OBJECTIVES: To study 18F-Choline PET/CT in the diagnosis and biopsy guide of prostate cancer (pCa) in patients with persistently high prostate-specific antigen (PSA) and previous negative prostate biopsy. To compare the clinical risk factors and metabolic variables as predictors of malignancy. METHODS: Patients with persistently elevated PSA in serum (total PSA >4ng/mL) and at least a previous negative or inconclusive biopsy were consecutively referred for a whole body 18F-Choline PET/CT. Patient age, PSA level, PSA doubling time (PSAdt) and PSA velocity (PSAvel) were obtained. PET images were visually (positive or negative) and semiquantitatively (SUVmax) reviewed. 18F-Choline uptake prostate patterns were defined as focal, multifocal, homogeneous or heterogeneous. Histology on biopsy using transrectal ultrasound-guided approach was the gold standard. Sensitivity (Se), specificity (Sp) and accuracy (Ac) of PET/CT for diagnosis of pCa were evaluated using per-patient and per-prostate lobe analysis. Receiver-operating-characteristic (ROC) curve analysis was used to assess the value of SUVmax to diagnose pCa. Correlation between PET/CT and biopsy results per-prostate lobe was assessed using the Chi-square test. Univariate and multivariate logistic regression analysis were applied to compare clinical risk factors and metabolic variables as predictors of malignancy. RESULTS: Thirty-six out of 43 patients with histologic confirmation were included. In 11 (30.5%) patients, pCa was diagnosed (Gleason score from 4 to 9). The mean values of patient age, PSA level, PSAdt and PSAvel were: 65.5 years, 15.6ng/ml, 28.1 months and 8.5ng/mL per year, respectively. Thirty-three patients had a positive PET/CT; 18 had a focal pattern, 7 multifocal, 4 homogeneous and 4 heterogeneous. Se, Sp and Ac of PET/CT were of 100%, 12% and 38% in the patient based analysis, and 87%, 29% and 14% in the prostate lobe based analysis, respectively. The ROC curve analysis of SUVmax showed an AUC of 0.568 (p=0.52). On a lobe analysis, poor agreement was observed between PET/CT findings and biopsy results (p=0.097). In the univariate/multivariate analysis, none of clinical and metabolic variables were statistically significant as predictor of pCa. CONCLUSION: Choline PET/CT is a suitable procedure for the detection of pCa in highly selected patients, however, a high rate of false positive should be expected.