Presentation Authors: Gregory Chesnut*, Emily Vertosick, Nicole Benfante, Daniel Sjoberg, James Eastham, Jonathan Fainberg, Vincent Laudone, Taehyoung Lee, Peter Scardino, Karim Touijer, Andrew Vickers, Behfar Ehdaie, New York, NY
Introduction: Active Surveillance (AS) protocols rely on multiple factors, including rectal exam, PSA, imaging, and biopsy to identify patients whose disease progresses. Some advocate following AS patients with MRI and only biopsying if new lesions appear. We evaluate whether an AS regimen based on MRI changes alone detects reclassification to Grade Group (GG) 2 or higher and could be used to avoid surveillance biopsies. We compare MRI-change only biopsies to our standard AS protocol.
Methods: All AS patients at our institution complete confirmatory biopsy and prostate MRI prior to enrollment. Men undergo laboratory and physical evaluation every 6 months, MRI every 18 months, and biopsy every 3 years. We identified a cohort of men initiated on AS between Jan 2013 and April 2017 and who had all values for AS per our protocol. Surveillance MRIs were reviewed, and lesions evaluated using composite Likert/PIRADS scoring. MRI findings and changes were assessed for association with upgrading. A univariate regression model was used to predict outcome of pathologic upgrading on 3-year biopsy.
Results: 201 men were initiated on AS, had complete data, did not have biopsy for cause, and were included in analysis. 61 (30%) of these men had GGâ‰¥2 at 3-year biopsy: 49 (80%) with GG2, 11 (18%) with GG3 and 1 (1.6%) with GG5. Among patients with a Likert/PIRADS score of â‰¥3 on 3-year MRI, 38% (95% CI 30%, 47%) had GGâ‰¥2 disease at 3 years, compared to 15% with Likert/PIRADS score < 3 (95% CI 8.0%, 26%, p=0.001). 57 men showed increase in Likert/PIRADS score, 24 men showed decreased score, and 120 men had stable regions of interest (ROI) on MRI. Increases in Likert/PIRADS score and findings of ECE on MRI were not associated with reclassification on univariate analysis (p=0.8 for both). Our model shows that AS biopsy only for Likert/PIRADS â‰¥ 3 lesions would avoid 353 biopsies per 1000 men at the cost of missing GG â‰¥ 2 diagnosis in 55 patients.
Conclusions: We show that presence of intermediate lesions on 3-year MRI is significantly associated with presence of GG â‰¥ 2 on surveillance biopsy. Patients with Likert/PIRADS < 3 still have up to 26% risk of harboring GG â‰¥ 2 disease. Changes in prostate imaging alone are not sufficiently reliable to guide AS biopsy and negative MRI cannot be used to avoid biopsy. We recommend an AS protocol involving PSA, MRI, and physical exam combined with repeat biopsy.
Source of Funding: NIH/NCI Cancer Center Support Grant P30 CA008748