Use of PIRADS 2.1 to predict capsular invasion in patients with radiologic T3a prostate cancer.

ABSTRACT

Objective: Using multi-parametric magnetic resonance imaging (mpMRI) to identify patients with clinical T3a (cT3a) who were overestimated on mpMRI with final pathological T2 (pT2). To suggest that the neurovascular bundle (NVB) can be preserved by evaluating the characteristics of patients according to their pathological grade among cT3a prostate cancer (PCa) patients using mpMRI. Materials and methods: Patients who underwent robot-assisted laparoscopic radical prostatectomy (RALP) were retrospectively analyzed and those patients with clinical T3aN0M0 were enrolled. These enrolled patients were divided into a localized cancer group with pT2 PCa and a locally advanced group with pT3a or higher. Factors affecting the diagnosis of localized PCa after RALP in patients with cT3a PCa were evaluated. Results: Among the preoperative parameters of patients with cT3a PCa, the prostate specific antigen density (PSAD) (OR 3.76, 95% CI 1.85-7.64, p<0.001), international society of urological pathology (ISUP) grade (p<0.05), and index lesion size (OR 1.44, 95% CI 1.85-7.64, p<0.001) were significantly associated with pathological locally advanced PCa. Optimal cut-off values for prediction of pT3a or higher were 0.36 ng/mL2 for PSAD (sensitivity 55.7%, specificity 70.8%), 1.77 cm for index lesion size (sensitivity 54.3%, specificity 66.0%), and 2.5 for ISUP grading (sensitivity 67.6%, specificity 53.2%). For prediction of pT3a or higher among patients with cT3a PCa, a nomogram was developed using ISUP grade, index lesion size, and PSAD on prostate biopsy (area under the curve 0.71, 95% CI 0.670-0.754, p<0.001). PSAD less than 0.36 index lesion size less than 1.77 cm, and biopsy ISUP grade 1-2 are highly likely to indicate that there is no actual extracapsular extension in cT3a PCa patients. Conclusions: PSAD, ISUP, and index lesion size showed significant associations with the classification of pathologic localized and locally advanced PCa in patients with cT3a PCa. A nomogram including these features can predict the diagnosis of locally advanced PCa in patients with cT3a PCa.