Prostate cancer: Utility of diffusion‐weighted imaging as a marker of side‐specific risk of extracapsular extension

Abstract

To assess the utility of diffusion-weighted imaging (DWI) findings as an indirect marker of side-specific risk of extracapsular extension (ECE) of prostate cancer. Fifty-one patients underwent 3T magnetic resonance imaging (MRI) before prostatectomy. Radiologists 1 and 2 (4 and 1 years experience) assessed each side for ECE using T2-weighted imaging (T2WI) and evaluated apparent diffusion coefficient (ADC) maps for the presence of apparent tumor in each lobe and to measure peripheral zone ADC. A uropathologist measured the extent of any ECE. In all, 28/102 lobes had ECE, of which 12 measured ≤1 mm, 11 measured >1 mm and ≤2 mm, and five measured >2 mm. Side-specific accuracies for detection of ECE for readers 1 and 2 were respectively: T2WI 68.6% and 74.5%; presence of apparent tumor on ADC map 66.7% and 60.8%; ADC value 75.5% and 69.6%. For ECE >2 mm, both readers achieved 100% sensitivity based on apparent tumor on ADC map or ADC values and 80% sensitivity using T2WI. For detection of ECE ≤2 mm, sensitivity for all combinations of the three methods and two readers ranged from 58.3%-81.8%, aside from assessment for ECE using T2WI by the less experienced reader, which exhibited sensitivity of 17.4%. Interreader agreement for the presence of ECE was 0.18 using T2WI, 0.37 using apparent tumor on ADC map, and 0.60 using ADC values. Compared with T2WI, DWI had comparable accuracy for side-specific assessment of ECE, greater sensitivity for ECE <2 mm for the less-experienced radiologist, and greater interreader agreement.