Presentation Authors: Michelle Yu*, Avinash Maganty, Liam C. Macleod, Pittsburgh, PA, Mina M. Fam, Neptune, NJ, Jonathan G. Yabes, Jathin Bandari, Alessandro Furlan, Pittsburgh, PA, Christopher P. Filson, Atlanta, GA, Benjamin J. Davies, Bruce L. Jacobs, Pittsburgh, PA
Introduction: Multi-parametric resonance imaging (mpMRI) has emerged to improve disease risk-stratification and decrease number of repeat biopsies in men on prostate cancer active surveillance (AS). However, the impact of mpMRI on cost of AS has not been established. Thus, we characterized the impact of mpMRI on cost of AS in the Medicare population.
Methods: Using SEER-Medicare data we identified men â‰¥66 years with localized grade group I-II prostate cancer diagnosed from 2008-2013. With an established algorithm, we classified men into active surveillance with and without mpMRI. We then determined costs of surveillance in each group using inflation-adjusted Medicare payments for surveillance-related procedures and their sequelae (i.e. PSA tests, prostate biopsies, post-biopsy complications and mpMRIs). Multivariable median regression compared cost and procedural-intensity for men who received mpMRI and those who did not.
Results: We identified 9,081 men on AS with median follow up of 45 months (IQR 29-64 months). 7,856 (87%) men did not receive a mpMRI and 1,225 (13%) did. On multivariable median regression, receipt of mpMRI was associated with an additional $449 (95%CI $391-$507) in Medicare payments per year. Younger age, treatment in the west or northeast, greater population density, and treatment later in the study period were associated with increased cost of AS.
Conclusions: Among Medicare beneficiaries on AS, mpMRI is associated with additional annual cost to Medicare. MpMRI may be a marker of more stringent AS, which is likely more costly than watchful waiting. Future studies are needed to determine optimal use of mpMRI during AS to maximize value.
Source of Funding: Bruce L. Jacobs is supported in part by the American Urological Association Data Grant and the University of Pittsburgh Physicians Foundation. The views expressed in this abstract do not reflect the views of the Federal government.