Deep learning has the potential to improve automatic interpretation of MPI as compared with current clinical methods.

Information on coronary physiology and myocardial blood flow (MBF) in patients with suspected angina is increasingly important to inform treatment decisions. A number of different techniques including myocardial perfusion imaging (MPI), noninvasive estimation of MBF, and coronary flow reserve (CFR), as well as invasive methods for CFR and fractional flow reserve (FFR) are now readily available. However, despite their incorporation into contemporary guidelines, these techniques are still poorly understood and their interpretation to guide revascularization decisions is often inconsistent. In particular, these inconsistencies arise when there are discrepancies between the various techniques. The purpose of this article is therefore to review the basic principles, techniques, and clinical value of MPI, FFR, and CFR-with particular focus on interpreting their agreements and disagreements.