Evaluation of the effectiveness of cognitive biopsy (CB) in patients with clinical suspicion of prostate cancer (PC), and at least one negative biopsy (TRB). Retrospective study of 144 patients with at least one previous TRB and magnetic resonance imaging (MRI). The MRI nodules were classified based on PI-RADS v2 grouping pZa, pZpl and pZpm as the peripheral zone(PZ), Tza, Tzp and CZ as the transitional zone (TZ), and the AS zones as the anterior zone (AZ). A biopsy was indicated for nodules ≥PI-RADS 3. Uni and multivariate analysis was undertaken (logistic regression) to identify variables relating to a PI-RADS 3 tumour on biopsy. The median age was 67 (IQR: 62-72) years, the median PSA was 8.2 (IQR: 6.2-12) ng/ml. A nodule was identified on MRI in the PZ in 97 (67.4%) cases, in the TZ in 29 (20.1%), and in the AZ in 41 (28.5%). PC was diagnosed on biopsy in 64 (44%) patients. The cancer rate in the PI-RADS 3 lesions was 17.5% (7/40), in the PI-RADS 4 47.3% (35/73), and in the PI-RADS 5 lesions it was 73.3% (22/29) (p=.0001). Multivariable analysis with variables that could influence the biopsy result in patients with PI-RADS 3: None (age, PSA, number of previous biopsies, rectal examination, PSAD, prostate volume or number of extracted cylinders) behaved as an independent tumour predictor. The diagnostic performance of CB in patients with at least one previous negative biopsy was 44%, increasing according to the PI-RADS grade, and low in PI-RADS 3. No clinical variable predictive of cancer was found in patients with PI-RADS 3. Copyright © 2018 AEU. Publicado por Elsevier España, S.L.U. All rights reserved.