Genomic amplifications identified by circulating tumor DNA analysis guide prognosis in metastatic castration-resistant prostate cancer

Purpose Analysis of circulating tumor DNA (ctDNA) in patients with metastatic prostate cancer (mPC) provides an opportunity to identify and monitor genomic alterations during a patient’s treatment course. We evaluated whether the presence of specific gene amplifications (GAs) and plasma copy number (PCN) alterations are associated with disease features. Methods This is a single-institution retrospective study of patients with mPC who underwent ctDNA profiling using Guardant360 ® (Guardant Health Inc.). This test identifies single nucleotide variants (SNVs) and GAs of select genes by next-generation sequencing. A total of 155 men with mPC were studied. Patients were stratified by GA status. The Kaplan-Meier method and multivariate cox regression models were used to estimate overall survival (OS) or failure-free survival (FFS) from either the date of GA detection or the initiation of systemic therapy. The chi-square test was used to evaluate associations between clinical factors and GAs. Results The presence of liver and/or lung metastases was associated with GAs of BRAF, CDK6, PI3KCA , and FGFR1 . Survival analyses were completed on a subset of 83 patients with metastatic castration-resistant prostate cancer (mCRPC). Median OS was improved in patients with 1 GA compared to patients with ≥2 GAs, whether determined from the date of initial GA(s) detection (14.9 mo vs. 8.9 mo) or date of therapy initiation nearest to GA detection (16.7 mo vs. 9.0 mo). Patients without GAs had not reached median OS. Patients with androgen receptor ( AR ) GA only were also found to have better median OS compared to patients with AR GA plus at least one other additional GA (19.3 mo vs. 8.9 mo). Patients with PIK3CA GA had significantly lower median OS compared to patients with GAs that did not have a PIK3CA GA (5.9 mo vs. 16.0 mo). In patients with AR and/or MYC GA(s), median OS improved in those with reduced AR or MYC PCN during therapy compared to those without such a reduction (25.1 mo vs. 15.9 mo). Conclusions The association of select GAs with survival provides an additional tool for assessing mCRPC prognosis and informing management. Serial monitoring of ctDNA GAs is also useful to guide prognosis and therapeutic response.