Value of prostate MRI in determining appropriate candidates for active surveillance.
Margolis, D; Tan, N; Natarajan, S; Chamie, K; Finley, D; Reiter, RE; Huang, J; Raman, S
Published in: Journal of Clinical Oncology
109 Background: The objective was to measure the added benefit of multi-parametric endorectal coil prostate MRI (eMRI) to traditional active (AS) criteria in delineating men with more advanced disease. Methods: We performed a retrospective study of 115 men who underwent robot-assisted laparoscopic prostatectomy (RALP) for their CaP with whole-mount pathological evaluation. All men underwent eMRI for surgical planning—including T2 weighted, diffusion-weighted (DWI), dynamic contrast-enhanced (DCE), and MR Spectroscopy—between July 2008 and March 2011. We examined the diagnostics of Epstein’s criteria (E-AS) for clinically insignificant CaP or in combination with MR parameters (eMRI-AS criteria) in predicting more advanced disease. E-AS criteria included Gleason score (GS) 3+3=6, <3 biopsy cores positive, PSA <10, ≤50% maximum % of cancer in any one core. The addition of apparent diffusion coefficient >0.85x10-3 mm2/sec, Ktrans<0.5 Hz, Kep>1.5 Hz, and normal MRSI defined the eMRI-AS cohort. Outcomes were stratified into low vs. high surgical risk. Low-risk disease was defined as having pT2, GS <3+4 with tumor size <1.5 cm, and negative surgical margins (SM). Conversely, higher-risk disease included men with GS ≥4+3, positive SM, ≥pT3 or those with GS 3+4 with tumor size ≥1.5 cm. Results: We identified 104 who met our inclusion criteria. We excluded 11 (9%) men due to post-biopsy hemorrhage (n=4), failed DCE (n=4), poor DWI (n=2), and two-year delay between MRI and RALP (n=1). Mean age was 60.6 years. Thirty-seven men (35%) satisfied E-AS, while 25 men (24%) satisfied eMRI-AS. On whole-mount sectioning 49 (47%) men were low surgical risk. E-AS to detect low surgical risk revealed a sensitivity, specificity, false positive, false negative and AUC of 67%, 73%, 33%, 27%, and 70%, respectively. eMRI-AS performance parameters were: 62.8%, 94.4%, 37%, 5% and 78%—significantly different than E-AS (p=0.04). Conclusions: eMRI-AS outperformed Epstein-AS in identifying poor candidates from going onto AS. If patients were identified as AS candidates based on traditional criteria, we found that 27% actually had higher risk disease; whereas the addition of eMRI reduced this number to 5%.
