The pharmacokinetics of 18 F-fluorodeoxythymidine (FLT), 18 F-FDG, 11 C-choline, and 18 F-fluoroethylcholine (FEC) in 2 hormoneindependent (PC-3, DU145) and 2 hormone-dependent (CWR22, PAC120) prostate cancer xenograft mouse models were evaluated by PET and compared by immunohistochemistry. Further investigation was performed to determine whether PET can detect early changes in tumor metabolism after androgen ablation therapy through surgical castration. Methods: PET was performed on 4 consecutive days. In addition, the CWR22 and PAC120 tumor models were surgically castrated after the baseline measurement and imaged again after castration. The tracer uptake was analyzed using time-activity curves, percentage injected dose per volume (%ID/cm 3 ), and tumor-to-muscle ratio (T/M). Results: Regarding the hormone-independent prostate tumor models, 18 F-FLT showed the best T/M and highest %ID/cm 3 in PC-3 (2.97 6 0.63 %ID/cm 3 ) and DU145 (2.06 6 0.75 %ID/cm 3 ) tumors. 18 F-FDG seemed to be the tracer of choice for delineation of the PC-3 tumors but not for the DU145 tumors. Using 11 C-choline (PC-3: 1.33 6 0.29 %ID/cm 3 , DU145: 1.60 6 0.27 %ID/cm 3 ) and 18 F-FEC, we did not find any significant uptake in the tumors, compared with muscle tissue. Regarding the hormone-dependent prostate tumor models, the CWR22 model showed a highly significant (P , 0.01) decrease in tumor 18 F-FDG uptake from 4.11 6 1.29 % ID/cm 3 to 2.19 6 1.45 %ID/cm 3 after androgen ablation therapy. However, the 18 F-FLT, 11 C-choline, or 18 F-FEC tracers did not provide sufficient uptake or reliable information about therapy response in CWR22 tumors. The PAC120 model showed a significant increase in 18 F-FLT tumor uptake (P 5 0.015) after androgen ablation therapy. The accumulation of 18 F-FEC (before: 2.32 6 1.01 %ID/cm 3 , after: 1.36 6 0.39 %ID/cm 3 ) was found to be the next highest after 18 F-FDG (before: 2.45 6 0.93 %ID/cm 3 , after: 2.18 6 0.65 %ID/cm 3 ) in PAC120 tumors before castration and is better suited for monitoring therapy response. Conclusion: This comprehensive study in 2 hormone-dependent and 2 hormone-independent prostate tumor mouse models shows that 18 F-FLT and 18 F-FDG are the most appropriate tracers for delineation of PC-3, DU145 (except 18 F-FDG), and CWR22 tumors, but not for PAC120 tumors. 18 F-FEC and 11 Ccholine, in particular, revealed insufficient T/M ratio in the prostate tumor models. The results may indicate that radiolabeled choline and choline derivatives compete with a high concentration of the precursor dimethylaminoethanol, resulting in reduced uptake in small-rodent tumor models, a hypothesis that is currently under investigation in our laboratory.