Maximal radial distance for predicting extraprostatic extension of prostate cancer: a histopathological-radiological study.

Background There is a lack of a reliable parameter for accurate prediction of extraprostatic extension (EPE) of prostate cancer on MRI. We introduce a new parameter for predicting EPE on MRI: the maximal radial distance (maxRADD). It corresponds to the largest diameter of a prostate cancer focus (PCF) perpendicular to the contact with the prostate pseudocapsule. We compared accuracy and reliability of maxRADD with the previously proposed maximal capsular contact length (maxCCL). Materials and methods In a single-centre study, we retrospectively included all consecutive patients undergoing prostate MRI between October 2018 and December 2020, followed by radical prostatectomy. One uropathologist and four radiologists determined maxRADD and maxCCL for each PCF. Accuracy in predicting EPE was assessed using the area under the curve (AUC), with optimal sensitivity/specificity cutoffs determined by the Youden J index using the histopathological findings as the gold standard. Correlations between histopathological and MRI values of maxRADD and maxCCL were analyzed using the Pearson correlation coefficient (r). Inter-/intra-reader agreements were evaluated using intraclass correlation coefficients (ICC). Results A total of 79 men (mean age ± standard deviation, 65 ± 6 years) were evaluated. Twenty-six (32.9%) patients had prostate cancer with EPE. On histopathology, there was no significant difference in the accuracy of predicting EPE between maxRADD and maxCCL (AUC, 0.92 vs. 0.91, p = 0.28). The optimal cutoffs were ≥ 10.7 mm (sensitivity: 70%; specificity: 90%) for maxRADD and ≥ 12.8 mm (sensitivity: 90%; specificity: 70%) for maxCCL. Pearson correlation showed a strong correlation for maxRADD values determined on MRI with their histopathological counterparts (all r ≥ 0.7) for all readers. On MRI, inter-/intra-reader agreements were good for both parameters but significantly higher for maxRADD than for maxCCL (ICC, 0.87 vs. 0.80, p < 0.001 and 0.90 vs. 0.88, p < 0.001, respectively). Conclusion MaxRADD permits assessing EPE with good accuracy and shows higher reliability compared with maxCCL. Reliable preoperative prediction of EPE improves treatment planning, potentially reducing positive surgical margins and improving outcomes for patients with prostate cancer.