To explore the influencing factors of clinically significant prostate cancer (CsPCa) in patients with PI-RADS score 3.

The data of 133 consecutive patients with the PI-RADS score 3 from January 2019 to December 2020 were retrospectively analyzed. All patients underwent 4-needle transperineal targeted biopsy and 12-needle systematic prostate biopsy (SB). The overall age was 66 (60-72) years, and the overall PSA value was 8.22 (5.95-11.41) ng/ml. All patients underwent multiparametric magnetic resonance imaging (mpMRI), and PI-RADS v2.0 score was 3. Patients were divided into two mutually exclusive groups: non CsPCa group and CsPCa group. The differences of lesion location, laterality, focality and sequence parameters of mpMRI between the two groups were compared, and multivariate binary logistic regression was used to analyze the independent predictors of PI-RADS score 3 in patients with CsPCa.

Biopsy results showed 57 cases of prostate cancer, including 41 cases of CsPCa, and 76 cases of non-prostate cancer. The detection rate of prostate cancer was 46.62 %(57/133), and the detection rate of CsPCa was 30.83 %(41/133). There were 41 cases in CsPCa group and 92 cases in non CsPCa group. There was no significant difference between CsPCa group and non CsPCa group in age [66 (58-70) years vs. 66 (60-72) years], body mass index [24.22 (21.82-25.71) kg/m² vs. 23.71 (21.99-26.12) kg/m²], PSA [9.39 (6.35-12.55) ng/ml vs. 7.67 (5.83-10.51) ng/ml], abnormal rate of rectal digital examination [21.95% (9/41) vs. 9.78% (9/92)] (all P > 0.05). There was significant difference in PSAD [0.40 (0.16-0.65) ng/ml² vs. 0.17 (0.12-0.24) ng/ml²] (P<0.05). In MRI, PI-RADS=3 lesions were mainly located in the transitional zone [46.62 %(62/133)]. In CsPCa group, MRI lesions were located in peripheral zone in 16 cases, transitional zone in 19 cases, and both areas in 6 cases. There were 16 cases on the right, 15 cases on the left and 10 cases on both sides. The lesions were diffused in 19 cases and localized in 22 cases. In the non CsPCa group, 41 lesions were located in the peripheral zone, 43 in the transitional zone, and 8 in both areas. There were 26 cases on the right, 35 cases on the left and 31 cases on both sides. The lesions were diffuse in 56 cases and localized in 36 cases. There was no significant difference in lesion location, side and diffusion degree between the two groups (P> 0.05). Compared with the non CsPCa group, the positive rate of all MRI sequences in CsPCa group was higher (82.93% vs. 40.22%, P < 0.001), the positive rate of T2 weighted imaging (T2WI) was higher (92.68% vs. 75.00%, P = 0.018), the positive rate of diffusion weighted imaging (DWI) was higher (90.24% vs. 56.52%, P < 0.001), the maximum diameter was larger[(0.67(0.30-1.19)mm vs. 0.48(0.20-0.62)mm, P < 0.001], and the apparent diffusion coefficient (ADC) was lower[0.70(0.61-0.87) vs. 1.10(0.86-1.50), P < 0.001]. Concurrently, PSAD and lesion ADC were important predictors of CsPCa in logistic regression model [mean 10 fold cross validation AUC: 0.78(95%CI 0.65-0.88)].

Most of the MRI lesions in patients with PI-RADS 3 were located in the transitional zone, and the MRI lesions in CsPCa were more obvious and diffusion limited. PSAD and ADC values are independent predictors for the diagnosis of CsPCa in patients with PI-RADS score 3, and the log₂PSAD-ADC prediction model is helpful to find CsPCa from patients with PI-RADS score 3 and protect patients from unnecessary biopsy.

Prostatic neoplasms; Carcinoma; Biopsy, needle; Magnetic resonance imaging; Diagnosis