OBJECTIVE To determine if sorafenib is associated with a 4-month probability of progression-free survival, which is consistent with 50%, as determined by clinical, radiographic, and prostate-specific antigen (PSA) criteria in patients with metastatic androgen-independent prostate cancer (AIPC). METHODS Patients with progressive metastatic AIPC were enrolled in an open-label, single-arm phase II study. Sorafenib was given continuously at a dose of 400 mg orally twice daily in 28-day cycles. Clinical assessment and PSA measurement were done every cycle whereas radiographic measurements were carried out every two cycles. RESULTS Twenty-two patients were enrolled in the study to date, completing a planned first stage of the trial. Baseline patient characteristics included a median age of 63.9 years (range, 50-77 years), Gleason score of 9 (range, 4-9.5), and PSA concentration of 53.3 ng/mL (range, 2-1,905 ng/mL). Fifty-nine percent of patients had received one prior chemotherapy regimen. Of the 21 patients with progressive disease, 13 progressed only by PSA criteria in the absence of evidence of clinical and radiographic progression. Two patients were found to have dramatic reduction of bone metastatic lesions as shown by bone scan, although they met PSA progression criteria at the time when scans were obtained. Toxicities likely related to treatment included one grade 3 hypertension; one grade 3 hand-foot syndrome; and grade 1/2 toxicities: fatigue, anorexia, hypertension, skin rash, nausea, and diarrhea. Results from in vitro studies suggested that PSA is not a good marker of sorafenib activity. The geometric mean exposure (AUC(0-12)) and maximum concentration (C(max)) were 9.76 h mg/L and 1.28 mg/L, respectively. The time to maximum concentration (t(max)) and accumulation ratio (after second dose) ranged from 2 to 12 h and 0.68 to 6.43, respectively. CONCLUSIONS Sorafenib is relatively well tolerated in AIPC with two patients showing evidence of improved bony metastatic lesions. Interpretation of this study is complicated by discordant radiographic and PSA responses. PSA may not be an adequate biomarker for monitoring sorafenib activity. Based on these observations, further investigation using only clinical and radiographic end points as progression criteria is warranted. Accrual to the second stage of trial is ongoing.