Abstract P2-08-01: Factors predicting relapse in early breast cancer patients with a pathological complete response after neoadjuvant therapy – Results of a pooled analysis based on the GBG meta-database

Abstract

Abstract Background Even though patients with a pCR following neoadjuvant chemotherapy have an excellent prognosis still some of these patients will eventually relapse. A better identification of pts with an increased risk of relapse despite a pCR would be helpful to select these patients for additional post-neoadjuvant treatment strategies. Thus, the rationale of this retrospective analysis was to identify factors predicting relapse despite a pCR. Methods This pooled retrospective analysis based on the GBG meta-database includes the neoadjuvant trials GeparTrio, GeparQuattro, GeparQuinto, GeparSixto and GeparSepto. In these trials 2188 (27%) of 7933 pts had a pCR according to ypT0/ypTis ypN0 Definition and were included. The primary endpoint was disease-free survival (DFS), secondary endpoints were distant DFS (DDFS) and overall survival (OS). A multivariate Cox proportional hazards model was used to report hazard ratios with 95% confidence interval (CI). The two-sided significance level was set to α=0.05. Endpoints were analysed for all pts and in subgroups defined by intrinsic subtypes. The potential risk factors intrinsic subtype (HER2 negative/hormone receptor (HR) positive, triple negative, HER2 positive/HR positive, HER2 positive/HR negative), histological tumor type (lobular vs other), grade (G1/G2 vs G3), KI67 (≤20% vs higher), initial cT and cN stadium (cT1 vs cT2 vs cT3/4; cN0 vs cN+), age (≤40 vs 41-59 vs ≥60), BMI (< 25 vs 25-29 vs ≥ 30), planned number of cycles of chemotherapy (≤6 vs > 6), menopausal status (pre- vs postmenopausal) and clinical response after 2-4 cycles (SD vs PR vs CR vs PD) were included as covariates in multivariate Cox regression models as well as study identification. Results From 2188 evaluable patients DFS, DDFS and OS events were observed in 290/197/130 pts respectively; the median follow-up over all studies was 59 months. In multivariate analysis including study and all potential risk factors DFS was significantly different with regard to the initial cN status (cN+ vs cN0, hazard ratio (HR) 1.70; 95% CI [1.2, 2.4], p=0.002). Of borderline significance was histological type (non-lobular vs lobular, HR 0.52 95% CI [0.3, 1.1]; p=0.076) and initial tumor stage (cT3/4 vs cT1, HR 1.61 95% CI [1.0, 2.7]; p=0.064). In terms of DDFS significant differences were seen for the initial cN status (cN+ vs cN0, HR 2.34; 95% CI [1.5, 3.6], p<0.001) and initial tumor stage (cT3/4 vs cT1, HR 1.83 95% CI [1.0, 3.3]; p=0.044); histological type was again close to significance (non-lobular vs lobular, HR 0.46 95% CI [0.2, 1.1]; p=0.067). Multivariate analysis showed significantly worse OS with initial cT3/4 tumors (cT3/4 vs cT1, HR 2.48 95%CI [1.1, 5.7]; p=0.030), and the lobular type (non-lobular vs lobular, HR 0.35 95% CI [0.1, 0.9]; p=0.026) and a trend for worse OS in pts with cN+ (cN+ vs cN0, HR 1.67 95% CI [1.0, 2.9]; p=0.067). Conclusions Initial tumor load before start of neoadjuvant chemotherapy (tumor stage and nodal status) and lobular subtype were predictors of long term outcome after a pCR following neoadjuvant chemotherapy. Intrinsic subtype, KI67, grade and planned number of cycles were not predictive for a relapse. Citation Format: Huober J, Schneeweiss A, Blohmer J-U, Denkert C, Tesch H, Hanusch CA, Salat C, Rhiem K, Rezai M, Solbach C, Fasching PA, Jackisch C, Mehta K, Nekljudova V, Seither F, von Minckwitz G, Loibl S, Untch M. Factors predicting relapse in early breast cancer patients with a pathological complete response after neoadjuvant therapy – Results of a pooled analysis based on the GBG meta-database [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-08-01.