MR-based attenuation correction for cardiac FDG PET on a hybrid PET/MRI scanner: comparison with standard CT attenuation correction.

Abstract

The aim of this study was to evaluate the feasibility of attenuation correction (AC) for cardiac (18)F-labelled fluorodeoxyglucose (FDG) positron emission tomography (PET) using MR-based attenuation maps. We included 23 patients with no known cardiac history undergoing whole-body FDG PET/CT imaging for oncological indications on a PET/CT scanner using time-of-flight (TOF) and subsequent whole-body PET/MR imaging on an investigational hybrid PET/MRI scanner. Data sets from PET/MRI (with and without TOF) were reconstructed using MR AC and semi-quantitative segmental (20-segment model) myocardial tracer uptake (per cent of maximum) and compared to PET/CT which was reconstructed using CT AC and served as standard of reference. Excellent correlations were found for regional uptake values between PET/CT and PET/MRI with TOF (n = 460 segments in 23 patients; r = 0.913; p < 0.0001) with narrow Bland-Altman limits of agreement (-8.5 to +12.6 %). Correlation coefficients were slightly lower between PET/CT and PET/MRI without TOF (n = 460 segments in 23 patients; r = 0.851; p < 0.0001) with broader Bland-Altman limits of agreement (-12.5 to +15.0 %). PET/MRI with and without TOF showed minimal underestimation of tracer uptake (-2.08 and -1.29 %, respectively), compared to PET/CT. Relative myocardial FDG uptake obtained from MR-based attenuation corrected FDG PET is highly comparable to standard CT-based attenuation corrected FDG PET, suggesting interchangeability of both AC techniques.