Categorical versus continuous circulating tumor cell enumeration as early surrogate marker for therapy response and prognosis during docetaxel therapy in metastatic prostate cancer patients

Background: Circulating tumor cell (CTCs) counts might serve as early surrogate marker for treatment efficacy in metastatic castration-resistant prostate cancer (mCRPC) patients. We prospectively assessed categorical and continuous CTC-counts for their utility in early prediction of radiographic response, progression-free (PFS) and overall survival (OS) in mCRPC patients treated with docetaxel. Methods: CTC-counts were assessed in 122 serial samples, as continuous or categorical (<5 vs. ≥5 CTCs) variables, at baseline (q0) and after 1 (q1), 4 (q4) and 10 (q10) cycles of docetaxel (3-weekly, 75 mg/m2) in 33 mCRPC patients. Treatment response (TR) was defined as non-progressive (non-PD) and progressive disease (PD), by morphologic RECIST or clinical criteria at q4 and q10. Binary logistic and Cox proportional hazards regression analyses were used as statistical methods. Results: Categorical CTC-count status predicted PD at q4 already after one cycle (q1) and after 4 cycles (q4) of chemotherapy with an odds ratio (OR) of 14.9 (p = 0.02) and 18.0 (p = 0.01). Continuous CTC-values predicted PD only at q4 (OR 1.04, p = 0.048). Regarding PFS, categorical CTC-counts at q1 were independent prognostic markers with a hazard ratio (HR) of 3.85 (95 % CI 1.1-13.8, p = 0.04) whereas early continuous CTC-values at q1 failed significance (HR 1.02, 95 % CI 0.99-1.05, p = 0.14). For OS early categorical and continuous CTC-counts were independent prognostic markers at q1 with a HR of 3.0 (95 % CI 1.6-15.7, p = 0.007) and 1.02 (95 % CI 1.0-1.040, p = 0.04). Conclusions: Categorical CTC-count status is an early independent predictor for TR, PFS and OS only 3 weeks following treatment initiation with docetaxel whereas continuous CTC-counts were an inconsistent surrogate marker in mCRPC patients. For clinical practice, categorical CTC-counts may provide complementary information towards individualized treatment strategies with early prediction of treatment efficacy and optimized sequential treatment.