Presentation Authors: Matthew Gay*, Joshua Langston, Douglas Kelly, Norfolk, VA
Introduction: Multiparametric MRI (mpMRI) transrectal ultrasound (TRUS) fusion-guided biopsy (Fbx) has been shown to be superior to standard 12-core TRUS-guided biopsy (Sbx) in detecting clinically significant (CS) prostate cancer (PCa). The effect of provider variability on prostate cancer detection using mpMRI Fbx has yet to be established. We hypothesize that there is no statistically significant difference between the yields of PCa detection using Fbx in the region of interest (ROI) on mpMRI between providers.
Methods: A retrospective chart review was conducted on the first 250 mpMRI-TRUS Fbx performed at our institution. All patients underwent mpMRI at 1 of 2 imaging centers. Suspicious lesions on mpMRI were scored according to PI-RADSv2. Three providers performed Sbx and Fbx according to standardized clinical pathways using UroNav Fusion Biopsy System (Phillips, Invivo). Patients were assigned to each provider randomly, based on the chronological order in which the consults were received. CS PCa was defined as a tumor containing Gleason score (GS) 3+4 or higher PCa. Fisher's exact tests were performed to assess if provider variability was significant.
Results: The overall (Fbx and Sbx) cancer detection rate (CDR) was 56% (n = 250). The CDR for Fbx was 45% and the CS CDR was 37%. The CDR for Sbx was 38% and the CS CDR was 28%. Fbx missed 21% PCa; however, only 11% CS PCa. Sbx missed 32% PCa and 25% CS PCa. In Fbx, PCa detection rates for PI-RADS 3/4/5 lesions were 17%, 53%, and 77%, respectively. In Fbx, CS PCa detection rates for PI-RADS 3/4/5 lesions were 11%, 42%, and 69%, respectively (Table 1). In Fbx, PCa detection rates for PI-RADS 3 lesions were 7%, 9%, and 30% (p = 0.036) for provider 1, 2, and 3, respectively. In Fbx, PCa detection rates for PI-RADS 4 lesions were 52%, 54%, and 53% (p = 0.99) for provider 1, 2, and 3, respectively. In Fbx, PCa detection rates for PI-RADS 5 lesions were 80%, 80%, and 72% (p=0.83) for provider 1, 2, and 3, respectively (Table 2).
Conclusions: In PI-RADS 4 and 5 lesions, provider variability did not affect detection of prostate cancer, suggesting that Fbx yield in PI-RADS 4 and 5 lesions is provider independent. There was variability in prostate cancer detection with Fbx in PI-RADS 3 lesions. Further research is needed to determine the etiology of this variation and methods to increase concordance between providers.