Performance of a blood-based liquid biopsy test to detect PSMA expression on circulating tumor cells in men with metastatic prostate cancer.

Abstract

256 Background: PSMA-targeted radioligand therapy has dramatically improved clinical outcomes in men with mCRPC. A blood-based liquid biopsy assay to detect PSMA expression on circulating tumor cells (CTCs) could be helpful to optimize therapeutic decisions and identify patients most likely to benefit from PSMA-targeted therapies. Here we describe the development of this PSMA blood-based assay and the application of the single cell sequencing (SCS) pipeline to identify copy number alterations and further characterize heterogeneity in late-stage disease. Methods: A 4-color immunofluorescence (IF) assay was developed to detect PSMA expression on CTCs on the Epic Sciences liquid biopsy platform. CTCs were identified as DAPI+, cytokeratin (CK)+, and CD45- with appropriate cell morphology. Cells were plated onto glass slides and subjected to IF staining followed by CTC identification using a high throughput scanning platform with digital pathology machine learning algorithms. Cell line cells (LNCaP, PC3s, 22Rv1) of varying PSMA expression levels were spiked into normal blood (CLC/WBC; 1:10000) to create laboratory derived slides. Assay specificity, sensitivity, limit of detection (LOD), precision, and overall accuracy, as well as subcellular localization of biomarker expression were reported. Clinical validation was performed in a cohort of high-risk mCRPC patients (N=24) with previously demonstrated PSMA expression by tissue biopsy, utilizing 3ml of blood. Additionally, a subset of CTCs (36 cells from 11 patients) underwent single-cell low-pass whole genome sequencing to identify copy number alterations. Results: The PSMA CTC assay (Epic Sciences) achieved analytical specificity, sensitivity, and overall accuracy of over 99%. Precision was evaluated using intra-class correlation (ICC) (0.98). The assay showed an LOD of 1 CLC/technical replicate. PSMA expression was specific to staining outside the nucleus and distinguished from CK cytoplasmic staining. Analysis of 24 mCRPC samples revealed that 21/24 (88%) had at least one conventional CTC/mL (range: 0-30 CTCs/mL), of which 16/21 (76%) were PSMA positive (range: 0-27 CTCs/mL) with considerable tumor heterogeneity. 36 sequenced CTCs (N=11 pts) exhibited genomic heterogeneity and frequent AR (3/11), MYC (7/11), and MYCN (5/11) gains or losses of PTEN (4/11), TP53 (7/11), BRCA2 (5/11), CHD1 (6/11), and/or RB1(5/11 pts). Conclusions: The PSMA CTC assay analytical results firmly support its feasibility to detect PSMA expression in CTCs from mCRPC patient samples. The assay demonstrates high sensitivity, specificity, and overall accuracy. Many ongoing trials are designed to target PSMA-positive tumors, positioning this assay as a robust research tool to identify and longitudinally monitor PSMA expression and explore heterogeneity, response, and resistance in these patient populations.