About sorafenib in castration-resistant prostate cancer.

In the study design of Chi et al. [1] is declared that sorafenib would be considered of interest if three or more prostatespecific antigen (PSA) responses were observed in the 25 patients accrued. In their trial, a PSA response occurred in one patient among 28 patients treated with sorafenib; therefore, sorafenib should be considered ineffective in chemonaive castration-resistant prostate cancer. The authors concluded that further study may be warranted, but needs to consider the limitations of PSA as an indicator of progression and response. In summary, Chi et al. believe that activity of this drug was missed in such a trial, also considering a recent finding about the effect of sorafenib on PSA [2]. Historically, post-therapy PSA decline has been proposed as a surrogate end point to evaluate new cytotoxic agents in metastatic castration-resistant prostate cancer [3]; PSA decline within 3 months of treatment initiation provided the highest degree of surrogacy for overall survival in a retrospective assessment of TAX327 trial [4] and similar results emerged from an analysis of data from SWOG 99-16 [5]. Criteria for PSA response were prospectively validated with taxanes/epothilones [6] and other drugs [7, 8]. At the same time, the necessity to verify the instrumental response after 4 weeks and PSA response after 12 weeks was pointed out [6]. The post-chemotherapy PSA surge syndrome suggests the opportunity of treatment continuation beyond 8 weeks despite an initial PSA elevation, with the obvious exception of a clinical disease progression [9]. Prostate Cancer Clinical Trials Working Group recommends, for noncytotoxic drugs, shifting the focus of designs from response to time to event end points, suggesting a phase II trial, possibly randomised, with TTP or TTF and not PSA response as primary end point, and that it will be also important keeping patients on trial until radiographic or symptomatic progression is documented [10]. To date, results of two other phase II trials of sorafenib in castration-resistant prostate cancer have been published [11, 12], both evidencing a limited activity of sorafenib. The first paper, additionally, reported a discrepancy between PSA response and clinical response: in this study, two patients with PSA progressive disease experienced a response on bone scan, and six of 17 with PSA progression presented a spontaneous reduction of PSA after suspension of sorafenib; letters to the editor Annals of Oncology