Chromosomal instability (CIN) in circulating tumor cells (CTC) predicts for taxane sensitivity in metastatic castration-resistant prostate cancer (mCRPC).

5026 Background: Few clinically validated biomarkers can inform therapy (tx) sequencing in mCRPC. Previously, we credentialed a biomarker of CIN in individual CTCs using phenotypic identification of cells with ≥9 large scale transitions (LSTs) and classification of CIN as high when detected in ≥3 CTC/mL (CIN3+). CIN3+ was associated with shorter radiographic progression survival (rPFS) and overall survival (OS) following taxane (T) or androgen receptor signaling inhibitor (ARSi) tx in mCRPC. Similarly, high CIN defined as ≥1 CTC/mL with ≥9 LSTs (CIN1+) predicts worse rPFS and OS following ARSi. Here, we explored if CIN1+ or CIN3+ predicted differential outcomes in pts on-tx with T or ARSi for mCRPC. Methods: We analyzed 208 banked CTC samples from 173 mCRPC pts collected within 30 days prior to starting a new tx with either T (n=88) or ARSi (n=120) in 2012-2017. CTCs were detected using Epic Sciences platform and CIN defined using the phenotypic classifier (PMID 32816908). Retrospective clinical annotation was completed for OS, defined as time from index tx start date to death, and rPFS, defined as index date to date of new or/and increasing lesion size in on-tx radiology report or death. Associations with rPFS and OS were studied according to samples’ CIN1+ and CIN3+ status. T vs. ARSi comparisons were adjusted using Cox models that incorporated pre-tx covariates, i.e. prior T or ARSi txs, total CTC count, presence of visceral disease, and prognostic lab values. Results: The median (range) follow-up from tx start to date of death or last follow-up was 7 (5-9) years. There were 53/120 (44%) pts starting ARSi that had prior ARSi txs and 41/88 (46%) pts starting taxane that had prior taxane txs. CTCs were detected in 184 (88%) samples, CIN1+ in 73 (35%; 64% pre-T, 36% pre-ARSi), and CIN3+ in 47 (22%; 60% pre-T,40% pre-ARSi). Compared to CIN1+, model results (Table) suggest that CIN3+ predicts a longer rPFS (p<0.01), and possibly OS (p=0.05), with T therapy. Compared to CIN1-, CIN3- also predicts a longer OS with ARSi therapy (p=0.04). Conclusions: CIN3+ CTC may be a more discriminating predictive biomarker than CIN1+ to help guide selection of T or ARSi tx in pts with mCRPC. Adequately powered prospective studies are planned. [Table: see text]