PROJECT SUMMARY It was long taken for granted that obstructive coronary artery disease (CAD) is the primary driver of angina and major adverse cardiac events. However, recent landmark studies have shown that up to 50% of the patients referred for diagnostic testing have ischemia with no obstructive CAD (INOCA). A large proportion of INOCA patients have coronary microvascular dysfunction (CMD), which even in the absence of flow-limiting stenoses can lead to myocardial ischemia and carries a high risk of adverse events. The reference standard for assessment of CMD is the functional coronary reactivity (CR) test, which is invasive. Despite key studies showing value of stratifying therapy based on CR testing, the practical utility of CR testing in the INOCA population is limited by its invasive nature, which carries serious risks even at experienced centers. Hence, a noninvasive approach that can detect and stage the severity of CMD would be invaluable for managing INOCA patients. Driven by this unmet need, prior studies have employed imaging approaches to index myocardial perfusion reserve (MPR) against CR; however, the association shown to date between MPR and CR impairment has been weak, likely due to the suboptimal sensitivity of MPR to subendocardial myocardial blood flow (MBF) deficits which is a hallmark of CMD. Studies using invasive microsphere-based methods have established a stress subendocardial-to-subepicardial (endo-epi) MBF gradient of larger than 1.0 in healthy animals, and shown that it decreases well below 1 under abnormally elevated microvascular resistance. However, noninvasive detection of endo-epi MBF gradients using existing imaging strategies is challenging because of the need to resolve MBF transmurally. We have developed new MRI strategies aimed at overcoming key barriers for accurate evaluation of endo-epi MBF gradients and applied them in preliminary animal and patient studies. Based on our preliminary data, we hypothesize that in the setting of CMD, impaired microvascular CR manifests as a stress-induced endo- epi MBF gradient, and the magnitude of this gradient significantly correlates with CMD severity. To test this hypothesis, we propose 3 specific aims. In Aim 1, we will develop a free-breathing artifact-free MRI technique optimized for high-resolution imaging of endo-epi MBF gradients, combined with a machine learning approach for fully-automated objective quantification of MBF gradients. In Aim 2, we will test the hypothesis that CMD severity can be staged on the basis of MRI-derived stress MBF gradient in a pig model of CMD. In Aim 3, we will test the hypothesis that CMD severity in INOCA patients is highly correlated with MRI-derived stress MBF gradient. This project brings together multiple interdisciplinary investigators with a strong collective track record in developing cardiac imaging strategies to advance a noninvasive approach for determining CMD severity based on the MRI-derived stress MBF gradient. Hence ...