Impact of dendritic cell vaccines added to neoadjuvant CT on pathological complete responses in early breast cancer patients according to PD-L1 expression.

Abstract

585 Background: Breast cancer (BC) in early stages exhibit a naïve and competent immune system that translates into a more prominent TIL infiltration and higher PD-L1 expression as compared to the advanced BC scenario were immunoescape and exhaustion are more prevalent. Expression of PD-L1 has been related to a better pCR when immune checkpoint inhibitors (IPI) have been added to neaodjuvant chemotherapy (NACT) in triple negative BC (TNBC). Our prior results shown dendritic cells vaccines (DCV) increased pCR in both TNBC and luminal B subtypes, with an absolute gain of 20% (p = 0.03) and a safe tolerance. Methods: Eighty-three HER2 negative BC patients with untreated stage II-III were included: 39 patients from the NCT01431196 trial that combine NACT with autologous DCV and 44 patients from a historic control group treated with the same NACT alone. NACT consists of dose dense Epirubicin plus Cyclophosphamide for 4 cycles sequenced to Docetaxel for 4 cycles. PD-L1 expression was measured in the membrane of tumoral cells with monoclonal rabbit anti PD-L1 28.8 pharmaDX (DAKO, Agilent Technologies) in FFPE samples at diagnosis. Primary endpoint was pathologic complete response (pCR) stratified by PD-L1 expression (positive or negative), while secondary endpoints were event-free survival (EFS) and overall survival (OS), also stratified by PD-L1 expression. Results: Both cohorts were well balanced in most of the features. Thirty-three percent of the tumors in the experimental group were PD-L1 positive, whereas 50% of them in the CG expressed PD-L1 (p = 0.06). Pathological CR was observed in 50% of the PD-L1 positive population, in contrast to a 2.8% in the PD-L1 negative in the NACT cohort (p < 0.01) as compared to the patients assigned to the DCV group (33.4% in PD-L1 positive vs 23.1% in PD-L1 negative population; p = 0.16). Among PD-L1 positive population, more pCR were seen in the CG than in the DCV group (50% vs 33.4%; p = 0.06). Within the PD-L1 negative population, more pCR were observed in the DCV group than in the CG (23.1% vs 2.8%; p < 0.05). With a median follow-up of 7 years, no significant differences were observed between the different subgroups neither in EFS (HR = 1.7; 0.42-6.8; p = 0.19) nor in OS (HR = 2.5; 0.56-11, p = 0.43). At 7 years, 20% and 14.4% of the patients relapsed according to the PD-L1 positive versus negative status respectively, and 10.78% versus 13.33% were dead. Conclusions: The benefit of DCV seems to be outstanding in the PD-L1 negative tumors that have a basal immune appropriate milieu. PD-L1 expression implies a more suppressed niche in which DCV are not able to stimulate antigen presentation and cell cytotoxic activity. PD-L1 positive population reach higher responses with both NACT±DCV than PD-L1 negative group, although the benefit seem to be higher in the NACT alone cohort. Further studies combining DVC+IPI together with NACT are needed. Clinical trial information: NCT01431196.