This is a single center, blinded (the mpMRI will be read by expert personnel not aware of the patient's diagnosis ), retrospective study. A total of 200 patients with suspicious prostate cancer (PCa) who underwent, between January 2014 and January 2022, mpMRI of the prostate and subsequent prostate biopsy will identified. All mpMRI images will be retrospectively collected and evaluated using the Quantib Prostate software. In particular, 200 mpMRI with endorectal coil will be read by the expert radiologist and by the Quantib prostate (Quantib B.V. Rotterdam, The Netherlands) software, thus evaluating the predictive characteristics of Quantib Prostate for the identification of suspicious lesions. The sensitivity, specificity, negative predictive value, positive predictive value and accuracy of Quantib Prostate vs. radiologist in detecting csPCa will be evaluated using the biopsy report as reference standard.

An increasing number of certified software using AI in radiology is becoming available. Nonetheless, the role of AI in the detection of prostate lesions at mpMRI in men with suspicious PCa is still under debate. Specifically, it is unknown whether the use of AI may help clinicians in the decision-making process with a reliable inter-observer agreement compared to assessments done by high-volume experienced dedicated uro-radiologists. In this context, the use of digital platforms available on the market such as Quantib Prostate (Quantib B.V. Rotterdam, The Netherlands) could help in the identification of patients at higher risk of clinically significant PCa (csPCa). Among the other platforms, Quantib Prostate has a validated algorithm for the evaluation of prostate MRI which has been already tested in the clinical setting and which holds a CE mark and authorization for clinical use. Moreover, the use of Quantib Prostate may lead to advantages in terms of time reduction, time-efficiency and diagnostic performance compared with the traditional radiologist approach, with a potential impact on report quality.For these reasons, among the various AI softwares which are currently available, we decided to select Quantib Prostate for these analyses.

The potential implication of this study includes:

1. Establishing inter-observer agreement between AI and traditional radiologist approach
2. Time reduction in the diagnosis of suspicious PCa
3. Error reduction within the imaging workflow

Inclusion Criteria:

1. Men at least 18 years of age referred with clinical suspicion of prostate cancer
2. Men who have undergone mpMRI and subsequent targeted plus random systematic prostate biopsies
3. Available pathological report with details regarding the site of positive cores and grade group
4. Serum PSA ≤ 20
5. Patient who underwent a 1.5T mpMRI with or without endorectal coil with the following protocol restrictions and according to recent European Guidelines, each case containing at least the following sequences, will be considered:

T2w

• No fat suppression
• At least axial acquisition
• Max voxel size 1.0x1.0x3.0mm DWI
• At least on b-value ≥800 s/mm^2 required, ≥1400 s/mm^2 preferred
• ADC map required
• Max voxel size: 2.0x2.0x4.0mm
• Axial scanning protocol only DCE scan
• Recommended temporal resolution <15s
• Axial scanning protocol only

Exclusion Criteria:

NA

A total of 200 patients who received mpMRI with endorectal coil at the Department of Radiology of the IRCCS Ospedale San Raffaele, Milan, Italy from January 2014 to January 2022 due to a clinical suspicion of PCa and subsequently underwent prostate biopsy will be retrospectively identified and will represent the study cohort.