RESULTS
The mean age, PSA, prostate volume and mean PSAd values for the 84 cases included in the study were 63.5±7.5 years, 11.68±17.34ng/ml, 62.4±38.08cm3, and 0.23±0.39ng/ml2, respectively. There were no statistically significant differences between malignant and benign diseases for age and PSA values. Prostate volume in the malignant group was found to be significantly lower and PSAd higher than the benign group (p<0.001,p<0.001)(Table-3 ).
Of the 106 lesions examined in this study from the 84 patients, 26(24.5%) were benign prostatic tissue, 36(33.96%) were prostatitis, 43(40.46%) were malignant lesions and 1(0.94%) was high-grade prostatic intraepithelial neoplasia. Among malignant lesions, 65.1% were localized in the PZ, 14% in the TZ, and 20.9% were diffuse cancers.
In the PI-RADS groups 1, 2, 3, 4, and 5, there were 5, 4, 54, 25, and 18 lesions identified respectively. No malignancy was detected in any of the 9 lesions who had PI-RADS-1 and 2. Systematic biopsy was performed to these patients with the decision of the clinician due to the increase in PSA, rectal examination findings and the age of the patient. Of the PI-RADS-3, 4, and 5 lesions, the PCa ratios were 22.2%, 56% and 94.45%, respectively.
Table-4 shows the statistical parameters in PZ and TZ when the cut-off value PI-RADS≥3 and PI-RADS≥4 taken as positive for cancer detection when T2W and DWI are independently evaluated. In TZ, there was no patient with a PI-RADS<3, so the diagnostic parameters for this variable could not be calculated.
The success of the PI-RADS score to predict cancer was found to have an AUC value of 0.764 (0.646-0.882) for PZ and 0.629 (0.347-0.910) for TZ when PI-RADS≥3 positive is considered. Evaluation by excluding the zonal anatomy and taking the PI-RADS score as ≥3 positive found the cancer prediction success had AUC values of 0.773 (0.683-0.864), 0.722 (0.621-0.824), 0.740 (0.641-0.838), 0.619 (0.514-0.724), and 0.764 (0.646-0.882) for T2W, DWI, DCE, T2W and DWI combinations (biparametric), and the T2W, DWI and DCE combination (MpMRI), respectively.
The sensitivity, specificity, NPV, and PPV values for prostate cancer detection according to PI-RADSv2.1 and regardless of the zone, for T2W and DWI independently, for biparametric and MpMRI assessment when PI-RADS≥3 and ≥4 taken as positive are summarized inTable-5 . When the results for cut-off value of ≥3 are compared with the results for cut-off value ≥4 inTable-4 andTable-5 , the differences between the two cut-off values were statistically significantly different(p<0.001). Accordingly, when the PI-RADS score cut-off value of ≥4 was taken as positive, the sensitivity and NPV moderately fall, while specificity and PPV increase.
All lesions in the ISUP>1 group (n=27) were evaluated as PI-RADS-4 or 5. While 25.47% of all lesions had ISUP>1 observed, 40.56% of the lesions (n=43) were identified to have PI-RADS-4 and 5. When PI-RADS-3 lesions are evaluated, 22.2% of these lesions received PCa. In this group, there were no lesions with ISUP>1 diagnosis. There was a positive correlation between PI-RADSv2.1 score with ISUP score and the correlation value was 0.814(p<0.001) (Table-6)
In PZ, for the ISUP grades 1, 2, 3, 4, and 5, there were 4(57.1%), 2(28.6%), 0(0%), 0(0%), and 1(14.3%) lesions which upgraded to PI-RADS group 4 with DWI score 3 and DCE positivity, identified respectively(Table-7) . For lesions with DWI score 4 and PIRADS 4, 0(%0), 2(33.3%), 2(33.3%), 2(33.3%), and 0 lesions identified respectively for the ISUP grades 1, 2, 3, 4, and 5. Lesions that are primary PI-RADS4, were observed to have higher ISUP grades. When grouped according to ISUP grade 1 and >1, there were significant differences in terms of the DWI score 3 and 4 percentages (p:0.03)(Figure-1 ,Figure-2 ).
The interobserver agreement kappa value(κ) for the PI-RADS score, without applying the cut-off value, was 0.562 which represents moderate agreement. When PI-RADS cut-off value of ≥3 positive is considered, kappa was 0.320, which represents fair agreement but when the cut-off value was determined as ≥4, the kappa is 0.770 which corresponds to a substantial agreement. Interobserver agreement for T2W was moderate with kappa:0.575 when PI-RADS≥3 positive taken as the cut-off and reached the substantial agreement with kappa: 0.814 when PI-RADS≥4 is taken. Interobserver agreement for DWI was fair with kappa:0.321, when PI-RADS≥3 is taken as the cut-off value but reached the substantial level when PI-RADS≥4 is taken as the cut-off(κ=0.757). For DCE investigation with positive and negative scores evaluation, interobserver agreement was at substantial levels with κ=0.721.