Abstract 2617: A multicenter project to test the validity and logistics surrounding the testing of AR-V7 mRNA expression in circulating tumor cells

• Published in: Cancer research (Chicago, Ill.) Ill.), 2018-07, Vol.78 (13_Supplement), p.2617-2617
• Main Authors: Sieuwerts, Anieta M., Mostert, Bianca, Vlugt-Daane, Michelle van der, Kraan, Jaco, Beaufort, Corine, Van, Mai, Prager, Wendy J., Laere, Bram De, Beije, Nick, Hamberg, Paul, Westgeest, Hans M., Tascilar, Metin, Dirix, Luc Y., Onstenk, Wendy, Wit, Ronald de, Lolkema, Martijn P., Mathijssen, Ron H., Martens, John W., Sleijfer, Stefan
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2018-2617

Introduction: The presence of the androgen receptor splice variant 7 (AR-V7) in circulating tumor cells (CTCs) has shown to be associated with resistance to anti-AR treatment with abiraterone or enzalutamide, but not to chemotherapy containing taxanes such as cabazitaxel in patients with metastatic castration-resistant prostate cancer (mCRPC). The primary objective of this project was to set up a validated multi-center pipeline to measure AR-V7 by RT-qPCR in RNA isolated from CellSearch-enriched CTCs. This pipeline included a) the pre-analytical phase (appropriate anonymized sample collection and sample preservation), b) the analytical phase (assay performance with high analytical sensitivity, precision and specificity and c) the post-analytical phase (data evaluation and reporting). Materials and Methods: CellSearch-enirched CTCs from metastatic castration-resistant prostate cancer (mCRPC) patients were characterized by RT-qPCR. To validate the pipeline, we determined the proportion of blood samples from mCRPC patients having three or more CTCs per 7.5 mL of blood for whom an AR-V7 status can be assayed and returned to the clinician within 11 days (design PRELUDE study protocol). Results: In the range of the RNA equivalent of 0.2-12 AR-V7 positive VCaP cells the coefficient of variation (CV) for AR-V7 was 9% (n=37). The limit of detection (LOD) was 0.3 and the limit of quantification (LOQ) 3 cells in the final RT-qPCR. In none of 17 healthy blood donors an AR-V7 signal was detected, showing high diagnostic specificity (100%). No differences were observed between AR-V7 data generated by four different technicians or in two different laboratories. For the 45 patients in the PRELUDE study, thirteen patients were not eligible due to poor RNA quality (n=1) or because the blood sample did not contain enough epithelial signal (n=12). Twenty-two patients were AR-V7 negative and ten AR-V7 positive. The median, 75th and 90th percentile reporting times from blood draw to dissemination of the test results were 7, 8 and 9 days, respectively. Conclusion: This AR-V7 test with validated cut-offs, a strong intra- and inter-laboratory performance and the ability to report the outcome within a clinically acceptable time frame for the large majority of patients, met our pre-specified objectives. The AR-V7 test is now ready to be used in prospective studies (including randomized controlled trials) to test its predictive value for outcome on post-docetaxel (anti-AR of cabazita