Abstract 5384: Analytic validation of an assay to detect androgen receptor splice variant ARv7 protein expression on circulating tumor cells from prostate cancer patients

• Published in: Cancer research (Chicago, Ill.) Ill.), 2020-08, Vol.80 (16_Supplement), p.5384-5384
• Main Authors: Campton, Daniel, Lo, Edward, Costandy, Lillian, Gardner, Brady, Houston, Ryan, Werbin, Jeffery L., Teplitz, Kyla, Sabath, Daniel E., Clein, Alisa C., Higano, Celestia S., Mojica, Tanisha M., Province, Kristin, Kaldjian, Eric P., Ramirez, Arturo B., George, Tad
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2020-5384

Circulating tumor cells (CTCs) can provide information on drug target expression, response to therapy, and disease prognosis from a non-invasive blood draw. Currently, investigating biomarkers on CTCs is difficult due to challenges of developing multiplexed assays that also identify rare cells. Presence of the androgen receptor splice variant ARv7 in prostate cancer cells is associated with resistance to second generation anti-androgen therapies. We report here the analytical validation of an immunofluorescence assay for characterization of ARv7 protein expression on CTCs using the RareCyte platform - an end-to-end platform that combines CTC sample preparation, multiparameter fluorescence staining, digital imaging, and single cell retrieval. Blood samples spiked with positive and negative cell lines for ARv7 expression were processed using the AccuCyte Sample Preparation System. Slides were auto-stained by immunofluorescence with the RarePlex ARv7 CTC Panel Kit comprised of a three-channel CTC detection base plus an ARv7 biomarker channel. The detection base consists of a nuclear dye, anti-CD45 antibody to exclude white blood cells, and cocktailed antibodies to cytokeratin (CK) and epithelial cell adhesion molecule (EpCAM). Stained slides were imaged with the CyteFinder Instrument. CTCs were identified using machine learning-based algorithms and confirmed by user review. Mean fluorescence intensity (MFI) measurements were used as a metric for ARv7 expression on confirmed CTCs. Analytic validation studies of the AR-V7 CTC assay were performed using 22RV1 (high), LNCaP (low), and BT-474 (negative) cell lines. Performance characteristics tested for ARv7 included accuracy, sensitivity, specificity, repeatability, and inter-stainer run coefficient of variation. Performance metrics for CTC recovery were calculated on spike-in and clinical samples. For recovery calculations, the number of CTCs found with the ARv7 assay was compared to the number of CTCs found with the CTC detection base assay. An ARv7 MFI threshold that segregated negative and positive cell lines was statistically defined. This threshold identified 80% of 22RV1 cells as positive for ARv7, 97% of BT-474 cells as negative, with an overall accuracy of 90%. When the assay was applied to clinical prostate cancer samples, staining with proper nuclear localization was observed. CTC recovery was at least as high with the ARv7 assay as with the base CTC detection assay. Citation Format: Daniel Campton, Ed