Distant disease-free survival as a surrogate endpoint for overall survival in randomised trials of neoadjuvant therapy for early breast cancer: a pooled analysis of GBG and AGO-B Study Group trials

Background The surrogacy value of distant disease-free survival for overall survival has not been validated in neoadjuvant randomised controlled trials (RCTs) for early breast cancer. Here, we assess the trial-level surrogacy value of distant disease-free survival for overall survival. Methods In this pooled analysis, we included individual patient data from RCTs of neoadjuvant therapy for early breast cancer conducted by the German Breast Group (GBG) and the German Gynecological Oncology Breast Study Group (AGO-B) with available data on distant disease-free survival and overall survival. We used the trial-level measure of surrogacy R 2trial from two-stage meta-analytical copula methods to quantify the association between treatment effects on overall survival and distant disease-free survival, overall and in prespecified clinical and pathological subgroups. According to ReSEEM guidelines, R 2trial values of 0·7 or higher represent strong correlations, values between 0·69 and 0·5 represent moderate correlations, and values of less than 0·5 represent weak correlations. Findings 11 RCTs, with a total of 15 neoadjuvant treatment comparisons and 12 247 patients, were included in the analysis. Overall, there was a strong association between copula model-based hazard ratios (HRs) for overall survival and copula model-based HRs for distant disease-free survival ( R 2trial=0·91 [95% CI 0·82–1·00]). No significant heterogeneity of results was observed across the majority of subgroups analysed (p heterogeneity >0·05 in all subgroups), with the exception of subgroups defined by tumour molecular features, such as tumour progesterone receptor status, HER2 status, and molecular subtypes. For molecular subtypes, the R 2trial for the association between distant disease-free survival and overall survival was higher than 0·7, indicating strong surrogacy in hormone receptor-negative and HER2-negative tumours ( R 2trial=0·89 [95% CI 0·75–1·00]) and hormone receptor-negative and HER2-positive tumours (0·73 [0·36–1·00]), and below the 0·5 threshold for weak surrogacy in hormone receptor-positive and HER2-negative tumours (0·33 [0·00–0·83]) and hormone receptor-positive and HER2-positive tumours (0·11 [0·00–0·55]; p heterogeneity =0·021). Interpretation With adequate follow-up, distant disease-free survival is a robust surrogate endpoint for predicting final overall survival outcomes in neoadjuvant RCTs for early breast cancer in most contexts. However, the distant disease-free survival surrogacy appears to be weak for the hormone receptor-positive and HER2-negative and for the hormone receptor-positive and HER2-positive molecular subtypes. These latter findings warrant further investigation in more recent RCTs enrolling higher-risk patient populations. Funding None.