Abstract RF2-02: Pooled analysis of the BrighTNess, CALGB 40603 (Alliance), and GeparSixto clinical trials identifies the impact of neoadjuvant carboplatin on pCR and survival in early-stage triple-negative breast cancer

We performed a pooled analysis of three randomized early-stage TNBC clinical trials that investigated the addition of carboplatin to NACT: BrighTNess (NCT02032277, n = 471), CALGB 40603 (NCT00861705, n = 351), and GeparSixto (NCT01426880, n = 262). We evaluated its impact on pCR, event-free survival (EFS), and overall survival (OS), and we explored germline BRCA1 and BRCA2 (gBRCA) mutation status as a potential biomarker. Additionally, we examined the prognostic and predictive value of eight published gene expression signatures (IgG, CD4, CD8, B-cell T-cell cooperativity, CD274, Immune1, Hypoxia core, VEGF). To analyze associations with pCR, we fit logistic mixed-effects regression models with study as a random effect. Associations with EFS and OS were evaluated using Cox proportional hazards models stratified by study. Multivariate models included age, tumor size, nodal status, gBRCA mutation status, and tumor grade as covariates.

For the pooled dataset (n = 1084), in multivariate models, the addition of neoadjuvant carboplatin was significantly associated with increased pCR rate (odds ratio [OR] = 1.89, 95% confidence interval [CI] = 1.41-2.55, p < 0.001) and EFS (hazard ratio [HR] = 0.71, 95% CI = 0.54-0.93, p = 0.01), but not OS (HR = 0.93, 95% CI = 0.65-1.31, p = 0.66). When considering only gBRCA mutant samples (n = 137), we found that the addition of carboplatin conveys a significant EFS benefit (HR = 0.50, 95% CI = 0.25-1.00, p = 0.05) but had no significant impact on pCR rate or OS. Among the eight gene expression signatures tested, all six immune signatures were associated with a higher pCR rate in multivariate models, with Benjamini-Hochberg (B-H) adjusted p < 0.05. Four signatures (IgG, CD8, CD274, and Immune1) were associated with improved EFS and OS (B-H adjusted p < 0.05). However, no interaction between carboplatin and any tested signature was significantly predictive of pCR, EFS, or OS. Whole transcriptome analyses are ongoing to identify gene expression features that may predict response to neoadjuvant carboplatin.