Trial Evaluating the Safety of 2 Schedules of Cabazitaxel in Elderly Men With mCRPC Previously Treated With a Docetaxel

Randomized, open-label, phase 3 trial in mCRPC patients aged ≥ 65 years. Number of subjects: Total:170 to 200 (85 to 100 per arm) Treatment: * Arm A : cabazitaxel 25 mg/m² on Day 1 of a 3-week cycle plus daily prednisone or * Arm B: cabazitaxel 16 mg/m² on Day 1 and Day 15 of a 4-week cycle plus daily prednisone. * Treatment will be continued for a maximum of 10 cycles unless there is documented disease progression or unacceptable toxicity. * Standard cabazitaxel premedication will be used * Prophylactic G-CSF (GRANOCYTE) will be injected from Day 3 to Day 7 after every administration cycle of cabazitaxel· All new hormonal treatment, including ODM-201, prior to study entry is allowed. * Patients who received Radium-223 are eligible for this study * Treatment with LHRH should not be discontinued. Exploratory assessments: CT-Scan (abdominal/pelvic/chest) or whole body MRI and Bone scan: at screening, every 3 months and EOT. FACT-P questionnaire:at C1D1,each subsequent visit and EOT Exploratory substudy Blood samples will be collected in France (4 or 6 sites) and the Netherlands (2 sites). Biomarker analysis will be conducted at the Urology and The Tumor Immunology Laboratory at Radboud UMC in NL. Biomarker schedule Arm A (25mg/m2): Baseline - Week 6 - Week 12 - at progression Arm B (16mg/m2): Baseline - Week 6 - Week 12 - at progression Optional sample points are at C1D8. Number of subjects: 50 Statistical analysis: A sample size of 77 to 90 evaluable patients per arm will achieve 80% power to detect a 20% difference in G3 neutropenia incidence between the 2 arms. The incidence in group cabazitaxel 25 mg/m2 q3w is assumed to be 32% and 12% on bi-weekly cabazitaxel arm. The test used is a two-sided Fisher's exact test at 0.05 significance level. Assuming 10% non-evaluable patients, 85 to 100 patients should be included in each arm for a total of 170 to 200. Patients will be stratified according to G8 score (\< 14 vs. ≥ 14), and age (\< 70 vs. ≥ 70) before randomization. Exploratory sub-study The trial is powered on a clinical endpoint, namely to detect a 20% difference in G3 nThe trial is powered on a clinical endpoint, namely to detect a 20% difference in G3 neutropenia incidence between arms (32% in arm A vs 12% arm B; power 80% with two-sided alpha of 5%, correcting for 10% non-evaluable patients (=17 patients). From the 153 to 180 evaluable patients, we have 76 to 90 patients in each arm, of which we expect 40-60 evaluable patients for translational studies (calculations performed on 25 per arm). In arm A, we expect 8 patients (32% of patients) with G3 neutropenia, and 17 patients that do not. In arm B, we expect 3 patients (12% of patients) with G3 neutropenia, and 22 patients that do not. For the MDSC analyses, we therefore will be comparing 11 patients with G3 neutropenia to 39 patients. For all continuous variables, including all immune subpopulations present in blood, mean (sd) will be presented if the distribution seems to be symmetric and in case of a skewed distribution the median and IQR. For categorical data, number and percentage will be presented. For comparison of continuous data linear regression analyses or correlation (Spearman or Pearson) will used. For comparison of continuous data with categorical data logistic regression analysis will be used. For comparison of two sets of categorical data the chi-square test of Fisher's exact test will be utilized. For the radiological PFS analyses the estimates of the hazard ratio and corresponding 95% confidence interval will be tested using a Cox Proportional hazard model. For the overall survival, a stratified log-rank test will be used to compare between groups.