This study extended a previously validated regional 7T 31P‑MRSI approach to patients with cardiovascular disease to map myocardial energetic status at the segmental level. The primary metabolic metric was the phosphocreatine‑to‑ATP ratio (PCr/ATP), measured with 31P magnetic resonance spectroscopic imaging and mapped across six mid‑ventricular AHA segments. The goal was to detect spatially heterogeneous energetic impairment and to assess local energetic–mechanical coupling, which is relevant for targeting metabolic therapies.
Twenty‑one participants underwent same‑day imaging: 9 healthy volunteers, 4 patients with type 2 diabetes, 5 with heart failure, and 3 with ischaemic cardiomyopathy. Each participant received 7T 31P‑MRSI and 3T cardiovascular magnetic resonance. PCr/ATP was assigned to the six mid‑ventricular AHA segments. From 3T CMR the investigators derived ejection fraction, global longitudinal strain, peak filling rate, and regional circumferential and radial strain to permit comparison between metabolic and mechanical measures.
Groups differed significantly in mean PCr/ATP (p = 0.005, eta2 = 0.58). Reported group means ranged from 1.78 ± 0.21 in healthy volunteers to 1.10 ± 0.28 in the heart failure group. These group differences indicate that global myocardial energetic reserve, as assessed by PCr/ATP at 7T, is lower in cardiovascular disease states represented in this cohort.
Mean PCr/ATP correlated with measures of global cardiac performance. Specifically, mean PCr/ATP correlated positively with ejection fraction (Spearman rs = 0.65, p = 0.002) and with peak filling rate (rs = 0.52, p = 0.023), and inversely with global longitudinal strain (rs = -0.63, p < 0.001). These relationships link reduced myocardial energetic reserve to impaired systolic and diastolic function in this dataset.
At the per‑segment level, statistically significant associations between segmental PCr/ATP and regional mechanical strain were confined to the septal segments (anteroseptal and inferoseptal). Circumferential strain in these septal segments correlated negatively with PCr/ATP (rs approximately -0.49 to -0.56), while radial strain correlated positively (rs approximately 0.43 to 0.57), with all reported associations meeting p < 0.05. No significant energetic–mechanical correlations were found in the non‑septal segments in this ungated acquisition, which the authors interpret as consistent with a displacement‑related measurement quality gradient affecting non‑septal regions.
Among the three participants with ischaemic cardiomyopathy, infarcted segments had markedly lower PCr/ATP than remote myocardium: 0.79 ± 0.20 in infarct segments versus 1.33 ± 0.22 in remote myocardium (p = 0.008, r = 0.62). All three ischaemic patients had remote PCr/ATP values below the healthy volunteer median; two patients had remote values below the healthy 5th percentile. These comparisons reached statistical significance by rank‑sum (p = 0.018) and Fisher exact test (p = 0.045), indicating energetic impairment extends beyond the infarcted tissue into structurally preserved remote myocardium in this small sample.
The acquisition used in this study was ungated. The authors report a displacement‑related gradient in measurement quality across myocardial segments, which likely limited detection of energetic–mechanical coupling to septal segments in this dataset. The small sample size, particularly in disease subgroups (notably only three ischaemic patients), limits generalisability and precludes definitive subgroup conclusions. Details such as the precise gating strategy alternatives, per‑segment signal‑to‑noise metrics, and wider population sampling were not reported beyond what is summarized above.
The results show that regional 7T 31P‑MRSI can detect segment‑level energetic differences across cardiovascular disease states and can identify spatially specific energetic–mechanical coupling under the conditions studied. The observation that energetic impairment in ischaemic cardiomyopathy extends into structurally preserved remote myocardium supports the potential value of regional metabolic imaging in designing and assessing metabolic cardiac therapies. The authors recommend larger studies using regional 31P‑MRSI to further investigate regional energetic–mechanical coupling as a therapeutic target.
The authors state that the 31P‑MRSI analysis code, including the adapted OXSA toolbox and AHA segment assignment pipeline, will be made publicly available at a repository upon acceptance. Anonymised participant data are available from the senior author on reasonable request, subject to institutional governance and a formal data transfer agreement.