Breathing-enhanced oxygenation-sensitive cardiac magnetic resonance imaging to identify coronary microvascular and vasomotor dysfunction in patients with ischemia and no obstructive CAD

Background Endothelial-dependent coronary microvascular dysfunction (CMVD), which requires invasive coronary function testing (ICFT) with intracoronary acetylcholine for diagnosis, is the predominant cause of ischemia in patients with no obstructive coronary artery disease (INOCA). CMVD however is rarely used for clinical decision-making, because currently available diagnostic procedures for assessing CMVD are either invasive or costly. Oxygenation-sensitive cardiac magnetic resonance (OS-CMR) has been introduced as a non-invasive, needle-free diagnostic test for endothelium-dependent coronary vascular function. Purpose This study aimed to validate OS-CMR combined with vasoactive breathing maneuvers against gold-standard ICFT in the assessment of endothelial-dependent CMVD in INOCA patients. Methods We conducted a prospective, single-center study that enrolled 27 INOCA patients diagnosed with endothelial-dependent CMVD by ICFT and 34 healthy participants. All participants underwent OS-CMR with a vasoactive breathing maneuver protocol-paced hyperventilation (inducing vasoconstriction) followed by an end-expiratory breath-hold (inducing vasodilation). OS-CMR derived biomarkers including the hyperventilation-induced myocardial oxygenation response (HV-MORE) and the breathing-induced myocardial oxygenation response (B-MORE) mean signal intensity (MI) were analyzed across the entire cardiac cycle and within the end-systolic and end-diastolic phases. Logistic regression with absolute shrinkage and selection operator (LASSO) penalization identified key predictive biomarkers of CMVD in INOCA patients. Results Twenty-seven patients and 34 healthy controls were included in the final analysis. Compared to healthy participants, INOCA-CMVD patients exhibited significantly impaired MORE across the cardiac cycle. End-diastolic myocardial oxygenation variability (end-diastolic HV-MOREMI-VAR and end-systolic B-MOREMI-AUC) emerged as robust predictors of endothelial-dependent CMVD. The LASSO-derived multivariate model, adjusting for age, demonstrated excellent discrimination (AUC 0.92), sensitivity (88.9%), specificity (82.1%), positive predictive value (82.8%), negative predictive value (88.5%), and accuracy (85.5%) in identifying endothelial-dependent CMVD in INOCA patients. Conclusion OS-CMR with vasoactive breathing maneuvers accurately identifies endothelial-dependent CMVD in INOCA patients with excellent diagnostic performance compared to ICFT.