Prospective Inclusion of Apparent Diffusion Coefficients in Multiparametric Prostate MRI Structured Reports: Discrimination of Clinically Insignificant and Significant Cancers.

ABSTRACT

OBJECTIVE: The purpose of this study was to assess the diagnostic performance of prospectively assigned mean apparent diffusion coefficient (ADC) values in multiparametric MRI (mpMRI) structured reports for discriminating clinically insignificant (International Society of Urological Pathology [ISUP] group 1) from clinically significant (ISUP groups 2-5) prostate cancer. MATERIALS AND METHODS: This single-center study included men with abnormal 3-T mpMRI findings (Prostate Imaging Reporting and Data System version 2 score, ≥ 3) who subsequently underwent radical prostatectomy or targeted biopsy with positive results. One of nine radiologists prospectively reported the mean ADC for each lesion during clinical interpretation. Lesions with ADC ≤ 0.700 mm2/s × 10-3 were flagged as concerning for clinically significant prostate cancer. The index lesion at MRI correlated with the site-concordant lesion at targeted biopsy or whole-mount histopathologic analysis. Logistic regression was used to evaluate the utility of mean ADC values for discriminating ISUP grade group 1 from groups 2-5. Diagnostic performance was assessed by ROC AUC. RESULTS: Among the 218 eligible men, those with ISUP group 2-5 lesions had lower mean ADC values than those with ISUP 1 lesions; overall, 0.598 vs 0.803 mm/s2 × 10-3 (p < 0.0001); peripheral zone (PZ), 0.597 vs 0.855 mm/s2 × 10-3 (p < 0.0001); transition zone (TZ), 0.600 vs 0.660 mm/s2 × 10-3 (p = 0.035). The AUC for the PZ was 0.91 and for the TZ was 0.70. The optimal ADC cutoff values were 0.682 mm/s2 × 10-3 for PZ lesions and 0.638 mm2/s × 10-3 for TZ lesions, resulting in sensitivity and specificity of 78% and 94% for the PZ and 72% and 67% for the TZ. CONCLUSION: ADC values estimated prospectively in a clinical setting can help differentiate clinically insignificant from clinically significant prostate cancer, facilitating prebiopsy and pretreatment risk stratification.