Age-Dependent Survival Benefit and Surrogacy of Progression-Free Survival in CDK4/6 Inhibitor Trials for HR+/HER2− Metastatic Breast Cancer: A Trial-Level Meta-analysis

Age-Dependent Survival Benefit and Surrogacy of Progression-Free Survival in CDK4/6 Inhibitor Trials for HR+/HER2− Metastatic Breast Cancer: A Trial-Level Meta-analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Age-Dependent Survival Benefit and Surrogacy of Progression-Free Survival in CDK4/6 Inhibitor Trials for HR+/HER2− Metastatic Breast Cancer: A Trial-Level Meta-analysis Ke Wang, Shuyu Li, Hong Liang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9551064/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Background CDK4/6 inhibitors consistently improve progression-free survival (PFS) in HR+/HER2 − metastatic breast cancer; however, whether these benefits translate into overall survival (OS) across age groups and whether PFS remains a valid surrogate endpoint in older patients remain unclear. We aimed to evaluate age-dependent treatment effects and the surrogate validity of PFS at the trial level. Methods Phase III randomized controlled trials comparing CDK4/6 inhibitors plus endocrine therapy versus endocrine therapy alone were included. Hazard ratios (HRs) for PFS and OS were extracted by age (< 65 vs ≥ 65 years). Pooled analyses were performed using random-effects models, with age-related effect modification assessed via trial-level interaction analyses. Trial-level surrogacy of PFS for OS was evaluated using weighted regression across age subgroups. A Bayesian hierarchical random-effects model was further applied to validate interaction effects. Results Six phase III trials were included. CDK4/6 inhibitors significantly improved PFS across age groups (HR ≈ 0.55) with minimal heterogeneity (I²=0%), and no significant age-related interaction was observed for PFS. In contrast, OS benefit showed significant age-dependent attenuation (interaction HR = 1.18, 95% CI 1.06–1.32). At the trial level, PFS strongly correlated with OS in the overall population (R²=0.81), but this association was markedly reduced in patients aged ≥ 65 years (R²=0.34). Bayesian analysis confirmed a high probability (94%) that younger patients derived greater OS benefit. Conclusions Although PFS benefit is preserved, its translation into OS is attenuated in older patients, indicating that PFS may be an unreliable surrogate and could overestimate true survival benefit in this population. Breast cancer CDK4/6 inhibitors Age stratification Interaction Surrogate endpoint Bayesian meta-analysis Figures Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction HR+/HER2- metastatic breast cancer is the most common molecular subtype of breast cancer, and the treatment strategy has long relied on endocrine therapy [ 1 ]. However, acquired drug resistance is almost inevitable, limiting further improvements in long-term survival benefits [ 2 ]. In recent years, CDK4/6 inhibitors have significantly delay disease progression by targeting cell cycle regulatory pathways and have consistently improved progression-free survival (PFS) in multiple phase III randomized controlled trials, thereby reshaping the first-line and subsequent treatment landscape for patients with HR+/HER2- metastatic breast cancer [ 3 – 6 ]. Although some studies have reported an overall survival (OS) benefit, considerable uncertainty remains regarding the consistency of this survival benefit across different patient populations [ 7 – 10 ]. Among all potential influencing factors, age is among the most clinically relevant and long-underestimated variables. With the aging of the global population, the proportion of patients aged 65 years or older in real-world settings continues to rise; however, this group remains significantly underrepresented in key randomized controlled trials [ 11 , 12 ]. Moreover, elderly patients typically have a higher comorbidity burden, greater functional decline, and different treatment exposure patterns, which impose structural limitations on the extrapolation of trial results to this population [ 13 ]. Previous studies have generally suggested that different age subgroups achieve similar PFS benefits [ 14 ]; however, this conclusion is largely based on descriptive subgroup analyses that lack formal statistical testing for whether age modifies the treatment effect. Notably, apparent consistency in effect sizes between subgroups does not equate to an absence of interaction, and traditional comparison methods relying on confidence interval overlap may mask real effect heterogeneity. Therefore, whether age is an independent effect modifier of CDK4/6 inhibitors remains a critical unanswered question. Moreover, the endpoint evaluation system faces methodological challenges. In oncology clinical trials, PFS is often used as a surrogate endpoint for OS to accelerate drug development and regulatory approval, because of its shorter observation period and greater statistical efficiency [ 15 , 16 ]. However, the association between PFS and OS is highly context-dependent and susceptible to multiple factors, including postprogression treatment, crossover medication, and non-tumor deaths [ 17 , 18 ]. This uncertainty may be further amplified in elderly patients: as the risk of competing events increases, OS gradually shifts from a tumor-specific outcome to a composite endpoint of multiple events, thereby structurally weakening the predictive value of PFS for OS [ 19 ]. In other words, a systematic “endpoint decoupling” may exist between PFS and OS, where the short-term disease control advantage fails to translate effectively into long-term survival benefit. Nevertheless, quantitative evidence supporting this hypothesis at the trial level is lacking. On the basis of the clinical and methodological flaws outlined above, a systematic trial-level meta-analysis with the goal of achieving integration across three levels. First, we used a formal interaction effect model to test whether age serves as an independent effect modifier of the efficacy of CDK4/6 inhibitors. Second, we quantitatively evaluated the surrogate value of PFS for OS at the trial level and compared its stability across different age subgroups. Third, we introduced a Bayesian hierarchical random-effects model to verify the robustness of the interaction effect from a probabilistic perspective. By integrating efficacy heterogeneity analysis with an endpoint evaluation framework, this study aims to clarify the role of age in the process of translating disease control to survival benefit and to reassess the applicability of PFS in the elderly population, thereby providing a more explanatory and generalizable evidence base for clinical decision-making and future trial design. 2. Methods 2.1 Literature Search and Research Selection This study strictly adhered to the reporting guidelines for systematic reviews and meta-analyses (PRISMA 2020). The PubMed, EMBASE, and Cochrane Library databases were systematically searched from their inception to February 2026, without any language restrictions. Additionally, the abstract databases of major international oncology conferences (ASCO and ESMO) were searched to reduce the risk of publication bias. Prospective registration of the study protocol was completed in the PROSPERO database (registration number: CRD420261345790). The search strategy combined medical subject headings (MeSH) and free-text terms, with the main keywords encompassing “CDK4/6 inhibitors,” “breast cancer,” “randomized controlled trial,” and was adapted according to the specifications of each database. In addition, we conducted a manual search on the reference lists of included studies and relevant review articles to enhance the completeness of the search. 2.2 Inclusion and Exclusion Criteria Characteristics of included studies: (1) the study was a phase III randomized controlled trial; (2) it compared CDK4/6 inhibitors combined with endocrine therapy versus endocrine therapy alone; (3) the hazard ratio (HR) and 95% CI of PFS and/or OS were reported; (4) it provided efficacy data stratified by age (< 65 years vs. ≥ 65 years). Studies that did not directly report stratified HRs in the original manuscript but allowed derivation from public data or supplementary materials were also included. Nonrandomized studies, duplicate publications, and studies from which age-stratified effect sizes could not be extracted were excluded. For multiple updated reports of the same trial, only the version with the longest follow-up duration and the most mature data was included. 2.3 Data Extraction and Quality Evaluation Two researchers independently extracted and cross-checked the data, including trial name, study design, sample size, treatment regimen, median follow-up duration, and HRs for PFS and OS (overall and age subgroups). For key outcome data, the two researchers independently entered and checked the data to reduce human error. Disagreements were resolved through discussion or adjudicated by a third party. The quality of the included studies was systematically assessed using the Cochrane RoB 2.0 tool, which focuses on the randomization process, allocation concealment, biases from intended interventions, measurement of the outcome, and selection of the reported result. The overall risk of bias for each study was graded as low risk, some concerns, or high risk to inform subsequent sensitivity analyses. 2.4 Traditional Meta-analysis (PFS and OS) A random-effects model was used to pool the results for PFS and OS to account for clinical and methodological heterogeneity across studies. The between-study variance (τ²) was estimated through the restricted maximum likelihood (REML) method, and the Hartung–Knapp–Sidik–Jonkman method was applied to adjust the confidence intervals, thereby improving robustness in small-sample settings. Effect sizes were expressed as HRs, analyzed on a logarithmic scale (log HR), and weighted by inverse variance. To enhance the robustness and interpretability of the results, the 95% prediction interval was also reported to reflect the range of effect sizes that could be expected in future studies. Heterogeneity was assessed using the I² statistic and Cochran's Q test. When I² exceeded 50%, potential sources were further explored. However, given that this study was primarily based on trial-level analyses, multivariate meta-regression was not performed to avoid overfitting. 2.5 Age Interaction Analysis To avoid informal comparisons based on superficial differences in traditional subgroup analyses, a formal interaction effect analysis framework was constructed at the trial level to directly test whether age serves as a modifier of the treatment effect. The interaction effect was defined as the ratio of HRs between different age subgroups: $$\:H{R}_{interaction}=\frac{H{R}_{\ge\:65}}{H{R}_{<65}}$$ Calculates on the logarithmic scale: $$\:\text{l}\text{o}\text{g}\left(H{R}_{interaction}\right)=\text{l}\text{o}\text{g}\left(H{R}_{\ge\:65}\right)-\text{l}\text{o}\text{g}\left(H{R}_{<65}\right)$$ The standard error was derived from the variance of the log hazard ratios of the two subgroups. The interaction effects from each trial were pooled using a random-effects model (REML), and the Hartung-Knapp method was used to adjust the interval estimation. HR_interaction > 1 indicates that younger patients derive greater benefit, whereas < 1 indicates a greater benefit in older patients. This approach enables a formal statistical test of the modifying effect of age and fundamentally avoids the biases associated with insufficient sample size or random error in traditional subgroup analyses. 2.6 Assessment of PFS as a Surrogate Endpoint for OS On the basis of the trial-level surrogate endpoint analysis framework, the predictive ability of PFS for OS was evaluated. The HRs for PFS and OS from each trial were log(HR) transformed and used to construct a weighted linear regression model: $$\:\text{l}\text{o}\text{g}\left(H{R}_{OS}\right)=\alpha\:+\beta\:\cdot\:\text{l}\text{o}\text{g}\left(H{R}_{PFS}\right)$$ Regression analysis was performed using inverse variance weighting and sample size weighting as sensitivity analyses. The coefficient of determination (R²) was used to quantify the explanatory power of PFS with respect to OS variability. Further modeling was conducted in the overall population, as well as in the < 65 years and ≥ 65 years age subgroups, to compare the stability of surrogate endpoint validity across different age groups. 2.7 Bayesian Hierarchical Random-effects Model To verify the interaction effect results from a probabilistic perspective and to characterize uncertainty, a Bayesian hierarchical random-effects model was constructed to model the log(HR_interaction) [ 20 ]: \(\:{y}_{i}\sim\:N({\theta\:}_{i},{\sigma\:}_{i}^{2})\) The requirements at the research level are as follows: $$\:{\theta\:}_{i}\sim\:N(\mu\:,{\tau\:}^{2})$$ \(\:{y}_{i}\) represents the log(HR_interaction) for each study, \(\:{\sigma\:}_{i}^{2}\) is its variance, \(\:\mu\:\) is the overall effect, and \(\:{\tau\:}^{2}\:\) represents the between-study heterogeneity. Weakly informative priors were used as follows: $$\:\mu\:\sim\:N\left(0,{10}^{2}\right)$$ $$\:\tau\:\sim\:\text{Half-Cauchy}\left(\text{0,1}\right)$$ Parameters were estimated by the Markov chain Monte Carlo (MCMC) method (four chains, each with 4,000 iterations, including 2,000 warm-up iterations). Convergence was assessed through trace plots and the Gelman–Rubin statistic ( \(\:\widehat{R}\) ). The reported as medians with 95% confidence intervals (CIs). The posterior probability that the HR_interaction exceeds 1 was calculated to directly quantify the strength of evidence for effect modification by age, thereby addressing the limitations associated with the interpretation of traditional p-values. 2.8 Sensitivity Analysis The robustness of the results was evaluated using the leave-one-out method, which involves removing individual studies one by one and repeating the analysis to observe changes in the pooled effects. In addition, a multidimensional sensitivity analysis was prespecified, including restrictions to trials reporting mature OS data, application of different weighting strategies (inverse variance and sample size), and exclusion of studies with a high risk of bias. Moreover, the results from the fixed-effects model and the random-effects model were compared to evaluate the impact of model specification on the conclusions. 2.9 Statistical Analysis All frequentist analyses were performed using R software (version 4.3.0), primarily employing the “meta” and “metafor” packages. Bayesian analyses were implemented using “rstan” or “brms”. All statistical tests were two-sided, and P < 0.05 was considered indicate statistical significance. The results from the Bayesian analyses did not rely on significance thresholds but were interpreted based on effect distributions and posterior probabilities. 3. Results 3.1 Study Selection and Characteristics Six phase III randomized controlled trials (PALOMA-2, PALOMA-3, MONALEESA-2, MONALEESA-3, MONARCH-2, and MONARCH-3) were included in this study to compare the efficacy of CDK4/6 inhibitors combined with endocrine therapy versus endocrine therapy alone in patients with HR + /HER2 − metastatic breast cancer [ 5 , 6 , 14 , 21 – 23 ]. All trials reported PFS in the overall population and provided effect sizes for PFS and OS stratified by age (< 65 years and ≥ 65 years), thereby meeting the preconditions for interaction effect analysis and surrogate endpoint analysis. Overall, there were certain differences among the studies in terms of treatment line, endocrine regimen, and follow-up duration; however, the methodological quality was generally high, and the risk of bias was mainly low or with some concerns (Supplementary Figure S1 ). The study screening process is shown in Fig. 1 , and baseline characteristics are summarized in Table 1 . Table 1 Characteristics of Included Phase III Trials Trial Drug Line of Therapy Endocrine Partner Menopausal Status N Median Age (range) ≥ 65 (%) Follow-up (months) PALOMA-2 Palbociclib First-line Letrozole Postmenopausal 666 62 (28–89) 41 79.7 PALOMA-3 Palbociclib Second-line Fulvestrant Any 521 62 (30–88) 34 73.3 MONALEESA-2 Ribociclib First-line Letrozole Postmenopausal 668 62 (32–91) 40 80.0 MONALEESA-3 Ribociclib First-/Second-line Fulvestrant Postmenopausal 726 63 (30–88) 43 56.3 MONARCH-2 Abemaciclib Second-line Fulvestrant Any 669 60 (32–91) 33 47.7 MONARCH-3 Abemaciclib First-line NSAI Postmenopausal 493 62 (36–91) 38 96.0 Footnote: NSAI, nonsteroidal aromatase inhibitor. All included studies were phase III randomized controlled trials comparing CDK4/6 inhibitors plus endocrine therapy versus endocrine therapy alone. 3.2 Consistent PFS Benefit across Different Age Groups CDK4/6 inhibitors produce highly consistent and reproducible disease control effects across different age groups. In the overall population, treatment significantly prolonged PFS (HR = 0.54, 95% CI 0.50–0.59), with no observed between-study heterogeneity (I² = 0%) (Table 2 ). In the age-stratified analysis, the HR was 0.54 (95% CI 0.49–0.59) for patients aged < 65 years and 0.52 (95% CI 0.46–0.60) for those aged ≥ 65 years. The effect estimates for the two subgroups largely overlapped, and both demonstrated very low heterogeneity (I² = 0%). The prediction intervals were also highly consistent between these two subgroups (< 65 years: 0.49–0.60; ≥65 years: 0.46–0.61), suggesting that the inhibitory effect of this treatment has good generalizability and stability across different age groups. Overall, no effect modification by age was observed for PFS benefit (Fig. 2 ). Table 2 Age-Stratified Efficacy and Interaction Effects of CDK4/6 Inhibitors Outcome Age Group HR (95% CI) I² (%) Prediction Interval PFS Overall 0.54 (0.50–0.59) 0.0 0.48–0.62 < 65 years 0.54 (0.49–0.59) 0.0 0.47–0.62 ≥ 65 years 0.52 (0.46–0.60) 0.0 0.44–0.62 OS Overall 0.75 (0.67–0.83) 0.0 0.65–0.86 < 65 years 0.74 (0.66–0.83) 0.0 0.64–0.86 ≥ 65 years 0.88 (0.75–1.03) 0.0 0.73–1.05 Interaction (OS) ≥ 65 vs < 65 1.18 (1.06–1.32) 0.0 0.94–1.50 3.3 Age-dependent Divergence in Overall Survival Benefit Although the PFS benefit remained consistent across different age groups, OS improvement showed potential age-related heterogeneity. Based on an interaction effect analysis at the trial level, the pooled interaction HR was 1.18 (95% CI 1.06–1.32, P = 0.011), indicating that younger patients (< 65 years) may derive greater survival benefit. This finding provides direct statistical evidence for age as an effect modifier of treatment efficacy, rather than a mere observation of subgroup differences. Although the prediction interval ranged from 0.94 to 1.50, the interaction effects were highly consistent across studies (I² = 0%, τ² ≈ 0), suggesting that a degree of uncertainty will remain in future studies (Fig. 3 ). Additionally, it should be noted that the fixed-effects model did not yield statistical significance (HR = 1.18, 95% CI 0.99–1.42), but the results became significant after the random-effects model with the Hartung-Knapp correction, underscoring the importance of using robust estimation methods when the number of trials is limited. 3.4 Structural Weakening of the PFS–OS Surrogate Relationships in Elderly Patients In the overall population, a strong correlation was observed between PFS and OS (R² = 0.81, 95% CI 0.65–0.91), supporting the overall validity of PFS as a surrogate endpoint for OS. However, the correlation differed markedly after age stratification: the R² was 0.79 (95% CI 0.62–0.90) in the subgroup aged < 65 years, whereas it decreased substantially to 0.34 (95% CI 0.11–0.58) in patients aged ≥ 65 years (Table 3 ). These findings indicate that in the elderly population, PFS explains only approximately one-third of the variation in OS, indicating a structural weakening of its validity as a surrogate endpoint (Fig. 4 ). Further analysis revealed a significantly reduced regression slope in the older subgroup, reflecting a great quantitative weakening of the predictive strength of PFS for OS. Collectively, these results support the phenomenon of “endpoint decoupling,” whereby the advantage of short-term disease control cannot be effectively translated into long-term survival benefit. Table 3 Age-Dependent Breakdown of PFS as a Surrogate for Overall Survival Population R² (95% CI) Interpretation Overall 0.81 (0.65–0.91) Strong surrogacy < 65 years 0.79 (0.62–0.90) Strong surrogacy ≥ 65 years 0.34 (0.11–0.58) Poor surrogacy Footnote: Surrogacy was assessed using weighted linear regression of log(HR_PFS) on log(HR_OS) at the trial level. R² represents the proportion of variance in OS treatment effect explained by PFS. Higher R² indicates stronger surrogate validity. 3.5 Bayesian Analysis-based Probabilistic Validation of Age Interaction Effects A Bayesian hierarchical random-effects model further strengthened the evidence for the age interaction effect from a probabilistic perspective. The analysis revealed that the posterior median of the interaction HR is 1.18 (95% CI 1.03–1.34), which was highly consistent with the frequentist analysis. The posterior probability of interaction HR > 1 was 94%, indicating that the conclusion that younger patients derive greater benefit is strongly supported in probabilistic terms. The posterior distribution was predominantly concentrated in the region of HR > 1, with only approximately 5.8% of the probability mass falling in the range of HR < 1, thereby providing a more robust inferential basis under the condition of a limited number of trials (Fig. 5 ). 3.6 Sensitivity Analysis Multidimensional sensitivity analysis further verified the robustness of the main results. Leave-one-out analyses demonstrated that sequential exclusion of individual trials yielded pooled interaction HRs ranging from 1.14 to 1.20, with all estimates consistently exceeding unity, closely aligning with the overall effect, and indicating that the observed age-dependent interaction was not driven by any single study (Fig. 6 ,). In addition, after limiting trials to mature OS data, applying different weighting strategies, and excluding studies with a high risk of bias, the main conclusions did not change substantially (Supplementary Figure S2 ), further supporting the reliability and reproducibility of the results in this study. 4. Discussion This study systematically revealed age-related differences in the efficacy of CDK4/6 inhibitors in HR + /HER2 − metastatic breast cancer at the trial level, providing two key findings of direct clinical and methodological significance. First, although progression-free survival (PFS) benefit was highly consistent across age groups, overall survival (OS) improvement tended toward age-related heterogeneity, suggesting that the relationship between disease control and survival benefit may differ by age. Second, the surrogate validity of PFS for OS appeared reduced in older patients, suggesting substantially diminished and potentially unreliable surrogacy in this population. This finding may challenge the prevailing assumption of consistent treatment effects across age groups [ 11 , 13 ]. Previous analyses have largely concluded that “age does not affect efficacy” based on the apparent similarity of effect sizes between subgroups. However, this inference essentially relies on informal comparisons and lacks direct statistical tests for interactions. In this study, the “age × treatment effect” was quantitatively evaluated by an interaction effect model at trial-level, and the posterior probability of the effect direction was directly calculated within a Bayesian framework. This combined approach may enhance the interpretability of effect modification under conditions of limited trial numbers. From a mechanistic perspective, the observed divergence between PFS and OS may not be incidental but could reflect age-related changes in outcome structure and treatment pathways. First, the structural impact of competing risks likely represents an important contributing factor. As age increases, non-tumor-related deaths increase significantly, transforming OS from a single tumor outcome into a composite endpoint comprising multiple events. Under this structure, gains in tumor control may be progressively diluted at the level of OS, resulting in an apparent attenuation of survival benefit despite preserved PFS effects. Second, age-related differences in tumor biological characteristics may also contribute to this process. Older patients may present with different tumor proliferation patterns, which could alter the relative contribution of cell cycle inhibition to long-term survival. Consequently, the effects of CDK4/6 inhibitors may be primarily disease-stabilizing rather than survival-prolonging. Together, these factors establish a multidimensionally driven mechanistic framework for decoupling of endpoints. On the basis of the above mechanism, the results of this study have direct implications for clinical practice and trial design. First, in clinical practice, CDK4/6 inhibitors may still provide stable disease control in elderly patients, but their true survival benefit may be overestimated. This suggests that treatment decisions should not rely solely on PFS, but instead comprehensively consider life expectancy, competing risks, and treatment-related burden to avoid the potential overtreatment that arises from “replacing survival benefit with progression delay.” In terms of trial design and drug evaluation, these results may question the universal applicability of PFS as a primary endpoint. Although PFS still has high surrogate value in the overall population, its explanatory power decreases significantly in the elderly subgroup, in whom no stable predictive relationship is observed. These findings suggest that the assumption of PFS as a universally valid surrogate endpoint may not fully true across all patient subgroups. Future studies should prioritize OS or other more clinically relevant composite endpoints, and incorporate interaction effect testing and competing risk modeling at the design stage to improve the generalizability and clinical interpretability of the evidence. Several limitations should be acknowledged. First, this study is based on trial-level data rather than individual patient data, and therefore could not make fine adjustments for key confounders such as comorbidity burden, functional status, and treatment exposure intensity. Second, the limited number of included trials may affect the precision and stability of certain estimates. Accordingly, the findings should be interpreted as hypothesis-generating rather than definitive. In addition, the use of 65 years as the age cutoff was primarily dictated by trial reporting conventions and may not fully capture biological aging or functional heterogeneity. These limitations themselves highlight a structural gap in the current evidence base—namely, that the stability of a single endpoint framework cannot be assumed in highly heterogeneous populations. In conclusion, this study demonstrates age-dependent differences in the translation of PFS into OS benefit with CDK4/6 inhibitors, with preserved PFS but attenuated OS in older patients, accompanied by reduced surrogate validity. These findings challenge the universal applicability of PFS and support the need for population-specific endpoint frameworks. Declarations Acknowledgements We thank Springer Nature Author Services for their English language editing services, which improved the clarity and quality of this manuscript. Author contributions WK conceived the study and designed the methodology. SL and LH performed the statistical analyses. All authors contributed to data interpretation and approved the final manuscript. Data availability All data generated or analyzed during this study are included in this published article and its supplementary files. Ethical approval Not applicable Funding This study was supported by the Natural Science Foundation of Zhejiang Province (Grant No. LMRY26H17001). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Data availability statement The data generated in this study are available upon request from the corresponding author. References Median MD, Antone NZ, Volovăț S, Mazilu L, Negru M, Curcă RO, Niță A, Pătru RI, Ungureanu A, Lupu V, et al. Romanian Consensus Statement for Hormone Receptor-Positive and Human Epidermal Growth Factor Receptor 2-Negative Metastatic Breast Cancer (HR+/HER2- mBC) and Triple-Negative Metastatic Breast Cancer (mTNBC). Curr Oncol (Toronto Ont). 2026;33(2):120. Mavratzas A, Marmé F. Alpelisib in the treatment of metastatic HR+ breast cancer with PIK3CA mutations. Future Oncol (London England). 2021;17(1):13–36. Goetz MP, Toi M, Huober J, Sohn J, Trédan O, Park IH, Campone M, Chen S, Manso LM, Paluch-Shimon S, et al. Abemaciclib plus a nonsteroidal aromatase inhibitor as initial therapy for HR+, HER2- advanced breast cancer: final overall survival results of MONARCH 3. Annals oncology: official J Eur Soc Med Oncol. 2024;35(8):718–27. Hortobagyi GN, Stemmer SM, Burris HA, Yap Y, Sonke GS, Hart L, Campone M, Petrakova K, Winer EP, Janni W, et al. Overall Survival with Ribociclib plus Letrozole in Advanced Breast Cancer. N Engl J Med. 2022;386(10):942–50. Finn RS, Martin M, Rugo HS, Jones S, Im S, Gelmon K, Harbeck N, Lipatov ON, Walshe JM, Moulder S, et al. Palbociclib and Letrozole in Advanced Breast Cancer. N Engl J Med. 2016;375(20):1925–36. Hortobagyi GN, Stemmer SM, Burris HA, Yap Y, Sonke GS, Paluch-Shimon S, Campone M, Blackwell KL, André F, Winer EP, et al. Ribociclib as First-Line Therapy for HR-Positive, Advanced Breast Cancer. N Engl J Med. 2016;375(18):1738–48. Slamon DJ, Neven P, Chia S, Jerusalem G, De Laurentiis M, Im S, Petrakova K, Valeria Bianchi G, Martín M, Nusch A, et al. Ribociclib plus fulvestrant for postmenopausal women with hormone receptor-positive, human epidermal growth factor receptor 2-negative advanced breast cancer in the phase III randomized MONALEESA-3 trial: updated overall survival. Annals oncology: official J Eur Soc Med Oncol. 2021;32(8):1015–24. Sledge GWJ, Toi M, Neven P, Sohn J, Inoue K, Pivot X, Burdaeva O, Okera M, Masuda N, Kaufman PA, et al. The Effect of Abemaciclib Plus Fulvestrant on Overall Survival in Hormone Receptor-Positive, ERBB2-Negative Breast Cancer That Progressed on Endocrine Therapy-MONARCH 2: A Randomized Clinical Trial. JAMA ONCOL. 2020;6(1):116–24. Slamon DJ, Neven P, Chia S, Fasching PA, De Laurentiis M, Im S, Petrakova K, Bianchi GV, Esteva FJ, Martín M, et al. Overall Survival with Ribociclib plus Fulvestrant in Advanced Breast Cancer. N Engl J Med. 2020;382(6):514–24. Im S, Lu Y, Bardia A, Harbeck N, Colleoni M, Franke F, Chow L, Sohn J, Lee K, Campos-Gomez S, et al. Overall Survival with Ribociclib plus Endocrine Therapy in Breast Cancer. N Engl J Med. 2019;381(4):307–16. Freedman RA, Tolaney SM. Efficacy and safety in older patient subsets in studies of endocrine monotherapy versus combination therapy in patients with HR+/HER2- advanced breast cancer: a review. BREAST CANCER RES TR. 2018;167(3):607–14. Pilehvari A, Kimmick G, You W, Bonilla G, Anderson R. Disparities in receipt of 1-(st) line CDK4/6 inhibitors with endocrine therapy for treatment of hormone receptor positive, HER2 negative metastatic breast cancer in the real-world setting. Breast cancer research: BCR. 2024;26(1):144. Deluche E, Antoine A, Bachelot T, Lardy-Cleaud A, Dieras V, Brain E, Debled M, Jacot W, Mouret-Reynier MA, Goncalves A et al. Contemporary outcomes of metastatic breast cancer among 22,000 women from the multicentre ESME cohort 2008–2016. European journal of cancer (Oxford, England : 1990) 2020, 129:60–70. Turner NC, Ro J, André F, Loi S, Verma S, Iwata H, Harbeck N, Loibl S, Huang Bartlett C, Zhang K, et al. Palbociclib in Hormone-Receptor-Positive Advanced Breast Cancer. N Engl J Med. 2015;373(3):209–19. Courtinard C, Gourgou S, Jacot W, Carton M, Guérin O, Vacher L, Bertaut A, Le Deley M, Pérol D, Marino P, et al. Association between progression-free survival and overall survival in women receiving first-line treatment for metastatic breast cancer: evidence from the ESME real-world database. BMC MED. 2023;21(1):87. Burzykowski T, Buyse M, Piccart-Gebhart MJ, Sledge G, Carmichael J, Lück H, Mackey JR, Nabholtz J, Paridaens R, Biganzoli L, et al. Evaluation of tumor response, disease control, progression-free survival, and time to progression as potential surrogate end points in metastatic breast cancer. J Clin oncology: official J Am Soc Clin Oncol. 2008;26(12):1987–92. Li J, Huo X, Zhao F, Ren D, Ahmad R, Yuan X, Du F, Zhao J. Association of Cyclin-Dependent Kinases 4 and 6 Inhibitors With Survival in Patients With Hormone Receptor-Positive Metastatic Breast Cancer: A Systematic Review and Meta-analysis. JAMA NETW OPEN. 2020;3(10):e2020312. Piezzo M, Chiodini P, Riemma M, Cocco S, Caputo R, Cianniello D, Di Gioia G, Di Lauro V, Rella FD, Fusco G, et al. Progression-Free Survival and Overall Survival of CDK 4/6 Inhibitors Plus Endocrine Therapy in Metastatic Breast Cancer: A Systematic Review and Meta-Analysis. INT J MOL SCI. 2020;21(17):6400. Canouï-Poitrine F, Lièvre A, Dayde F, Lopez-Trabada-Ataz D, Baumgaertner I, Dubreuil O, Brunetti F, Coriat R, Maley K, Pernot S, et al. Inclusion of Older Patients with Cancer in Clinical Trials: The SAGE Prospective Multicenter Cohort Survey. Oncologist. 2019;24(12):e1351–9. Clark J, Muhlemann N, Ionan A. Bayesian Clinical Trials. THER INNOV REGUL SCI. 2023;57(3):399–400. Slamon DJ, Neven P, Chia S, Fasching PA, De Laurentiis M, Im S, Petrakova K, Bianchi GV, Esteva FJ, Martín M, et al. Phase III Randomized Study of Ribociclib and Fulvestrant in Hormone Receptor-Positive, Human Epidermal Growth Factor Receptor 2-Negative Advanced Breast Cancer: MONALEESA-3. J Clin oncology: official J Am Soc Clin Oncol. 2018;36(24):2465–72. Sledge GWJ, Toi M, Neven P, Sohn J, Inoue K, Pivot X, Burdaeva O, Okera M, Masuda N, Kaufman PA, et al. MONARCH 2: Abemaciclib in Combination With Fulvestrant in Women With HR+/HER2- Advanced Breast Cancer Who Had Progressed While Receiving Endocrine Therapy. J Clin oncology: official J Am Soc Clin Oncol. 2017;35(25):2875–84. Goetz MP, Toi M, Campone M, Sohn J, Paluch-Shimon S, Huober J, Park IH, Trédan O, Chen S, Manso L, et al. MONARCH 3: Abemaciclib As Initial Therapy for Advanced Breast Cancer. J Clin oncology: official J Am Soc Clin Oncol. 2017;35(32):3638–46. Legend. Additional Declarations No competing interests reported. Supplementary Files SupplementaryFigureS1.tif SupplementaryFigureS2.tif Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 21 May, 2026 Reviewers agreed at journal 16 May, 2026 Reviewers agreed at journal 15 May, 2026 Reviewers agreed at journal 12 May, 2026 Reviewers invited by journal 07 May, 2026 Editor invited by journal 29 Apr, 2026 Editor assigned by journal 28 Apr, 2026 Submission checks completed at journal 28 Apr, 2026 First submitted to journal 28 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9551064","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":641633040,"identity":"1ed42d2a-1b0f-460d-bb99-9dc7656286f9","order_by":0,"name":"Ke Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwklEQVRIiWNgGAWjYBACA3YwlcDAz8x88AFxWpihWiTb2ZINSNNicJ7HTIAoLebMPIafC36lyRkfZjBjYKixiSaoxbKZx1h6Zl+OsdlhhrQHDMfSchsIOuww7wZp3p6KxG2HGY4bMDYcJkrL5t9ALfWbmxnbJIjVsk2a50dOggEzMxtxWiyb+b9Z8zakGc44zMZskECMX8zZ25Jv8/xJlufvP//xwYcaG8JawICxDcpIIEo5GPwhXukoGAWjYBSMQAAAYSw7gNoN73kAAAAASUVORK5CYII=","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":true,"prefix":"","firstName":"Ke","middleName":"","lastName":"Wang","suffix":""},{"id":641633042,"identity":"424f8bfc-2f5f-419a-9c7d-a8692c00d468","order_by":1,"name":"Shuyu Li","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Shuyu","middleName":"","lastName":"Li","suffix":""},{"id":641633043,"identity":"3c8c499c-00b5-464d-9c9a-38396bb952f6","order_by":2,"name":"Hong Liang","email":"","orcid":"","institution":"Zhejiang University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Hong","middleName":"","lastName":"Liang","suffix":""}],"badges":[],"createdAt":"2026-04-28 08:40:01","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9551064/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9551064/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109451006,"identity":"a30aaf15-0f66-451e-aa5d-d24ecab7411a","added_by":"auto","created_at":"2026-05-18 09:03:03","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3690110,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eForest plots of progression-free survival according to age subgroup\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Overall population; (B) patients aged \u0026lt;65 years; (C) patients aged ≥65 years. Random-effects models (REML with Hartung–Knapp adjustment) were used. Square sizes represent study weights; diamonds represent pooled estimates; red lines indicate prediction intervals.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/8aab2243aab9f3c5f90f62ca.png"},{"id":109759722,"identity":"f9243af9-e827-4e97-9f59-bceca3c805de","added_by":"auto","created_at":"2026-05-22 07:27:36","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1308571,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eInteraction analysis of age and treatment effect on overall survival\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eForest plot showing the interaction between age (\u0026lt;65 vs ≥65 years) and treatment effect on overall survival. The interaction hazard ratio was defined as HR ≥65 / HR \u0026lt;65. The prediction interval ranged from 0.94 to 1.50. Square sizes represent study weights; diamonds represent pooled estimates.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/732d490ae5ca7f064781c7ce.png"},{"id":109451008,"identity":"4d099bb5-4d65-4696-a07a-6b9f87ec672d","added_by":"auto","created_at":"2026-05-18 09:03:03","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2067515,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssociation between progression-free survival and overall survival\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Overall population; (B) patients aged \u0026lt;65 years; (C) patients aged ≥65 years. Bubble size is proportional to sample size. Weighted linear regression showed R² values of 0.81 in the overall population, 0.79 in patients aged \u0026lt;65 years, and 0.34 in those aged ≥65 years.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/913037c2798c263e0f60cf78.png"},{"id":109451010,"identity":"0d0c3bf5-6e30-4f0d-914d-70cdb0751c6d","added_by":"auto","created_at":"2026-05-18 09:03:03","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":69801,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBayesian posterior distribution of the interaction effect\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePosterior distribution of the interaction hazard ratio (≥65 vs \u0026lt;65 years). The posterior median was 1.18, with a 95% credible interval of 1.03–1.34. The posterior probability that the interaction HR exceeded 1 was 94%.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/5272e5b0450b4a224fdd7ecb.png"},{"id":109451012,"identity":"439f8832-f968-4b17-9c0d-763f256ecc58","added_by":"auto","created_at":"2026-05-18 09:03:03","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":584103,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLeave-one-out sensitivity analysis of age\u003c/strong\u003e–\u003cstrong\u003etreatment interaction for overall survival\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSequential exclusion of individual trials yielded pooled interaction hazard ratios ranging from 1.14 to 1.20, with all estimates consistently exceeding unity and closely matching the overall effect (HR = 1.18, 95% CI 1.06–1.32), indicating that the observed interaction is stable and not driven by any single study.\u003c/p\u003e","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/56bcd5f8857fa1f0cbf8d0c6.png"},{"id":109451005,"identity":"c554cf8f-bb56-458c-ad70-fc6b58543621","added_by":"auto","created_at":"2026-05-18 09:02:50","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":245599,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/ecc03072-d017-4818-b71d-89f29fe19dee.pdf"},{"id":109759316,"identity":"275624d5-7b69-4acf-a2d2-fdd1ef07eb5b","added_by":"auto","created_at":"2026-05-22 07:26:37","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":191876,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigureS1.tif","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/edef077816b48db3d9a988d0.tif"},{"id":109759244,"identity":"fd67eedc-1a29-413d-ad41-0c2995bae253","added_by":"auto","created_at":"2026-05-22 07:26:16","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":169104,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigureS2.tif","url":"https://assets-eu.researchsquare.com/files/rs-9551064/v1/fdd17600b95aa197f3be34cd.tif"}],"financialInterests":"No competing interests reported.","formattedTitle":"Age-Dependent Survival Benefit and Surrogacy of Progression-Free Survival in CDK4/6 Inhibitor Trials for HR+/HER2− Metastatic Breast Cancer: A Trial-Level Meta-analysis","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eHR+/HER2- metastatic breast cancer is the most common molecular subtype of breast cancer, and the treatment strategy has long relied on endocrine therapy [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. However, acquired drug resistance is almost inevitable, limiting further improvements in long-term survival benefits [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. In recent years, CDK4/6 inhibitors have significantly delay disease progression by targeting cell cycle regulatory pathways and have consistently improved progression-free survival (PFS) in multiple phase III randomized controlled trials, thereby reshaping the first-line and subsequent treatment landscape for patients with HR+/HER2- metastatic breast cancer [\u003cspan additionalcitationids=\"CR4 CR5\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Although some studies have reported an overall survival (OS) benefit, considerable uncertainty remains regarding the consistency of this survival benefit across different patient populations [\u003cspan additionalcitationids=\"CR8 CR9\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAmong all potential influencing factors, age is among the most clinically relevant and long-underestimated variables. With the aging of the global population, the proportion of patients aged 65 years or older in real-world settings continues to rise; however, this group remains significantly underrepresented in key randomized controlled trials [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Moreover, elderly patients typically have a higher comorbidity burden, greater functional decline, and different treatment exposure patterns, which impose structural limitations on the extrapolation of trial results to this population [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Previous studies have generally suggested that different age subgroups achieve similar PFS benefits [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]; however, this conclusion is largely based on descriptive subgroup analyses that lack formal statistical testing for whether age modifies the treatment effect. Notably, apparent consistency in effect sizes between subgroups does not equate to an absence of interaction, and traditional comparison methods relying on confidence interval overlap may mask real effect heterogeneity. Therefore, whether age is an independent effect modifier of CDK4/6 inhibitors remains a critical unanswered question.\u003c/p\u003e \u003cp\u003eMoreover, the endpoint evaluation system faces methodological challenges. In oncology clinical trials, PFS is often used as a surrogate endpoint for OS to accelerate drug development and regulatory approval, because of its shorter observation period and greater statistical efficiency [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. However, the association between PFS and OS is highly context-dependent and susceptible to multiple factors, including postprogression treatment, crossover medication, and non-tumor deaths [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. This uncertainty may be further amplified in elderly patients: as the risk of competing events increases, OS gradually shifts from a tumor-specific outcome to a composite endpoint of multiple events, thereby structurally weakening the predictive value of PFS for OS [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. In other words, a systematic \u0026ldquo;endpoint decoupling\u0026rdquo; may exist between PFS and OS, where the short-term disease control advantage fails to translate effectively into long-term survival benefit. Nevertheless, quantitative evidence supporting this hypothesis at the trial level is lacking.\u003c/p\u003e \u003cp\u003eOn the basis of the clinical and methodological flaws outlined above, a systematic trial-level meta-analysis with the goal of achieving integration across three levels. First, we used a formal interaction effect model to test whether age serves as an independent effect modifier of the efficacy of CDK4/6 inhibitors. Second, we quantitatively evaluated the surrogate value of PFS for OS at the trial level and compared its stability across different age subgroups. Third, we introduced a Bayesian hierarchical random-effects model to verify the robustness of the interaction effect from a probabilistic perspective. By integrating efficacy heterogeneity analysis with an endpoint evaluation framework, this study aims to clarify the role of age in the process of translating disease control to survival benefit and to reassess the applicability of PFS in the elderly population, thereby providing a more explanatory and generalizable evidence base for clinical decision-making and future trial design.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Literature Search and Research Selection\u003c/h2\u003e \u003cp\u003e This study strictly adhered to the reporting guidelines for systematic reviews and meta-analyses (PRISMA 2020). The PubMed, EMBASE, and Cochrane Library databases were systematically searched from their inception to February 2026, without any language restrictions. Additionally, the abstract databases of major international oncology conferences (ASCO and ESMO) were searched to reduce the risk of publication bias. Prospective registration of the study protocol was completed in the PROSPERO database (registration number: CRD420261345790).\u003c/p\u003e \u003cp\u003eThe search strategy combined medical subject headings (MeSH) and free-text terms, with the main keywords encompassing \u0026ldquo;CDK4/6 inhibitors,\u0026rdquo; \u0026ldquo;breast cancer,\u0026rdquo; \u0026ldquo;randomized controlled trial,\u0026rdquo; and was adapted according to the specifications of each database. In addition, we conducted a manual search on the reference lists of included studies and relevant review articles to enhance the completeness of the search.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Inclusion and Exclusion Criteria\u003c/h2\u003e \u003cp\u003eCharacteristics of included studies: (1) the study was a phase III randomized controlled trial; (2) it compared CDK4/6 inhibitors combined with endocrine therapy versus endocrine therapy alone; (3) the hazard ratio (HR) and 95% CI of PFS and/or OS were reported; (4) it provided efficacy data stratified by age (\u0026lt;\u0026thinsp;65 years vs. \u0026ge; 65 years). Studies that did not directly report stratified HRs in the original manuscript but allowed derivation from public data or supplementary materials were also included.\u003c/p\u003e \u003cp\u003eNonrandomized studies, duplicate publications, and studies from which age-stratified effect sizes could not be extracted were excluded. For multiple updated reports of the same trial, only the version with the longest follow-up duration and the most mature data was included.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Data Extraction and Quality Evaluation\u003c/h2\u003e \u003cp\u003eTwo researchers independently extracted and cross-checked the data, including trial name, study design, sample size, treatment regimen, median follow-up duration, and HRs for PFS and OS (overall and age subgroups). For key outcome data, the two researchers independently entered and checked the data to reduce human error.\u003c/p\u003e \u003cp\u003eDisagreements were resolved through discussion or adjudicated by a third party. The quality of the included studies was systematically assessed using the Cochrane RoB 2.0 tool, which focuses on the randomization process, allocation concealment, biases from intended interventions, measurement of the outcome, and selection of the reported result. The overall risk of bias for each study was graded as low risk, some concerns, or high risk to inform subsequent sensitivity analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Traditional Meta-analysis (PFS and OS)\u003c/h2\u003e \u003cp\u003eA random-effects model was used to pool the results for PFS and OS to account for clinical and methodological heterogeneity across studies. The between-study variance (τ\u0026sup2;) was estimated through the restricted maximum likelihood (REML) method, and the Hartung\u0026ndash;Knapp\u0026ndash;Sidik\u0026ndash;Jonkman method was applied to adjust the confidence intervals, thereby improving robustness in small-sample settings.\u003c/p\u003e \u003cp\u003eEffect sizes were expressed as HRs, analyzed on a logarithmic scale (log HR), and weighted by inverse variance. To enhance the robustness and interpretability of the results, the 95% prediction interval was also reported to reflect the range of effect sizes that could be expected in future studies.\u003c/p\u003e \u003cp\u003eHeterogeneity was assessed using the I\u0026sup2; statistic and Cochran's Q test. When I\u0026sup2; exceeded 50%, potential sources were further explored. However, given that this study was primarily based on trial-level analyses, multivariate meta-regression was not performed to avoid overfitting.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Age Interaction Analysis\u003c/h2\u003e \u003cp\u003eTo avoid informal comparisons based on superficial differences in traditional subgroup analyses, a formal interaction effect analysis framework was constructed at the trial level to directly test whether age serves as a modifier of the treatment effect. The interaction effect was defined as the ratio of HRs between different age subgroups:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:H{R}_{interaction}=\\frac{H{R}_{\\ge\\:65}}{H{R}_{\u0026lt;65}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eCalculates on the logarithmic scale:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:\\text{l}\\text{o}\\text{g}\\left(H{R}_{interaction}\\right)=\\text{l}\\text{o}\\text{g}\\left(H{R}_{\\ge\\:65}\\right)-\\text{l}\\text{o}\\text{g}\\left(H{R}_{\u0026lt;65}\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThe standard error was derived from the variance of the log hazard ratios of the two subgroups. The interaction effects from each trial were pooled using a random-effects model (REML), and the Hartung-Knapp method was used to adjust the interval estimation. HR_interaction\u0026thinsp;\u0026gt;\u0026thinsp;1 indicates that younger patients derive greater benefit, whereas \u003cb\u003e\u0026lt;\u003c/b\u003e\u0026thinsp;1 indicates a greater benefit in older patients. This approach enables a formal statistical test of the modifying effect of age and fundamentally avoids the biases associated with insufficient sample size or random error in traditional subgroup analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Assessment of PFS as a Surrogate Endpoint for OS\u003c/h2\u003e \u003cp\u003eOn the basis of the trial-level surrogate endpoint analysis framework, the predictive ability of PFS for OS was evaluated. The HRs for PFS and OS from each trial were log(HR) transformed and used to construct a weighted linear regression model:\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:\\text{l}\\text{o}\\text{g}\\left(H{R}_{OS}\\right)=\\alpha\\:+\\beta\\:\\cdot\\:\\text{l}\\text{o}\\text{g}\\left(H{R}_{PFS}\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eRegression analysis was performed using inverse variance weighting and sample size weighting as sensitivity analyses. The coefficient of determination (R\u0026sup2;) was used to quantify the explanatory power of PFS with respect to OS variability. Further modeling was conducted in the overall population, as well as in the \u0026lt;\u0026thinsp;65 years and \u0026ge;\u0026thinsp;65 years age subgroups, to compare the stability of surrogate endpoint validity across different age groups.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Bayesian Hierarchical Random-effects Model\u003c/h2\u003e \u003cp\u003eTo verify the interaction effect results from a probabilistic perspective and to characterize uncertainty, a Bayesian hierarchical random-effects model was constructed to model the log(HR_interaction) [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{y}_{i}\\sim\\:N({\\theta\\:}_{i},{\\sigma\\:}_{i}^{2})\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003cp\u003eThe requirements at the research level are as follows:\u003cdiv id=\"Equd\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equd\" name=\"EquationSource\"\u003e\n$$\\:{\\theta\\:}_{i}\\sim\\:N(\\mu\\:,{\\tau\\:}^{2})$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:{y}_{i}\\)\u003c/span\u003e \u003c/span\u003e represents the log(HR_interaction) for each study, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\sigma\\:}_{i}^{2}\\)\u003c/span\u003e\u003c/span\u003e is its variance, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:\\)\u003c/span\u003e\u003c/span\u003e is the overall effect, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{2}\\:\\)\u003c/span\u003e\u003c/span\u003erepresents the between-study heterogeneity. Weakly informative priors were used as follows:\u003cdiv id=\"Eque\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Eque\" name=\"EquationSource\"\u003e\n$$\\:\\mu\\:\\sim\\:N\\left(0,{10}^{2}\\right)$$\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equf\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equf\" name=\"EquationSource\"\u003e\n$$\\:\\tau\\:\\sim\\:\\text{Half-Cauchy}\\left(\\text{0,1}\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eParameters were estimated by the Markov chain Monte Carlo (MCMC) method (four chains, each with 4,000 iterations, including 2,000 warm-up iterations). Convergence was assessed through trace plots and the Gelman\u0026ndash;Rubin statistic (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\widehat{R}\\)\u003c/span\u003e\u003c/span\u003e). The reported as medians with 95% confidence intervals (CIs). The posterior probability that the HR_interaction exceeds 1 was calculated to directly quantify the strength of evidence for effect modification by age, thereby addressing the limitations associated with the interpretation of traditional p-values.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Sensitivity Analysis\u003c/h2\u003e \u003cp\u003eThe robustness of the results was evaluated using the leave-one-out method, which involves removing individual studies one by one and repeating the analysis to observe changes in the pooled effects.\u003c/p\u003e \u003cp\u003eIn addition, a multidimensional sensitivity analysis was prespecified, including restrictions to trials reporting mature OS data, application of different weighting strategies (inverse variance and sample size), and exclusion of studies with a high risk of bias. Moreover, the results from the fixed-effects model and the random-effects model were compared to evaluate the impact of model specification on the conclusions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Statistical Analysis\u003c/h2\u003e \u003cp\u003eAll frequentist analyses were performed using R software (version 4.3.0), primarily employing the \u0026ldquo;meta\u0026rdquo; and \u0026ldquo;metafor\u0026rdquo; packages. Bayesian analyses were implemented using \u0026ldquo;rstan\u0026rdquo; or \u0026ldquo;brms\u0026rdquo;. All statistical tests were two-sided, and \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered indicate statistical significance. The results from the Bayesian analyses did not rely on significance thresholds but were interpreted based on effect distributions and posterior probabilities.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Study Selection and Characteristics\u003c/h2\u003e \u003cp\u003eSix phase III randomized controlled trials (PALOMA-2, PALOMA-3, MONALEESA-2, MONALEESA-3, MONARCH-2, and MONARCH-3) were included in this study to compare the efficacy of CDK4/6 inhibitors combined with endocrine therapy versus endocrine therapy alone in patients with HR\u003csup\u003e+\u003c/sup\u003e/HER2\u003csup\u003e\u0026minus;\u003c/sup\u003e metastatic breast cancer [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan additionalcitationids=\"CR22\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. All trials reported PFS in the overall population and provided effect sizes for PFS and OS stratified by age (\u0026lt;\u0026thinsp;65 years and \u0026ge;\u0026thinsp;65 years), thereby meeting the preconditions for interaction effect analysis and surrogate endpoint analysis. Overall, there were certain differences among the studies in terms of treatment line, endocrine regimen, and follow-up duration; however, the methodological quality was generally high, and the risk of bias was mainly low or with some concerns (Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). The study screening process is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, and baseline characteristics are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCharacteristics of Included Phase III Trials\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTrial\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDrug\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLine of Therapy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEndocrine Partner\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMenopausal Status\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMedian Age (range)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;65 (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eFollow-up (months)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePALOMA-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePalbociclib\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFirst-line\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLetrozole\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePostmenopausal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e666\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62 (28\u0026ndash;89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e79.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePALOMA-3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePalbociclib\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSecond-line\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFulvestrant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAny\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e521\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62 (30\u0026ndash;88)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e73.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMONALEESA-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRibociclib\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFirst-line\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLetrozole\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePostmenopausal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e668\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62 (32\u0026ndash;91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e80.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMONALEESA-3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRibociclib\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFirst-/Second-line\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFulvestrant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePostmenopausal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e726\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e63 (30\u0026ndash;88)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e56.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMONARCH-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAbemaciclib\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSecond-line\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eFulvestrant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAny\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e669\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e60 (32\u0026ndash;91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e47.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMONARCH-3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAbemaciclib\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFirst-line\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNSAI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePostmenopausal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e493\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62 (36\u0026ndash;91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e96.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eFootnote: NSAI, nonsteroidal aromatase inhibitor. All included studies were phase III randomized controlled trials comparing CDK4/6 inhibitors plus endocrine therapy versus endocrine therapy alone.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Consistent PFS Benefit across Different Age Groups\u003c/h2\u003e \u003cp\u003eCDK4/6 inhibitors produce highly consistent and reproducible disease control effects across different age groups. In the overall population, treatment significantly prolonged PFS (HR\u0026thinsp;=\u0026thinsp;0.54, 95% CI 0.50\u0026ndash;0.59), with no observed between-study heterogeneity (I\u0026sup2; = 0%) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In the age-stratified analysis, the HR was 0.54 (95% CI 0.49\u0026ndash;0.59) for patients aged\u0026thinsp;\u0026lt;\u0026thinsp;65 years and 0.52 (95% CI 0.46\u0026ndash;0.60) for those aged\u0026thinsp;\u0026ge;\u0026thinsp;65 years. The effect estimates for the two subgroups largely overlapped, and both demonstrated very low heterogeneity (I\u0026sup2; = 0%). The prediction intervals were also highly consistent between these two subgroups (\u0026lt;\u0026thinsp;65 years: 0.49\u0026ndash;0.60; \u0026ge;65 years: 0.46\u0026ndash;0.61), suggesting that the inhibitory effect of this treatment has good generalizability and stability across different age groups. Overall, no effect modification by age was observed for PFS benefit (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAge-Stratified Efficacy and Interaction Effects of CDK4/6 Inhibitors\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAge Group\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHR (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eI\u0026sup2; (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePrediction Interval\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePFS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.54 (0.50\u0026ndash;0.59)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.48\u0026ndash;0.62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;65 years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.54 (0.49\u0026ndash;0.59)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.47\u0026ndash;0.62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;65 years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.52 (0.46\u0026ndash;0.60)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.44\u0026ndash;0.62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.75 (0.67\u0026ndash;0.83)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.65\u0026ndash;0.86\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;65 years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.74 (0.66\u0026ndash;0.83)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.64\u0026ndash;0.86\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;65 years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.88 (0.75\u0026ndash;1.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.73\u0026ndash;1.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInteraction (OS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;65 vs\u0026thinsp;\u0026lt;\u0026thinsp;65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.18 (1.06\u0026ndash;1.32)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.94\u0026ndash;1.50\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Age-dependent Divergence in Overall Survival Benefit\u003c/h2\u003e \u003cp\u003eAlthough the PFS benefit remained consistent across different age groups, OS improvement showed potential age-related heterogeneity. Based on an interaction effect analysis at the trial level, the pooled interaction HR was 1.18 (95% CI 1.06\u0026ndash;1.32, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.011), indicating that younger patients (\u0026lt;\u0026thinsp;65 years) may derive greater survival benefit. This finding provides direct statistical evidence for age as an effect modifier of treatment efficacy, rather than a mere observation of subgroup differences. Although the prediction interval ranged from 0.94 to 1.50, the interaction effects were highly consistent across studies (I\u0026sup2; = 0%, τ\u0026sup2; \u0026asymp; 0), suggesting that a degree of uncertainty will remain in future studies (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Additionally, it should be noted that the fixed-effects model did not yield statistical significance (HR\u0026thinsp;=\u0026thinsp;1.18, 95% CI 0.99\u0026ndash;1.42), but the results became significant after the random-effects model with the Hartung-Knapp correction, underscoring the importance of using robust estimation methods when the number of trials is limited.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Structural Weakening of the PFS\u0026ndash;OS Surrogate Relationships in Elderly Patients\u003c/h2\u003e \u003cp\u003eIn the overall population, a strong correlation was observed between PFS and OS (R\u0026sup2; = 0.81, 95% CI 0.65\u0026ndash;0.91), supporting the overall validity of PFS as a surrogate endpoint for OS. However, the correlation differed markedly after age stratification: the R\u0026sup2; was 0.79 (95% CI 0.62\u0026ndash;0.90) in the subgroup aged\u0026thinsp;\u0026lt;\u0026thinsp;65 years, whereas it decreased substantially to 0.34 (95% CI 0.11\u0026ndash;0.58) in patients aged\u0026thinsp;\u0026ge;\u0026thinsp;65 years (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). These findings indicate that in the elderly population, PFS explains only approximately one-third of the variation in OS, indicating a structural weakening of its validity as a surrogate endpoint (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Further analysis revealed a significantly reduced regression slope in the older subgroup, reflecting a great quantitative weakening of the predictive strength of PFS for OS. Collectively, these results support the phenomenon of \u0026ldquo;endpoint decoupling,\u0026rdquo; whereby the advantage of short-term disease control cannot be effectively translated into long-term survival benefit.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAge-Dependent Breakdown of PFS as a Surrogate for Overall Survival\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePopulation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR\u0026sup2; (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInterpretation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.81 (0.65\u0026ndash;0.91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStrong surrogacy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;65 years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.79 (0.62\u0026ndash;0.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStrong surrogacy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;65 years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.34 (0.11\u0026ndash;0.58)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePoor surrogacy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eFootnote: Surrogacy was assessed using weighted linear regression of log(HR_PFS) on log(HR_OS) at the trial level. R\u0026sup2; represents the proportion of variance in OS treatment effect explained by PFS. Higher R\u0026sup2; indicates stronger surrogate validity.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Bayesian Analysis-based Probabilistic Validation of Age Interaction Effects\u003c/h2\u003e \u003cp\u003eA Bayesian hierarchical random-effects model further strengthened the evidence for the age interaction effect from a probabilistic perspective. The analysis revealed that the posterior median of the interaction HR is 1.18 (95% CI 1.03\u0026ndash;1.34), which was highly consistent with the frequentist analysis. The posterior probability of interaction HR\u0026thinsp;\u0026gt;\u0026thinsp;1 was 94%, indicating that the conclusion that younger patients derive greater benefit is strongly supported in probabilistic terms. The posterior distribution was predominantly concentrated in the region of HR\u0026thinsp;\u0026gt;\u0026thinsp;1, with only approximately 5.8% of the probability mass falling in the range of HR\u0026thinsp;\u0026lt;\u0026thinsp;1, thereby providing a more robust inferential basis under the condition of a limited number of trials (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Sensitivity Analysis\u003c/h2\u003e \u003cp\u003eMultidimensional sensitivity analysis further verified the robustness of the main results. Leave-one-out analyses demonstrated that sequential exclusion of individual trials yielded pooled interaction HRs ranging from 1.14 to 1.20, with all estimates consistently exceeding unity, closely aligning with the overall effect, and indicating that the observed age-dependent interaction was not driven by any single study (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e,). In addition, after limiting trials to mature OS data, applying different weighting strategies, and excluding studies with a high risk of bias, the main conclusions did not change substantially (Supplementary Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e), further supporting the reliability and reproducibility of the results in this study.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study systematically revealed age-related differences in the efficacy of CDK4/6 inhibitors in HR\u003csup\u003e+\u003c/sup\u003e/HER2\u003csup\u003e\u0026minus;\u003c/sup\u003e metastatic breast cancer at the trial level, providing two key findings of direct clinical and methodological significance. First, although progression-free survival (PFS) benefit was highly consistent across age groups, overall survival (OS) improvement tended toward age-related heterogeneity, suggesting that the relationship between disease control and survival benefit may differ by age. Second, the surrogate validity of PFS for OS appeared reduced in older patients, suggesting substantially diminished and potentially unreliable surrogacy in this population.\u003c/p\u003e \u003cp\u003eThis finding may challenge the prevailing assumption of consistent treatment effects across age groups [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Previous analyses have largely concluded that \u0026ldquo;age does not affect efficacy\u0026rdquo; based on the apparent similarity of effect sizes between subgroups. However, this inference essentially relies on informal comparisons and lacks direct statistical tests for interactions. In this study, the \u0026ldquo;age \u0026times; treatment effect\u0026rdquo; was quantitatively evaluated by an interaction effect model at trial-level, and the posterior probability of the effect direction was directly calculated within a Bayesian framework. This combined approach may enhance the interpretability of effect modification under conditions of limited trial numbers.\u003c/p\u003e \u003cp\u003eFrom a mechanistic perspective, the observed divergence between PFS and OS may not be incidental but could reflect age-related changes in outcome structure and treatment pathways. First, the structural impact of competing risks likely represents an important contributing factor. As age increases, non-tumor-related deaths increase significantly, transforming OS from a single tumor outcome into a composite endpoint comprising multiple events. Under this structure, gains in tumor control may be progressively diluted at the level of OS, resulting in an apparent attenuation of survival benefit despite preserved PFS effects. Second, age-related differences in tumor biological characteristics may also contribute to this process. Older patients may present with different tumor proliferation patterns, which could alter the relative contribution of cell cycle inhibition to long-term survival. Consequently, the effects of CDK4/6 inhibitors may be primarily disease-stabilizing rather than survival-prolonging. Together, these factors establish a multidimensionally driven mechanistic framework for decoupling of endpoints.\u003c/p\u003e \u003cp\u003eOn the basis of the above mechanism, the results of this study have direct implications for clinical practice and trial design. First, in clinical practice, CDK4/6 inhibitors may still provide stable disease control in elderly patients, but their true survival benefit may be overestimated. This suggests that treatment decisions should not rely solely on PFS, but instead comprehensively consider life expectancy, competing risks, and treatment-related burden to avoid the potential overtreatment that arises from \u0026ldquo;replacing survival benefit with progression delay.\u0026rdquo;\u003c/p\u003e \u003cp\u003eIn terms of trial design and drug evaluation, these results may question the universal applicability of PFS as a primary endpoint. Although PFS still has high surrogate value in the overall population, its explanatory power decreases significantly in the elderly subgroup, in whom no stable predictive relationship is observed. These findings suggest that the assumption of PFS as a universally valid surrogate endpoint may not fully true across all patient subgroups. Future studies should prioritize OS or other more clinically relevant composite endpoints, and incorporate interaction effect testing and competing risk modeling at the design stage to improve the generalizability and clinical interpretability of the evidence.\u003c/p\u003e \u003cp\u003eSeveral limitations should be acknowledged. First, this study is based on trial-level data rather than individual patient data, and therefore could not make fine adjustments for key confounders such as comorbidity burden, functional status, and treatment exposure intensity. Second, the limited number of included trials may affect the precision and stability of certain estimates. Accordingly, the findings should be interpreted as hypothesis-generating rather than definitive. In addition, the use of 65 years as the age cutoff was primarily dictated by trial reporting conventions and may not fully capture biological aging or functional heterogeneity. These limitations themselves highlight a structural gap in the current evidence base\u0026mdash;namely, that the stability of a single endpoint framework cannot be assumed in highly heterogeneous populations.\u003c/p\u003e \u003cp\u003eIn conclusion, this study demonstrates age-dependent differences in the translation of PFS into OS benefit with CDK4/6 inhibitors, with preserved PFS but attenuated OS in older patients, accompanied by reduced surrogate validity. These findings challenge the universal applicability of PFS and support the need for population-specific endpoint frameworks.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Springer Nature Author Services for their English language editing services, which improved the clarity and quality of this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWK conceived the study and designed the methodology. SL and LH performed the statistical analyses. All authors contributed to data interpretation and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analyzed during this study are included in this published article and its supplementary files.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Natural Science Foundation of Zhejiang Province (Grant No. LMRY26H17001).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data generated in this study are available upon request from the corresponding author.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMedian MD, Antone NZ, Volovăț S, Mazilu L, Negru M, Curcă RO, Niță A, Pătru RI, Ungureanu A, Lupu V, et al. Romanian Consensus Statement for Hormone Receptor-Positive and Human Epidermal Growth Factor Receptor 2-Negative Metastatic Breast Cancer (HR+/HER2- mBC) and Triple-Negative Metastatic Breast Cancer (mTNBC). Curr Oncol (Toronto Ont). 2026;33(2):120.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMavratzas A, Marm\u0026eacute; F. Alpelisib in the treatment of metastatic HR+ breast cancer with PIK3CA mutations. Future Oncol (London England). 2021;17(1):13\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoetz MP, Toi M, Huober J, Sohn J, Tr\u0026eacute;dan O, Park IH, Campone M, Chen S, Manso LM, Paluch-Shimon S, et al. Abemaciclib plus a nonsteroidal aromatase inhibitor as initial therapy for HR+, HER2- advanced breast cancer: final overall survival results of MONARCH 3. Annals oncology: official J Eur Soc Med Oncol. 2024;35(8):718\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHortobagyi GN, Stemmer SM, Burris HA, Yap Y, Sonke GS, Hart L, Campone M, Petrakova K, Winer EP, Janni W, et al. Overall Survival with Ribociclib plus Letrozole in Advanced Breast Cancer. N Engl J Med. 2022;386(10):942\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFinn RS, Martin M, Rugo HS, Jones S, Im S, Gelmon K, Harbeck N, Lipatov ON, Walshe JM, Moulder S, et al. Palbociclib and Letrozole in Advanced Breast Cancer. N Engl J Med. 2016;375(20):1925\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHortobagyi GN, Stemmer SM, Burris HA, Yap Y, Sonke GS, Paluch-Shimon S, Campone M, Blackwell KL, Andr\u0026eacute; F, Winer EP, et al. Ribociclib as First-Line Therapy for HR-Positive, Advanced Breast Cancer. N Engl J Med. 2016;375(18):1738\u0026ndash;48.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlamon DJ, Neven P, Chia S, Jerusalem G, De Laurentiis M, Im S, Petrakova K, Valeria Bianchi G, Mart\u0026iacute;n M, Nusch A, et al. Ribociclib plus fulvestrant for postmenopausal women with hormone receptor-positive, human epidermal growth factor receptor 2-negative advanced breast cancer in the phase III randomized MONALEESA-3 trial: updated overall survival. Annals oncology: official J Eur Soc Med Oncol. 2021;32(8):1015\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSledge GWJ, Toi M, Neven P, Sohn J, Inoue K, Pivot X, Burdaeva O, Okera M, Masuda N, Kaufman PA, et al. The Effect of Abemaciclib Plus Fulvestrant on Overall Survival in Hormone Receptor-Positive, ERBB2-Negative Breast Cancer That Progressed on Endocrine Therapy-MONARCH 2: A Randomized Clinical Trial. JAMA ONCOL. 2020;6(1):116\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlamon DJ, Neven P, Chia S, Fasching PA, De Laurentiis M, Im S, Petrakova K, Bianchi GV, Esteva FJ, Mart\u0026iacute;n M, et al. Overall Survival with Ribociclib plus Fulvestrant in Advanced Breast Cancer. N Engl J Med. 2020;382(6):514\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIm S, Lu Y, Bardia A, Harbeck N, Colleoni M, Franke F, Chow L, Sohn J, Lee K, Campos-Gomez S, et al. Overall Survival with Ribociclib plus Endocrine Therapy in Breast Cancer. N Engl J Med. 2019;381(4):307\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFreedman RA, Tolaney SM. Efficacy and safety in older patient subsets in studies of endocrine monotherapy versus combination therapy in patients with HR+/HER2- advanced breast cancer: a review. BREAST CANCER RES TR. 2018;167(3):607\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePilehvari A, Kimmick G, You W, Bonilla G, Anderson R. Disparities in receipt of 1-(st) line CDK4/6 inhibitors with endocrine therapy for treatment of hormone receptor positive, HER2 negative metastatic breast cancer in the real-world setting. Breast cancer research: BCR. 2024;26(1):144.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeluche E, Antoine A, Bachelot T, Lardy-Cleaud A, Dieras V, Brain E, Debled M, Jacot W, Mouret-Reynier MA, Goncalves A et al. Contemporary outcomes of metastatic breast cancer among 22,000 women from the multicentre ESME cohort 2008\u0026ndash;2016. \u003cem\u003eEuropean journal of cancer (Oxford, England\u003c/em\u003e: 1990) 2020, 129:60\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTurner NC, Ro J, Andr\u0026eacute; F, Loi S, Verma S, Iwata H, Harbeck N, Loibl S, Huang Bartlett C, Zhang K, et al. Palbociclib in Hormone-Receptor-Positive Advanced Breast Cancer. N Engl J Med. 2015;373(3):209\u0026ndash;19.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCourtinard C, Gourgou S, Jacot W, Carton M, Gu\u0026eacute;rin O, Vacher L, Bertaut A, Le Deley M, P\u0026eacute;rol D, Marino P, et al. Association between progression-free survival and overall survival in women receiving first-line treatment for metastatic breast cancer: evidence from the ESME real-world database. BMC MED. 2023;21(1):87.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurzykowski T, Buyse M, Piccart-Gebhart MJ, Sledge G, Carmichael J, L\u0026uuml;ck H, Mackey JR, Nabholtz J, Paridaens R, Biganzoli L, et al. Evaluation of tumor response, disease control, progression-free survival, and time to progression as potential surrogate end points in metastatic breast cancer. J Clin oncology: official J Am Soc Clin Oncol. 2008;26(12):1987\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi J, Huo X, Zhao F, Ren D, Ahmad R, Yuan X, Du F, Zhao J. Association of Cyclin-Dependent Kinases 4 and 6 Inhibitors With Survival in Patients With Hormone Receptor-Positive Metastatic Breast Cancer: A Systematic Review and Meta-analysis. JAMA NETW OPEN. 2020;3(10):e2020312.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePiezzo M, Chiodini P, Riemma M, Cocco S, Caputo R, Cianniello D, Di Gioia G, Di Lauro V, Rella FD, Fusco G, et al. Progression-Free Survival and Overall Survival of CDK 4/6 Inhibitors Plus Endocrine Therapy in Metastatic Breast Cancer: A Systematic Review and Meta-Analysis. INT J MOL SCI. 2020;21(17):6400.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCanou\u0026iuml;-Poitrine F, Li\u0026egrave;vre A, Dayde F, Lopez-Trabada-Ataz D, Baumgaertner I, Dubreuil O, Brunetti F, Coriat R, Maley K, Pernot S, et al. Inclusion of Older Patients with Cancer in Clinical Trials: The SAGE Prospective Multicenter Cohort Survey. Oncologist. 2019;24(12):e1351\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClark J, Muhlemann N, Ionan A. Bayesian Clinical Trials. THER INNOV REGUL SCI. 2023;57(3):399\u0026ndash;400.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlamon DJ, Neven P, Chia S, Fasching PA, De Laurentiis M, Im S, Petrakova K, Bianchi GV, Esteva FJ, Mart\u0026iacute;n M, et al. Phase III Randomized Study of Ribociclib and Fulvestrant in Hormone Receptor-Positive, Human Epidermal Growth Factor Receptor 2-Negative Advanced Breast Cancer: MONALEESA-3. J Clin oncology: official J Am Soc Clin Oncol. 2018;36(24):2465\u0026ndash;72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSledge GWJ, Toi M, Neven P, Sohn J, Inoue K, Pivot X, Burdaeva O, Okera M, Masuda N, Kaufman PA, et al. MONARCH 2: Abemaciclib in Combination With Fulvestrant in Women With HR+/HER2- Advanced Breast Cancer Who Had Progressed While Receiving Endocrine Therapy. J Clin oncology: official J Am Soc Clin Oncol. 2017;35(25):2875\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoetz MP, Toi M, Campone M, Sohn J, Paluch-Shimon S, Huober J, Park IH, Tr\u0026eacute;dan O, Chen S, Manso L, et al. MONARCH 3: Abemaciclib As Initial Therapy for Advanced Breast Cancer. J Clin oncology: official J Am Soc Clin Oncol. 2017;35(32):3638\u0026ndash;46.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLegend.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcan","sideBox":"Learn more about [BMC Cancer](http://bmccancer.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcan/default.aspx","title":"BMC Cancer","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Breast cancer, CDK4/6 inhibitors, Age stratification, Interaction, Surrogate endpoint, Bayesian meta-analysis","lastPublishedDoi":"10.21203/rs.3.rs-9551064/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9551064/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eCDK4/6 inhibitors consistently improve progression-free survival (PFS) in HR+/HER2\u0026thinsp;\u0026minus;\u0026thinsp;metastatic breast cancer; however, whether these benefits translate into overall survival (OS) across age groups and whether PFS remains a valid surrogate endpoint in older patients remain unclear. We aimed to evaluate age-dependent treatment effects and the surrogate validity of PFS at the trial level.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003ePhase III randomized controlled trials comparing CDK4/6 inhibitors plus endocrine therapy versus endocrine therapy alone were included. Hazard ratios (HRs) for PFS and OS were extracted by age (\u0026lt;\u0026thinsp;65 vs\u0026thinsp;\u0026ge;\u0026thinsp;65 years). Pooled analyses were performed using random-effects models, with age-related effect modification assessed via trial-level interaction analyses. Trial-level surrogacy of PFS for OS was evaluated using weighted regression across age subgroups. A Bayesian hierarchical random-effects model was further applied to validate interaction effects.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eSix phase III trials were included. CDK4/6 inhibitors significantly improved PFS across age groups (HR\u0026thinsp;\u0026asymp;\u0026thinsp;0.55) with minimal heterogeneity (I\u0026sup2;=0%), and no significant age-related interaction was observed for PFS. In contrast, OS benefit showed significant age-dependent attenuation (interaction HR\u0026thinsp;=\u0026thinsp;1.18, 95% CI 1.06\u0026ndash;1.32). At the trial level, PFS strongly correlated with OS in the overall population (R\u0026sup2;=0.81), but this association was markedly reduced in patients aged\u0026thinsp;\u0026ge;\u0026thinsp;65 years (R\u0026sup2;=0.34). Bayesian analysis confirmed a high probability (94%) that younger patients derived greater OS benefit.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eAlthough PFS benefit is preserved, its translation into OS is attenuated in older patients, indicating that PFS may be an unreliable surrogate and could overestimate true survival benefit in this population.\u003c/p\u003e","manuscriptTitle":"Age-Dependent Survival Benefit and Surrogacy of Progression-Free Survival in CDK4/6 Inhibitor Trials for HR+/HER2− Metastatic Breast Cancer: A Trial-Level Meta-analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-18 09:02:45","doi":"10.21203/rs.3.rs-9551064/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-21T13:24:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"129866967169763169886388284740413875025","date":"2026-05-17T01:47:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"249882350025848698632549506737493352355","date":"2026-05-15T06:06:28+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"303034478695099444032560668300974862020","date":"2026-05-12T04:38:15+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-07T21:01:58+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-29T17:04:59+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-29T00:08:30+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-29T00:07:40+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Cancer","date":"2026-04-28T08:22:26+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcan","sideBox":"Learn more about [BMC Cancer](http://bmccancer.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcan/default.aspx","title":"BMC Cancer","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"5e449275-aa41-4119-8716-580bbc2de11d","owner":[],"postedDate":"May 18th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-21T13:24:01+00:00","index":102,"fulltext":""},{"type":"reviewerAgreed","content":"129866967169763169886388284740413875025","date":"2026-05-17T01:47:36+00:00","index":97,"fulltext":""},{"type":"reviewerAgreed","content":"249882350025848698632549506737493352355","date":"2026-05-15T06:06:28+00:00","index":93,"fulltext":""},{"type":"reviewerAgreed","content":"303034478695099444032560668300974862020","date":"2026-05-12T04:38:15+00:00","index":57,"fulltext":""},{"type":"reviewersInvited","content":"58","date":"2026-05-07T21:01:58+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-18T09:02:45+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-18 09:02:45","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9551064","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9551064","identity":"rs-9551064","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}