Number of cores needed to diagnose prostate cancer during MRI targeted biopsy decreases after the learning curve

Abstract

Introduction: We hypothesized that the number of cores needed to detect prostate cancer would decrease with increasing MRI-targeted biopsy (TBx) experience. Methods: All patients undergoing TBx at our institution from May 2017 to August 2019 were enrolled in a prospectively maintained database. Five biopsy cores were obtained from each lesion ≥3 on PI-RADS v2.0 followed by a systematic 12-core biopsy. To assess learning curve, the study population was divided into quartiles by sequential biopsies. Clinically significant prostate cancer (csPC) was defined as Gleason Grade Group 2 or higher. Results: 377 patients underwent prostate biopsy (533 lesions); 233 lesions (44%) were positive for prostate cancer and 173 lesions (32%) were csPC. There was a significant decline in the number of cores required for diagnosing any cancer (P < 0.001) and csPC (P < 0.05) after the first quartile. There was no difference when stratifying by PI-RADS score or lesion volume. Within the first quartile, limiting the biopsy to 3 cores would miss 16.2% of csPC, decreasing to 6.6% after approximately 100 patients. Conclusion: MRI TBx is associated with a learning curve of approximately 100 cases. Four or 5 cores should be considered during the initial experience, but thereafter, 3 cores per lesion is sufficient to detect csPC.