Extracapsular Extension in Prostate Cancer: Added Value of Diffusion-Weighted MRI in Patients With Equivocal Findings on T2-Weighted Imaging

OBJECTIVE. The objective of our study was to evaluate diffusion-weighted imaging (DWI) and T2-weighted imaging for predicting extracapsular extension (ECE) in patients with prostate cancer. MATERIALS AND METHODS. One hundred seventeen patients underwent preoperative 3-T MRI and radical prostatectomy. Two radiologists evaluated ECE with T2-weighted imaging based on the Prostate Imaging Reporting and Data System (PI-RADS). Then, the apparent diffusion coefficient (ADC) of the tumor calculated from two b values (0, 1000 s/ mm2) was measured. Interreader agreement of T2-weighted imaging scores was assessed with weighted kappa statistics. We compared T2-weighted imaging scores and ADC values between patients with ECE and those without ECE using the unpaired Student t test and evaluated their association with ECE using logistic regression analyses and ROC curves incorporating prostate-specific antigen value, Gleason score, clinical stage, and greatest percentage involved core length. The ADC values were also tested for differences between patients with ECE and those without ECE in subgroups stratified by the T2-weighted imaging criteria shown to have a high specificity for ECE. RESULTS. Fifty (42.7%) patients had ECE. There was substantial agreement for T2-weighted imaging scores (kappa = 0.613). T2-weighted imaging scores and ADCs were significantly different in patients with ECE and those without ECE (p < 0.001). Multivariate analysis showed that the greatest percentage involved core length, T2-weighted imaging score, and ADC (p < 0.05) were independently predictive of ECE. The AUCs for these variables (range, 0.733-0.770) were not significantly different except for the AUC for clinical stage (0.552). The use of a high specificity (92.5%) setting for ECE divided the patients into the following groups: patients with a T2-weighted imaging score of >= 4 (n = 20) and patients with a T2-weighted imaging score of < 4 (n = 97). In patients with a T2-weighted imaging score of = 4, the ADC was not significantly different between patients with ECE and those without ECE (p = 0.555). However, among patients with a T2-weighted imaging score of = 3, the ADC value was significantly lower in patients with ECE (mean +/- SD, 0.794 +/- 0.116) than in those without ECE (1.027 +/- 0.339) (p < 0.001). CONCLUSION. T2-weighted imaging scores and ADCs were independently associated with ECE, and the ADC had incremental value in patients without a high suspicion of ECE on T2-weighted imaging.