Assessment of Genetic Alterations in Postmenopausal Women with Hormone Receptor-Positive, HER2-Negative Advanced Breast Cancer from the BOLERO-2 Trial by Next-Generation Sequencing

Abstract

Introduction: In 724 postmenopausal women with HR + , HER2– advanced breast cancer, everolimus (EVE) plus exemestane significantly improved progression-free survival, response rate, and clinical benefit rate versus placebo plus exemestane (BOLERO-2 phase 3; NCT00863655; September 2012). Although benefits were seen in all prospectively defined subgroups, some variations were observed, partially due to genetic differences in molecular determinants of EVE sensitivity and interactions between the estrogen receptor (ER) and mTOR pathways. Methods: Next-generation sequencing was used to comprehensively assess genetic alterations in archival tumor specimens. DNA was extracted and successfully analyzed from 230 tumors (>350 x coverage). Coding regions of 182 cancer-related genes were analyzed for sequence and copy number variations. Predicted germline event alterations were discarded. Results: All patients had ≥1 genetic alteration; 98% had ≥2. 173 genes were altered in ≥1 tumor sample. Alteration number/sample varied (1-24; average, 9). Sequence variations (N = 1565) were primarily point mutations (85%), deletions (9%), and insertions (6%). Frequently mutated genes were PIK3CA (49%), TP53 (24%), and ARID1A (16%). ESR1, NF1, and IGFR1 mutations were seen in 9%, 5%, and 4% of samples, respectively. Of 112 patients with PIK3CA mutations, 55 and 43 had mutations in exon 20 and exon 9, respectively. PIK3CA and AKT1 (6%) mutations were mutually exclusive. Copy-number variations included 524 amplifications (most frequent: CCND1 [32%], FGFR1 [18%]) and 27 bi-allelic deletions. About 2/3 of samples had ≥1 amplification event. However, 10% of samples had ≥8 sequence variations but no gene amplification. Common mutation rates were comparable with those from the Cancer Genome Atlas. However, higher mutation frequencies were seen in some genes in BOLERO-2, likely due to the absence of germline variation data. Conclusions: These results demonstrate the feasibility of large-scale sequencing in a phase 3 trial and generate potential hypotheses for developing targeted therapy combinations in this patient population.