Indentification of a CTC-based gene expression signature predicting resistance to abiraterone and enzalutamide in mCRPC.

Abstract

5072 Background: Circulating tumor cell (CTC)-based detection of AR-V7 has been shown to be one potential marker for predicting response to 2nd generation androgen receptor (AR) therapies. However, the apparent rarity of AR-V7 positivity is indicative of the importance of other drivers of resistance in this setting. We sought to utilize a multiplex gene expression platform for assessing CTCs in order to determine other predictive biomarkers of response. Methods: Whole blood (~5mL) was obtained from 37 patients with mCRPC starting enzalutamide (n=16) or abiraterone (n=21). CTCs were isolated using anti-EpCAM-conjugated magnetic beads. Following cell lysis, mRNA was extracted followed by multiplex qRT-PCR for 92 prostate cancer-related genes. Samples were considered CTC-positive based on a previously established set of epithelial markers. We identified genes associated with PSA and radioclinical progression free survival (PFS) using Cox regression analysis. Muli-gene models were tested using ROC analysis. Results: We identified 20 patients (54%) with detectable CTCs, and patients were followed for a median of 10 months (IQR 3.9-19.4 months). Seven genes were associated with both PSA PFS and radioclinical PFS in the Cox analyses (Table). Combing the 7 genes into a single model gave AUC values of 0.88 for PSA PFS and 0.89 for radioclinical PFS. In comparison, the AR-V7-only model resulted in AUC values of 0.65 and 0.66. Conclusions: We identified seven prostate cancer-related genes that can be determined from CTCs and appear to predict short time to progression in men with mCRPC being treated with 2nd generation hormonal therapies. While this is a small cohort and prospective validation is needed, these findings highlight the potential role for this approach in helping guide therapy choice. [Table: see text]