Improved myocardial perfusion PET imaging using artificial neural networks.

Myocardial perfusion (MP) PET imaging plays a key role in risk assessment and stratification of patients with coronary artery disease. In this work, we proposed a patch-based artificial neural network (ANN) fusion approach that integrates information from the maximum-likelihood (ML) and the post-smoothed ML reconstruction to improve MP PET imaging. To enhance quantification and tasked-based MP defect detection, the proposed method fused features from patches of the ML and the post-smoothed ML reconstructed images with different noise levels and spatial resolution. Using the XCAT phantom, we simulated three MP PET datasets, one with normal perfusion and the other two with non-transmural and transmural regionally reduced perfusion of the left ventricular (LV) myocardium. The proposed ANN fusion technique was quantitatively evaluated in terms of noise-bias and noise-contrast tradeoff, and compared with the post-smoothed ML reconstruction. Using the channelized Hotelling observer, we evaluated the detectability of the non-transmural and transmural defects through a receiver operating characteristic analysis. The quantitative results demonstrated that the ANN enhancement method reduced bias and improved contrast while reaching comparable noise to that of the post-smoothed ML reconstruction. Moreover, the ANN fusion technique significantly improved the defect detectability of both non-transmural and transmural defects. In addition to the simulation study, we further evaluated the ANN enhancement method on patient data. Compared with the post-smoothed ML reconstruction, the ANN fusion method improved the tradeoff between noise and mean on the LV myocardium, indicating its potential clinical value in MP PET imaging.