Copy number aberration analysis to predict response to neoadjuvant antiHER2 therapy: results from the NeoALTTO phase III clinical trial

Purpose: The heterogeneity of response to anti-HER2 agents represents a major challenge in patients with HER2-positive breast cancer. To better understand the sensitivity and resistance to trastuzumab and lapatinib, we investigated the role of copy number aberrations (CNAs) in predicting pathological complete response (pCR) and survival outcomes in the NeoALTTO trial. Experimental design: The neoadjuvant phase III NeoALTTO trial enrolled 455 HER2-positive early-stage breast cancer patients. DNA samples from 269 patients were assessed for genome-wide copy number profiling. Recurrent CNAs were found with GISTIC2.0. Results: Copy number aberrations estimates were obtained for 184 patients included in NeoALTTO. Among those, matched transcriptome and whole-exome data were available for 154 and 181 patients, respectively. A significant association between gene copy number and pCR was demonstrated for ERBB2 amplification. Nevertheless, ERBB2 amplification ceased to be predictive once ERBB2 expression level was considered. GISTIC2.0 analysis revealed 159 recurrent CNA regions. Lower copy number levels of the 6q23-24 locus predicted absence of pCR in the whole cohort and in the ER-positive subgroup. 6q23-24 deletion was significantly more frequent in TP53 wild-type (WT) compared to TP53 mutated tumors, with copy number levels resulting significantly associated with lack of pCR only in the TP53 WT subgroup. Interestingly, a gene-ontology analysis highlighted several immune processes correlated to 6q23-24 copy number. Conclusions: Our analysis identified ERBB2 copy number as well as 6q23-24 CNAs as predictors of response to anti-HER2-based treatment. ERBB2 expression outperformed ERBB2 amplification. The complexity of the 6q23-24 region warrants further investigation.