Objective To compare the capability of radionuclide myocardial perfusion/metabolic imaging (hereinafter referred to as ¹⁸F-FDG PET) and cardiac magnetic resonance (CMR) in the assessment of myocardial viability among patients with ischemic heart failure.

Methods A total of 285 consecutive patients with ischemic heart failure combined with cardiac dysfunction who underwent ¹⁸F-FDG PET and CMR imaging simultaneously (within a month) at Fuwai Hospital, Chinese Academy of Medical Sciences from March 2016 to December 2019 were included in the retrospective analysis. The patients comprised 260 males and 25 females, aged (57.8±10.0) years. The 17-segment method was used to compare the agreement and differences of ¹⁸F-FDG PET and CMR in the identification of viable myocardium and infarcted myocardium in each myocardial segment. Spearman correlation coefficient and Kendall′s tau-b correlation coefficient were used to analyze the correlation between the two methods in the identification of different myocardial viabilities. The Bland-Altman method was used to evaluate the diagnostic agreement between the two methods in assessing infarcted myocardium.

Results Of the 285 patients with a total of 4845 myocardial segments, 3376 (69.7%) myocardial segments with wall motion abnormality on ¹⁸F-FDG PET were included in the analysis. The total content of infarcted myocardium measured through ¹⁸F-FDG PET and CMR was 17.6%±13.0% and 19.2%±13.4%, respectively, and a significant correlation existed between them (r=0.67, P<0.001). By contrast, ¹⁸F-FDG PET may underestimate the content of infarcted myocardium compared with CMR (The 95%CI of the total difference in infarcted myocardium content was −12.7%–27.8%, and the mean difference was 7.5%). Among the 697 segments with motion abnormality detected through CMR as transmural infarctions, most (90.4%, 630/697) showed transmural or non-transmural infarctions detected through ¹⁸F-FDG PET, showing a high agreement between the two methods. Among the 1371 segments detected as non-transmural infarctions through CMR, as many as 686 (50.0%) did not show infarct changes on ¹⁸F-FDG PET. At the individual level, no significant correlation (r=−0.09, P=0.12) exists between the total infarcted myocardium content detected through CMR and the total hibernating myocardium content detected through ¹⁸F-FDG PET in each patient. Among the 1371 segments detected as non-transmural infarctions through CMR, 838 (61.1%) had hibernating myocardium detected through ¹⁸F-FDG PET. Moreover, among the 697 segments detected as transmural infarctions through CMR, 286 (41.0%) had hibernating myocardium detected through ¹⁸F-FDG PET. Segments with non-transmural infarctions detected through CMR contained more hibernating myocardium than those with transmural infarctions (61.1% vs. 41.0%, χ²=66.207, P<0.001).

Conclusions ¹⁸F-FDG PET and CMR have their strengths in assessing myocardial viability in patients with ischemic heart failure. CMR is superior to ¹⁸F-FDG PET in terms of identifying infarcted myocardium, whereas ¹⁸F-FDG PET is more sensitive in detecting viable myocardium with the ability to distinguish hibernating myocardium. Combining ¹⁸F-FDG PET and CMR will provide more comprehensive information on myocardial viability.