Impact of transmurality and location on the diagnostic performance of 13N-ammonia PET in the detection of myocardial fibrosis – A hybrid PET/MR study

The influence of fibrosis transmurality and location on the diagnostic performance of positron emission tomography (PET) myocardial perfusion imaging (MPI) for detecting myocardial fibrosis remains unclear. One hundred patients with suspected or known coronary artery disease (n=50) or with a history of myocarditis (n=50) underwent simultaneous cardiac magnetic resonance (CMR) and 13N-ammonia PET imaging on a hybrid PET/MR device. CMR served as the reference with late gadolinium enhancement defining fibrosis of the left ventricular myocardium. PET MPI was assessed for fibrosis qualitatively (i.e. visually), semiquantitatively and quantitatively (both threshold-based). From CMR, for every segment, the transmurality and location (subendocardial, midmyocardial, or subepicardial) of fibrosis were recorded. The sensitivity and specificity of qualitatively analyzed PET to detect fibrosis were 83.6% and 84.9% per patient and 52.9% and 96.4% per segment, respectively, outperforming the threshold-based semiquantitative and quantitative analyses. On a per-segment basis, the sensitivity of PET was increasing with transmurality and was dependent on location with higher sensitivities for subendocardial than subepicardial or midmyocardial fibrosis (p<0.001). Overall median fibrosis extent, as depicted by PET, was significantly smaller than by CMR (2.0 [IQR 1.0-5.0] vs. 3.0 [IQR 2.0-6.0] segments, p<0.001). PET yields high sensitivity for the detection of LV myocardial fibrosis of ischemic or inflammatory etiology on a per-patient basis. However, compared to CMR, the overall fibrosis extent is significantly underestimated because, on a segmental level, sensitivities are significantly lower for fibrosis with lower transmurality and those located subepicardial and midmyocardial, as compared to subendocardial.