Introduction:
The Prostate Imaging Reporting and Data System (PI-RADS) is a structured reporting schema that helps determine the risk of clinically significant (CS) cancer on prostate multiparametric magnetic resonance imaging (mpMRI). PI-RADS 5 lesions are considered to be at the highest risk for CS cancer. However, a significant proportion of PI-RADS 5 lesions do not demonstrate CS cancer on the MRI-US fusion biopsy (FBx). In this study, we look to identify the common reasons behind the findings of benign or non-CS cancer on FBx of PI-RADS 5 lesions.
Methods:
A retrospective review of patients who underwent MR fusion biopsy from January 2014 to March 2020 was performed. Patients who had a PI-RADS 5 score on MRI were included on analysis. Demographic and pathological data were included. Clinically significant cancer was determined to be Gleason score of 7 or higher. Patients with no significant findings on biopsy were subsequently reviewed for potential causes. The causes considered included MRI reading errors, FBx registration errors, presence of inflammation, atrophy, prostatitis, atypical small acinar proliferation, prostatic intraepithelial neoplasia and non-CS cancer. Following the study, a logistic regression analysis was completed looking at age, race, digital rectal exam, total PSA, and prostate volume in patients with false positives and patients with true PI-RADS 5 lesions.
Results:
Six hundred and sixty-five FBx were performed during the study period. Out of those, 176 patients had FBx due to the presence of one or more PI-RADS 5 lesions. CS cancer on FBx was found in 135 (76.7%) patients. Out of the remaining 41 patients, 34 (82.9%) had true PI-RADS 5 lesions. In this group, 26 (76.5%) had false positives (no finding of clinically significant cancer on standard template biopsy). Causes for false positive findings included chronic inflammation (23.1%), benign biopsy (23.1%), acute inflammation/prostatitis (15.4%), and atrophy (15.4%) within the prostate. Gleason 6 was found in 46.2% of these patients. In patients with a true PI-RADS 5 lesion, 8 (23.5%) patients had FBx registration errors. Logistic regression analysis was completed, and significant predictors of false positives included younger age (p=0.01), low total PSA value (p=0.001), and high prostate volume (p=0.001).
Conclusion:
The PI-RADS scoring system is not a perfect measure of clinically significant prostate cancer. Our study showed that 76.5% of patients with a negative biopsy and confirmed PI-RADS 5 lesion had no clinically significant cancer. Additional research is required to determine further parameters to help us rectify these patients and avoid unnecessary biopsy.
Funding: N/A
Image(s) (click to enlarge):
THE EFFICACY OF PI-RADS 5 IN PREDICTING PRESENCE OF CLINICALLY SIGNIFICANT PROSTATE CANCER ON MRI-FUSION BIOPSY
Category
Prostate Cancer > Other
Description
Poster #162
-
Presented By: Farzaan Kassam
Authors:
Farzaan Kassam
Brandon Levin
Tianyuan Guan
Marepalli Rao, PhD
Juliana Tobler, MD
Sadhna Verma, MD
Abhinav Sidana, MD, MPH