency of co-occurring mutations in TP53-mutated and wild-type tumors by different mutations (PIK3CA, GATA3, BRCA2, NOTCH3, ESR1, KMT2D, LTK, SPEN, NOTCH1, MAP3K1) Association of age and TP53 mutation on CDK4/6i treatment duration in ER-positive/HER2-negative breast cancer Given the critical role of the host’s hormonal status in influencing treatment response in ER-positive/HER2-negative breast cancer, we analyzed the association between TP53 mutation on CDK4/6i treatment duration across different age groups: Adolescent and Young Adults (AYA, 15–39 years), perimenopausal (40–54 years), menopausal (55–64 years), and older (65 years and above) groups (Table S1). We observed significant differences in treatment duration with CDK4/6i across these groups. Overall, menopausal patients tended to have longer treatment durations on CDK4/6i, while the AYA group showed significantly shorter treatment durations (Fig. 5A, p = 0.0019). To further investigate this, we analyzed the timing of CDK4/6i initiation across age groups. The AYA group was found to have a significantly lower proportion of patients starting CDK4/6i in the later treatment lines compared to the other groups (Fig. 5B, p = 0.026). To understand the rationale behind these observed differences, we assessed the prevalence of TP53 mutation in each age group. We found a consistent decline in TP53 mutation prevalence with increasing age, with the AYA group having a significantly higher proportion of TP53 mutation compared to the other age groups (Fig. 5C). Finally, we compared the CDK4/6i treatment duration between TP53-mutated and wild-type groups within each age group. While no significant difference was observed in the AYA group, a significant shortening of treatment duration in TP53-mutated tumors was observed in the perimenopausal, menopausal, and older groups (Fig. 5D, p = 0.28, p < 0.0001, p = 0.002, p = 0.012, respectively). These results indicate age-related differences in treatment duration and TP53 mutation prevalence in this cohort. However, formal interaction testing did not detect a significant interaction between TP53 mutation and age group (likelihood ratio test, p = 0.89). Age-related differences in treatment duration and TP53 mutation frequency in ER-positive/HER2-negative breast cancer. A Kaplan–Meier curve comparing CDK4/6i treatment duration across different age groups: adolescent and young adults (AYA), perimenopausal, menopausal, and older. B Bar plot showing the distribution of CDK4/6i treatment lines across age groups. C Bar plot showing the prevalence of TP53 mutation across age groups. D Kaplan–Meier curve comparing CDK4/6i treatment duration between TP53-mutated and wild-type groups across different age groups. Numbers of patients in each age group and TP53 subgroup are provided in supplementary Table S1 Discussion In this study, we investigated the relationship between TP53 mutation and treatment duration in ER-positive/HER2-negative breast cancer using real-world data from METABRIC and C-CAT cohort. TP53 mutation was significantly associated with poor prognosis in the METABRIC cohort, where patients with TP53-mutated tumors exhibited worse DFS, DSS, and OS. Functional analysis revealed that TP53-mutated tumors enriched cell proliferation and immune-related gene sets. When investigating the relationship between TP53 mutation and treatment duration in CDK4/6i using C-CAT cohort, we found that TP53 mutation was significantly associated with shorter treatment durations for CDK4/6i, but not ICIs. This association remained significant after adjustment for available clinical variables. To further understand the impact of TP53 mutation on CDK4/6i treatment duration, we stratified patients by treatment line. TP53 mutation was associated with significantly shorter treatment durations in the 1–3 line and 4–7 line groups. However, no significant difference was observed in patients who received CDK4/6i after 8 lines of treatment, that is, heavily pretreated patients. In addition, we examined the association between TP53 mutations and treatment duration separately for abemaciclib and palbociclib. Although shorter treatment duration was observed in TP53-mutated tumors in both groups, the pattern varied by type of CDK4/6 inhibitor and across treatment lines, and these subgroup analyses were exploratory. We found no significant differences between TP53 truncating mutation and SNVs and their association with treatment duration on CDK4/6i. TP53 mutation classified as pathogenic did not significantly differ from non-pathogenic mutations in terms of their relationship with CDK4/6i treatment duration. Finally, age-related factors were found to play a significant role in treatment outcomes. TP53 mutation was most prevalent in the AYA group, and treatment durations with CDK4/6i were shorter in AYA compared with the other age groups. In this study, we observed a significant association between TP53 mutation and shorter treatment durations with CDK4/6i. Notably, the differences in treatment duration were most pronounced during the mid-treatment lines, particularly 4–7 lines of treatment in metastatic setting. While we did not observe statistically significant differences in the later lines of treatment, a consistent trend toward shorter treatment duration remained, indicating that tumor biology may evolve with prolonged treatment exposure. Interestingly, shorter treatment duration in TP53-mutated tumors was observed in both abemaciclib and palbociclib groups. However, the pattern of association varied across treatment lines. Because these analyses were exploratory and not based on formal interaction testing, potential differences between CDK4/6 inhibitors should be interpreted with caution. In contrast, no association was detected between TP53 mutation and ICI treatment duration in this cohort. However, the TP53 wild-type group was relatively small compared with the mutation group, likely reflecting the real-world use of ICI in HR-positive/HER2-negative breast cancer, where patients are typically selected by the biomarker. Therefore, this finding should be interpreted with caution, and further studies are required to clarify this relationship. While no significant differences were observed between truncation and SNV mutations, or between pathogenic and non-pathogenic mutations, these findings suggest that the presence of TP53 mutation itself, rather than mutation subtype, may be associated with differences in treatment duration. Specifically, TP53 loss may be a key factor influencing the response to CDK4/6i by impairing critical regulatory pathways involved in cell cycle progression. Existing studies have suggested that TP53 deficiency can promote cell cycle progression through CDK2-mediated phosphorylation of p130, leading to disruption of the DREAM complex and potential bypass of CDK4/6 inhibition [26, 27]. These findings provide a potential biological framework that may help explain the shorter treatment duration observed in TP53-mutated tumors in our study. Furthermore, amplification of MDM2 has been shown to be associated with reduced sensitivity to CDK4/6i, highlighting the potential role of the TP53/MDM2 axis in modulating treatment response [28]. However, in this study, we only differentiated between truncation and SNV mutations, which limits the granularity of mutation classification. A more refined classification of TP53 mutations may reveal additional associations with treatment outcomes. Therefore, further studies incorporating detailed classification of TP53 mutations and their functional impact will be important to better understand the biological basis of treatment response. Beyond mutation-level classification, recent studies have highlighted additional and complementary mechanisms by which TP53 regulates ER signaling and luminal cell identity [29]. In our analysis, GSEA demonstrated that estrogen response pathways were significantly enriched in TP53 wild-type tumors compared to TP53-mutated tumors, suggesting a greater reliance on ER-driven transcriptional programs in the wild-type setting. Consistent with this observation, Aylon et al. demonstrated that wild-type p53 transcriptionally regulates chromatin modifiers that facilitate chromatin accessibility at ER-binding sites, thereby promoting ER-dependent transcriptional programs and maintaining luminal differentiation in breast cancer. In this context, TP53 mutation may impair ER-driven transcription and reduce reliance on canonical estrogen-dependent cell cycle progression, including the cyclin D-CDK4/6 axis [29]. These findings suggest that TP53 mutation may promote resistance to CDK4/6 inhibitors through multiple, non-mutually exclusive mechanisms, including both disruption of cell cycle control via the DREAM complex and altered epigenetic and transcriptional regulation of ER signaling. This integrated framework may help explain the reduced clinical benefit of CDK4/6 inhibitors observed among TP53-mutated tumors in our study. Consistent with this biological complexity, we did not observe significant differences in treatment duration based on TP53 mutation type or pathogenicity classification. This may indicate that simple categorization of TP53 alterations is insufficient to capture the functional heterogeneity of TP53 signaling, and that additional layers of regulation, such as isoform diversity or interaction with TP53 family members, may contribute to treatment response. In addition to canonical TP53 mutations, it is important to recognize that TP53 encodes multiple isoforms with distinct biological functions. Isoforms such as Δ40p53α, and Δ133p53β have been implicated in tumor progression, regulation of senescence, and clinical outcomes in breast cancer [30]. Furthermore, other members of the p53 family, including TP63, have been reported to partially compensate for TP53 loss and may influence transcriptional programs related to proliferation and metastasis. These findings highlight the complexity of TP53 pathway regulation beyond mutation status alone. However, due to the lack of isoform-level and TP53 family expression data in the datasets used in this study, we were unable to directly assess these factors. Future studies incorporating isoform-specific and functional analyses will be required to better understand how these layers of regulation contribute to treatment response. Additionally, we observed distinct co-mutation patterns according to TP53 status, including differences in PI3K, GATA3, and ESR1alterations. Notably, PI3K mutations were more frequent in the TP53-mutant group, suggesting that the co-occurring genomic alterations may influence treatment response in a combinatorial manner [31]. This interplay between multiple genetic alterations highlights the need for a more personalized approach to treatment, where therapies are tailored based on a comprehensive genetic profile that includes TP53 mutation and other co-occurring mutations. Until recently, P53 was considered a “undruggable target.” However, given the activation of CDK2 by TP53 loss, targeting CDK2 becomes crucial and has emerged as a potential pathway to overcome resistance associated with TP53 mutation. For example, co-targeting CDK2 and CDK4/6 has shown promise in preclinical models by enhancing senescence and suppressing cellular proliferation [32, 33] in preclinical models. Learning from these findings, several clinical trials have investigated this approach such as NCT05905341, and NCT05304962. NCT05304962 using RGT-419B showed a partial response of 50% in patients with metastatic HR positive HER2 negative breast cancer after progression on CDK4/6 inhibitors in the first-line setting. NCT03519178 investigating ebvaciclib showed a durable control rate of 48% (monotherapy) or 67% when combined with letrozole and fulvestrant [34]. This integrated approach may potentially enhance treatment outcomes by targeting specific molecular pathways relevant to individual tumor biology. The particularly interesting finding was the higher frequency of ESR1 mutation in TP53 wild-type tumors. ESR1 mutation is considered a marker of resistance to estrogen therapy [35], yet their presence at a higher frequency in TP53 wild-type tumors, which are less resistant to CDK4/6i, may seem paradoxical at first glance. However, this apparent contradiction highlights the complexity of cancer treatment response. Recent research has highlighted a notable relationship between germline variants in the ESR1 gene and somatic TP53 mutation in breast cancer. A study by Li et al. found a significant inverse relationship between ESR1 mutation and TP53 mutation in estrogen receptor-positive metastatic breast cancers, with TP53-mutated tumors exhibiting diminished ER activity [35]. Additionally, genome-wide association studies identified strong association between ESR1 locus variants, including rs9383938 and rs9479090, and the presence of TP53 mutation, particularly gain-of-function mutation [36, 37]. The intricate interplay of signaling pathways involved in treatment response underscores the difficulty in evaluating treatment outcomes based solely on TP53 or ESR1 mutation. While TP53 mutation plays a significant role in conferring treatment duration to CDK4/6i and other therapies, other genetic alterations, such as ESR1 mutation, contribute to through different mechanisms. The fact that TP53 mutation enhances estrogen response signaling, while ESR1 mutation is linked to reduced sensitivity to hormone therapy, demonstrates the complex interaction of these factors in treatment response. This complexity suggests that relying solely on TP53 mutation to predict treatment responses may be insufficient. As shown in this study, the integration of clinical and genomic data from real-world cohorts provides a more comprehensive approach to understanding the multifaceted nature of treatment response. By analyzing genetic factors, as well as clinical variables such as age and hormonal status, we gain deeper insight into how these factors influence treatment outcomes. This emphasizes the importance of conducting clinical research that explores the biological aspects of cancer from multiple angles, and the need for an integrated and multidimensional analysis to develop more personalized and effective treatment strategies. The observed discrepancy in the frequency of TP53 mutation, which was significantly higher in the AYA group compared to previous reports, is a notable finding. In the METABRIC cohort, which includes clinicopathological and genomic data from primary breast cancers treated surgically, TP53 mutation rates in the AYA group were reported to be approximately 25% 38. However, in our study of C-CAT cohort, the TP53 mutation rate in the AYA group was approximately 70%. This difference may reflect differences in cohort composition, as C-CAT is focused on metastatic breast cancer (mBC) and is based on clinical genomic testing in real-world practice. These findings support further investigation of TP53 mutation frequency in metastatic breast cancer cohort and of potential differences between primary and metastatic disease. In the age-stratified analyses, no significant association between TP53 mutation and CDK4/6i treatment duration was observed in the AYA group, whereas such an association was detected in the other age groups. At the same time, the AYA group showed shorter treatment duration overall than the other age groups. However, formal interaction testing did not detect a significant interaction between TP53 mutation and age group. Therefore, these subgroup findings should be interpreted cautiously as descriptive observations. Although our previous research has revealed that ER-positive/HER2-negative breast cancer in the AYA group exhibits enhanced cell proliferation, DNA repair, and other pro-cancerous signaling compared to other age groups, suggesting that AYA breast cancer may have unique biological characteristics [38], the present study was not designed to determine the mechanisms underlying these age-related patterns. Nonetheless, these real-world data provide additional context for understanding treatment duration patterns in the AYA population. Given the difficulty in conducting precise evaluations using preclinical models, the use of real-world clinical data in this study provides additional insights into breast cancer, particularly in the context of the AYA population. This also highlights the need for further clinical studies to evaluate potential age-related differences in treatment outcomes. A critical limitation of this study is the inability to distinguish between germline and somatic TP53 alterations within the C-CAT dataset. In particular, the AYA subgroup may include patients with hereditary cancer predisposition syndromes such as Li-Fraumeni syndrome, which is characterized by germline TP53 mutations and early-onset breast cancer. Therefore, the higher prevalence of TP53 mutations observed in the AYA group, as well as the lack of a significant association between TP53 mutation and CDK4/6i treatment duration in this subgroup, may in part reflect underlying biological heterogeneity, including differences in mutation origin (germline versus somatic), tumor biology, and clinical behavior. Future studies incorporating germline genetic information will be essential to clarify these relationships. One of the most intriguing findings of this study is the relationship between TP53 mutation and not only cell proliferation but also immune-related signaling pathways. While this relationship has been explored in various contexts [39, 40], the connection between TP53 mutation and immune pathway activation in hormone receptor-positive breast cancer, which is generally considered to have low immunogenicity, has not been thoroughly investigated. Our results suggest that, even in this immunologically less active subtype, TP53 mutation may contribute to the activation of immune-related pathways. However, despite the enrichment of immune-related signaling pathways in TP53-mutated tumors, we did not observe a significant association with ICI treatment duration. This finding underscores the complexity of immune evasion mechanisms in ER-positive/HER2-negative breast cancer, indicating that activation of immune-related pathways alone is insufficient to predict the effectiveness of ICIs. This is similar to another study using the data from Memorial Sloan Kettering Cancer Center (MSKCC) - where the benefit with ICI is confined to patients with TP53 mutations in bladder cancer but worse outcomes in colorectal cancer [41], suggesting that predictive role of TP53 mutation differs by tumor subtype. Other factors, such as the tumor microenvironment, or mutations in immune-related pathways, may play a significant role in influencing the therapeutic efficacy of ICI. Clinically, while ICI is primarily indicated for TNBC, genomic profiling may identify other populations, including those with microsatellite instability, or high tumor mutation burden, who could benefit from ICI therapy, even in patients with metastatic hormone receptor-positive breast cancer. Specifically, among tumors with TP53 mutations, there may be a differential immune response that warrants further investigation. Future studies should focus on the association between TP53 mutation and ICI treatment outcomes in TNBC, a subtype known for its higher immunogenicity and aggressive behavior using RWD, especially as ICI are FDA approved in patients with TNBC. Understanding the role of TP53 mutations in shaping tumor-immune interactions and treatment response will be important for the development of more effective therapeutic strategies for TNBC. In addition to hormonal influences, age-related changes in immune function may also contribute to the observed differences in treatment outcomes across age groups. The immune system undergoes significant remodeling with age, which can affect tumor-immune interactions and response to therapy. Consistent with this, previous studies have demonstrated that breast cancer in AYA exhibits distinct biological features, including altered immune and proliferative signaling, compared to tumors in older patients [38]. Notably, TP53-mutated tumors have been reported to generate neoantigens that can elicit immune responses [42, 43], although the magnitude and quality of such responses depend on factors including the specific TP53 mutation and the host HLA genotype. In our study, TP53-mutated tumors demonstrated enrichment of immune-related gene sets, supporting a link between TP53 status and tumor immunogenicity. However, due to the lack of HLA typing and mutation-specific immunogenicity data in our dataset, we were unable to directly assess these effects. Future studies integrating genomic, immunological, and host HLA status will be necessary to clarify the role of immune heterogeneity in age-dependent treatment responses. This study has several important limitations that should be acknowledged. First, although the METABRIC dataset is a large and comprehensive resource, it is retrospective and may not fully capture the dynamic nature of TP53 mutation or the tumor microenvironment in response to treatment over time. The clinical data available from METABRIC are based on a cohort of primarily patients from the Western world, and there may be differences in treatment regimens and outcomes when compared to the C-CAT cohort, which includes Japanese mBC patients. It should also be noted that while the C-CAT cohort primarily consists of patients with mBC, some of the patients may have received CDK4/6i for treatment of early breast cancer in the adjuvant setting, although genomic testing is performed for mBC. The inability to distinguish whether the CDK4/6i treatment was initiated for primary BC or mBC is a limitation of the C-CAT dataset. Second, the use of RWD from the C-CAT cohort introduces inherent biases associated with observational studies. While RWD offers valuable insights into the treatment outcomes of a diverse patient population, the lack of randomization and potential confounding factors could influence the observed associations. Specifically, the influence of unmeasured factors such as comorbidities, lifestyle factors, or prior treatments could not be fully controlled in this analysis. The definition of treatment line was based on the C-CAT database as recorded, and the criteria used to define line of therapy were not uniformly specified. This may have introduced heterogeneity in the classification of treatment lines. Additionally, ER and HER2 status were obtained from the dataset as recorded. However, the timing of receptor assessment (primary tumor versus metastatic lesion) was not uniformly specified. Although most cases are likely based on primary tumor data, the possibility that some cases reflect metastatic biopsy cannot be excluded. Therefore, the analysis may have included patients whose tumor phenotype differed at the time of treatment, which could influence the interpretation of treatment outcomes for ICI, and explains the use of ICI for ER-positive/HER2-negative breast cancer, although it is mainly approved for TNBC. Further studies using more uniformly defined receptor status from primary and metastatic tumors are needed to better assess these associations. Third, while the study identified significant associations between TP53 mutation and treatment duration for CDK4/6i, the mechanisms underlying these associations were not fully explored in this work. Further molecular and functional studies are needed to better understand how TP53 mutation contributes to treatment response to CDK4/6i and their interplay with ICIs, particularly at the genomic and proteomic levels. Additionally, the dataset does not allow differentiation between germline and somatic TP53 alterations. As a result, a subset of patients, particularly in the AYA group, may include individuals with hereditary TP53 mutations such as Li-Fraumeni syndrome. This limitation may contribute to the observed heterogeneity in TP53 mutation prevalence and treatment outcomes across age groups. Furthermore, this study focused on treatment duration as a surrogate marker of treatment efficacy and response, as long-term survival data were not available in the datasets analyzed. While treatment duration is a relevant measure of response, it may not fully capture the long-term effectiveness of treatment or the complex interplay between treatment and disease progression. Finally, TP53 mutation was classified into broad categories (truncating versus SNV and pathogenic versus non-pathogenic), which may not fully capture their functional heterogeneity. In addition, rare variants may be underrepresented. Future studies with more detailed functional annotation are warranted. Despite these limitations, this study provides important insights into the relationship between TP53 mutation and treatment response, particularly CDK4/6i therapy, and highlights the need for personalized treatment strategies that consider genetic mutations and age-related factors in treatment decisions for ER-positive/HER2-negative breast cancer. Conclusions This study shows that TP53 mutation is associated with shorter CDK4/6i treatment duration in ER-positive/HER2-negative breast cancer, while no significant association was observed with ICI treatment duration in this real-world cohort. Age-related differences in treatment duration and TP53 mutation prevalence were observed, particularly in the AYA group; however, no significant interaction between TP53 mutation and age group was detected. These findings highlight the importance of considering both genomic and age-related factors in the interpretation of treatment outcomes. In addition, the interplay between TP53 mutation and co-occurring genetic alterations, including ESR1 and PI3K, underscores the complexity of treatment response in ER-positive/HER2-negative breast cancer and supports the need for more personalized therapeutic strategies. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Not applicable. Author contributions Conceptualization, I.E., K.T., M.O., M.S.; Methodology, M.O., M.S.; Formal Analysis, M.O.; Writing-Original Draft Preparation, M.O.; Writing-Review & Editing, K.K., K.N., S.G., M.S., K.T., T.I., I.E., Y.A. Funding This research was supported Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS), grant number 23K14600 (M.O.), and 21K15535 (A.Y.), and the National Institutes of Health; R37CA248018, R01CA250412, R01CA251545, R01EB029596, and the US Department of Defense BCRP; W81XWH-19-1-0674 and W81XWH-19-1-0111 (K.T.), as well as by Roswell Park Comprehensive Cancer Center and National Cancer Institute cancer center support grant P30CA016056. Data availability Publicly available datasets analyzed in this study include METABRIC, accessible via the cBioPortal for Cancer Genomics. The C-CAT dataset is available upon application to the Center for Cancer Genomics and Advanced Therapeutics (C-CAT) and approval by relevant ethical committees in Japan (F250900014). Additional data generated in this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate The METABRIC dataset is publicly available and therefore exempt from ethical approval. The use of C-CAT data in this study was reviewed and approved by the Ethical Advisory Committee of YCU (IRB approval number: F250900014). All data analyzed were fully de-identified in accordance with applicable ethical guidelines. Consent for publication Not applicable, as this study used de-identified secondary data. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement Publicly available datasets analyzed in this study include METABRIC, accessible via the cBioPortal for Cancer Genomics. The C-CAT dataset is available upon application to the Center for Cancer Genomics and Advanced Therapeutics (C-CAT) and approval by relevant ethical committees in Japan (F250900014). Additional data generated in this study are available from the corresponding author upon reasonable request.