To determine a prognostic model derived from prostate cancer-enhanced transcripts in whole blood of castration-resistant prostate cancer (CRPC) patients and explore its applicability as a surrogate of treatment response.METHODS.
Six out of twenty-three selected transcripts were identified as specific for detection of metastatic prostate cancer cells in peripheral blood using quantitative polymerase chain reaction (qPCR). Their prognostic value was explored in whole blood samples of a training cohort (n = 22 CRPC patients, New York, USA). A resulting 2-gene panel (2GP) including KLK2 and TMPRSS2 was validated in an independent cohort with pre- and post-treatment blood draws after 9–16 weeks of systemic treament (n = 86 CRPC patients, Munich, Germany). Overall survival (OS), prostate-specific antigen progression-free survival (PSA-PFS), and clinical PFS were analyzed. Kaplan–Meier and cox regression analyses were performed.RESULTS.
An unfavorable 2GP (≥1 marker positive) identified patients with poor survival (median OS 10.0 months [95%CI 5.7–14.2] vs. not reached; P = 0.023). This was validated in an independent cohort at pre-treatment (median OS 7.8 [95%CI 6.5–9.2] vs. 17.3 months [95%CI 10.7–23.8]; P = 0.004) and post-treatment blood draw (median OS 5.0 [95%CI 0.0–10.0] vs. 18.0 months [95%CI 9.5–26.6]; P = 0.003). The 2GP independently predicted OS on multivariate analysis (hazard ratio 2.1 [95%CI 1.1–4.0]; P = 0.034) and performed better than PSA decline at correlation with OS. Conversion to favorable 2GP during treatment correlated with improved OS (7.8 to 20.9 months), PSA-PFS (2.8 to 12.0 months), and clinical PFS (4.6 to 8.0 months).CONCLUSIONS.
The established 2GP is prognostic for survival at pre- and post-treatment blood draw in CRPC patients and conversion to favorable 2GP predicts treatment benefit. Prostate 76:1160–1168, 2016. © 2016 Wiley Periodicals, Inc.