Presentation Authors: Daniel Artenstein*, Aaron Krug, Los Angeles, CA, Eugene Rhee, San Diego, CA, Ron Loo, Downey, CA, Stephen Williams, Riverside, CA, Polina Reyblat, Rex Parker, David Finley, Los Angeles, CA
Introduction: Prostate cancer diagnosis has undergone a sea-change with the development of powerful ancillary tests to distill out malignant from benign disease. Multiparametric MRI (MRI) is useful for the detection of larger or higher-grade tumors but can miss small lesions. Confirm MDX (MDX), an epigenetic tissue test, analyzes biopsy negative prostate tissue for tumor-associated DNA promoter hypermethylation to geographically localize occult tumor. The intersection between imaging and epigenetic testing is incompletely understood. We evaluated whether the tumor location predicted by MDX correlated with MRI tumor suspicious regions (ROI).
Methods: After IRB approval, consecutive patients referred within the past year with elevated PSA and prior negative systematic biopsy within 24 months prospectively underwent MRI at 1.5 Tesla (no endorectal coil) and MDX. MRI was centrally re-reviewed. PIRADS v.2 score was assigned ROIs. PIRADS 3 or higher lesions were considered abnormal (MRI+). Geographical concordance between MDX and MRI was defined as positivity in the same anatomic location.
Results: 113 patients underwent MDX and MRI. MDX was abnormal (MDX+) in 50.4% of patients. 28.1% of MDX+ had MRI+ [PIRADS 3, 4, 5 (56.25%, 18.75%, 25%, respectively). Geographical concordance between MDX and MRI was found in these PIRADS strata at 55.6%, 100% and 100%, respectively. Of patients with PIRADS 3 lesions and geographically discordant MDX, 75% of lesions were in the transition zone whereas all patients with concordant MDX were in the peripheral zone. To date, 9 of 16 patients in the MDX+MRI+ group underwent biopsy, 4 were found to have cancer [(2) Gleason 6, (2) Gleason 7]. The only predictive factor that distinguished MRI+ from MRI- patients in the MDX+ cohort was PSA density [(PSAD) 0.21 vs. 0.14, p=0.034.
Conclusions: MRI at 1.5T successfully located a suspicious target in the same anatomic location predicted by MDX for all PIRADS 4 and 5 lesions. For PIRADS 3 lesions, geographical concordance was more likely in the peripheral zone than the transition zone. Elevated PSAD was a predictor of abnormal MRI in patients with abnormal Confirm MDX. MDX captured some patients with a negative MRI, perhaps due to small tumor volume and limitations of MRI at 1.5T.