Implementing molecular characterisation of prostate cancer tissue from patients recruited to the multi-centre STAMPEDE trial: The STRATOSPHERE consortium

Abstract

Background: STAMPEDE is a multi‐arm, multi‐stage (MAMS) platform protocol. It has recruited more than 10,000 patients since 2005. Hormone‐sensitive metastatic or locally advanced prostate cancer patients starting androgen deprivation therapy (ADT) have been randomised between a number of experimental treatment pairings, including docetaxel or abiraterone, shown to improve overall survival. Funded by PCUK, the STRATOSPHERE consortium (Stratification for rational treatment‐oncomarker pairings of STAMPEDE patients starting long‐term hormone treatment) has established a framework to enable molecular analyses of archival tissue from these patients. We present the optimised work‐flow and preliminary data on feasibility of next‐generation sequencing (NGS) of FFPE archival samples. Methods: Approximately 95% of patients participating in STAMPEDE have consented to donate their archival tissue for research. Retrospective tumour block acquisition started in March 2017 and is ongoing. To date, FFPE blocks from 1310 (35%) consenting Abstract: ADT and abiraterone and docetaxel +/‐ zoledronic acid comparison arms (3909 total participants) have been retrieved from 64/112 STAMPEDE sites. Results: Morphological review has identified cancer in 86% of cores. DNA and RNA from alternate sections have been extracted using the Zymo Quick‐DNA and Quick‐ RNA FFPE kits. The mean amount of DNA obtained per core was 23ng (range: 0‐82ng) from a mean of 8 sections (range 2‐14) at 10um thickness per block. DNA has been subjected to low‐pass whole genome sequencing and analyses of the first 100 cores from participants recruited to the SOC arm is ongoing. RNA is being extracted from cores with >50% cellularity and>60% tumour content to assess transcriptomic changes and integrate with genomic data. Conclusions: The STRATOSPHERE consortium is addressing the challenge of exploiting FFPE material with low‐output, poor‐quality nucleic acids from a clinicallyimportant randomised cohort with long‐term follow‐up. Molecular data will be linked to clinical outcome to test differences in treatment effect by molecular sub‐type.