Abstract P4-12-03: Role of primary tumor molecular characteristics in identifying relapses in HER2-positive breast carcinomas adjuvantly treated with trastuzumab

Abstract

Abstract Trastuzumab, a recombinant humanized monoclonal antibody directed to the HER2 protein, has shown survival benefits in women with HER2-positive breast cancer, and treatment is now FDA-approved in combination with chemotherapy. Nonetheless, some patients relapse after treatment, underscoring the need to identify patients for whom chemotherapy + trastuzumab is adequate versus patients requiring additional drugs. To search for factors predictive of relapse in HER2-positive breast carcinoma patients treated adjuvantly with trastuzumab, we conducted gene expression profiling analysis in 53 cases selected among a cohort of 243 patients (32 of whom relapsed) treated in the clinic with chemotherapy + trastuzumab in our institute during 2006-2009, with median follow-up of 32 months; the 53 cases comprised 23 of the relapsed and 30 of the non-relapsed patients with similar clinico-pathological characteristics (size, lymph node involvement and estrogen receptor positivity) for whom sufficient material was available. RNA extracted from formalin-fixed, paraffin-embedded (FFPE) was profiled using HumanHT12_v4 wgDASL expression BeadChips (Illumina). A Cox's proportional hazard model was used to estimate the association between gene expression and relapse-free survival (RFS), and a multivariate permutation test was used to check the proportion of false discoveries. Analysis identified 330 probes corresponding to 308 unique genes as significantly associated to RFS (a<0.005; permutation test p<0.01), showing an enrichment for gene ontology biological processes such as cell cycle regulation of immune cell proliferation and activation, cytoskeleton organization, ER pathway and ion transport. We also tested the feasibility of developing a predictive model including the gene expression data. Based on 10-fold cross-validation, our model was able to stratify patients into two groups (high and low risk), with a 4-year RFS probability of 14.5% in high-risk versus 87% in low-risk tumors (HR = 8.33, 95% CI = 3.53-18.18, p<0.0001). To test whether this signature identifies patients with intrinsic poor prognosis independent of trastuzumab treatment, we conducted in silico analysis on a dataset containing 144 HER2-positive breast carcinomas treated with chemotherapy alone; results revealed an inverse association between risk groups and prognosis, with an overall survival probability of 83% in our high-risk group and 63% in the low-risk group (HR = 0.60, 95% CI = 0.37-0.94, p = 0.0249). Together, our results suggest the usefulness of molecular characteristics of primary tumors as indicators of early relapse in patients treated adjuvantly with trastuzumab and as tools to identify patients for whom additional treatments are required. Supported by AIRC. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-12-03.