Presentation Authors: Marshall Strother*, Lauren Schwartz, Lisa Jones, Philadelphia, PA, Colin Sperling, Glassboro, NJ, Mark Rosen, Alan Wein, Thomas Guzzo, Philadelphia, PA
Introduction: The PRECISION trial recently demonstrated the superiority of a novel algorithm over traditional systematic biopsy for the detection of clinically significant (CS) prostate cancer in biopsy-naive men. In this algorithm, all men with clinically suspected disease undergo prostate MRI followed by MRI fusion biopsy (MRF) if a lesion is detected. Systematic biopsy is omitted entirely. Concerns remain regarding the safety this algorithm, especially at centers inexperienced with MRF.
Methods: We retrospectively reviewed all biopsy-naive men who underwent prostate MRI followed by 12-core sextant biopsy following initiation of an MRF program at our institution. All men with a PIRADS 3 or greater lesion also underwent concurrent software-assisted MRF biopsy. The proportion of men who would have been diagnosed with CS (Gleason score â‰¥3+4) disease by the PRECISION algorithm was compared to systematic biopsy alone. Characteristics of men diagnosed with CS disease on systematic biopsy but not the PRECISION algorithm (PRECISION failures) were examined.
Results: 143 biopsy-naive men underwent biopsy between August 2016 and June 2018. 98 men had positive MRIs, and 84 were diagnosed with CS cancer. Concordance between systematic biopsies and the PRECISION algorithm is shown in Table 1. Systematic biopsies detected significantly more CS disease (80/143 vs. 58/143, p < 0.005). PRECISION failures with positive MRIs (n=11) had statistically smaller lesions on MRI (58 vs 162 mm2, p < 0.005). Lesion size was the most predictive variable on multivariate logistic regression but did not achieve significance (p=0.163). PRECISION failures with negative MRIs (n=14) had higher PSAs (9.1 vs 5.0, p=0.042), smaller prostates (35.1 vs 61.6, p < 0.005), and more family history (7/14 vs 3/31 p=.005), but only prostate size approached significance on multivariate analysis (p=0.061). Control chart analysis demonstrated no outliers or evidence of a learning curve.
Conclusions: The PRECISION algorithm missed 25/84 CS cancers at our institution and was inferior to systematic biopsy alone. Men with small lesions on MRI and men with small prostates and negative MRIs were especially at risk for failure. MRF-based screening algorithms should be validated at individual institutions prior to adoption.