A New Perspective on the Role of MRI in Refining PI-RADS 3 and PI-RADS 4 Categories

Introduction : Prostate cancer (PCa) remains one of the leading causes of cancer-related mortality. Traditional diagnostic methods, including PSA testing and transrectal ultrasound (TRUS), have limitations in detecting early-stage PCa. MRI plays a key role in diagnosis, particularly in identifying clinically significant tumors. However, the PI-RADS v2.1 system has limitations in the quantitative assessment of pathological changes. Purpose : To develop and evaluate quantitative MRI criteria for the differential diagnosis of PI-RADS 3 and PI-RADS 4 lesions, and to identify statistically significant predictors of cancer presence in lesions. Materials and Methods : A retrospective analysis was conducted on 105 patients (aged 49–78 years, median age 65) with suspected PCa from 2019 to 2024. Patients were grouped based on clinical history: negative biopsy (34 %), elevated PSA and negative MRI dynamics (33 %), and rising PSA without ultrasound changes (32 %). Prostate MRI was performed using 1.5 T and 3 T scanners (bi-or multiparametric), with assessment according to PI-RADS v2.1. PI-RADS 4 lesions were identiĮed in 73.3 % of cases, and PI-RADS 3 lesions in 26.7 %. All patients underwent stereotactic MRI-targeted biopsy with sampling from target lesions and systematic biopsy. Statistical analysis included frequency tables, chi-square and Fisher͛ s tests, t-test, Mann–Whitney test, and ROC analysis. Results : PCa was conĮrmed in 78 % of cases (predominantly Gleason 7, ISUP grade 2 adenocarcinomas). PI-RADS 3–4 lesions were most often located in the peripheral zones (73.3 %, posterior regions). New quantitative parameters were introduced: the signal intensity ratio on high b-value DWI (b 1 /b 2 ) and on ADC maps (ADC 2 /ADC 1 ). For all lesions, a b1/b2 threshold х1.56 showed sensitivity of 81.7 % and speciĮcity of 91.3 %; ADC 2 /ADC 1 >2.65 showed 52.4 % sensitivity and 82.6 % speciĮcity. For PI-RADS 4 lesions, thresholds of b1/b2 >1.57 (87.1 % sensitivity, 85.7 % speciĮcity) and ADC2/ADC1 >2.2 (81.4 % and 85.7 %, respectively) were significant predictors of cancer. For PI-RADS 3 lesions, these ratios were not significant. Dynamic contrast enhancement (performed in 34.2 % of cases) helped to reĮne lesion localization but did not influence the biopsy decision; type 3 curves correlated with cancer in PI-RADS 4 lesions (95 %). Discussion : The introduction of quantitative criteria (b 1 /b 2 and ADC 2 /ADC 1 ) allows for a more objective assessment of PI-RADS 4 lesions, reducing the dependence on radiologist experience. These parameters are signiĮcant predictors of clinically significant PCa, particularly in PI-RADS 4. Further research is needed to identify reliable quantitative markers for PI-RADS 3 lesions. Biparametric MRI (without contrast) was suĸcient for biopsy decision-making in most cases (65.7 %), making the method more accessible and safer. Quantitative analysis may assist in selecting between bi- and multiparametric MRI protocols. Conclusion : MRI with quantitative assessment of b 1 /b 2 and ADC 2 /ADC 1 improves the detection of clinically signiĮcant PCa, especially for PI-RADS 4 lesions. The b 1 /b 2 and ADC 2 /ADC 1 ratios are reliable cancer predictors for PI-RADS 4 and reduce diagnostic subjectivity. Using the proposed quantitative thresholds can support more informed decision-making regarding the need for biopsy in patients with equivocal MRI findings (PI-RADS 3 and 4).