[Feasibility study of intelligent structural-functional analysis of cardiomyopathy based on four-dimensional modeling...

Objective: To construct and preliminarily validate the four-dimensional cardiac mesh model based on cardiac magnetic resonance (CMR) imaging for the evaluation of cardiac structure and function in patients with cardiomyopathy. Methods: A total of 70 patients who underwent CMR examinations were enrolled from the UK Biobank database as the training set. A four-dimensional mesh model of the cardiac left and right ventricles was constructed using CMR images that passed quality control. Additionally, a validation set consisting of 30 patients who underwent CMR examinations at the First Medical Center of the Chinese People's Liberation Army General Hospital from March 2021 to October 2024 were included. These patients were classified into three groups based on clinical diagnosis: dilated cardiomyopathy (DCM) group, hypertrophic cardiomyopathy (HCM) group, and control group (no cardiomyopathy). The trained model was applied to analyze the CMR images of the validation set patients, calculating various cardiac structure and function parameters, including left ventricular volume index, left ventricular ejection fraction, left myocardial mass index, and strain. Differences in these parameters between the DCM or HCM groups and the control group were compared, and linear mixed models were used for statistical analysis of inter-group differences. Results: The age of patients in the training set was (62.2±10.3) years, with 58 (83%) males. The age of patients in the validation set was (56.5±13.9) years, with 21 (70%) males, including 8 patients in the HCM group, 9 in the DCM group, and 13 in the control group. The left ventricular ejection fraction in the DCM group was lower than that in the control group ((39.34±3.73) % vs. (53.56±2.98) %, P=0.006 3). The linear mixed model analysis showed that the left ventricular volume index and various strain parameters in the DCM group were significantly higher than those in the control group (all P<0.001). The left myocardial mass index in the HCM group was higher than that in the control group ((94.21±9.10) g/m² vs. (56.36±11.33) g/m², P<0.001). The analysis results of the linear mixed model were consistent with it (P<0.001). Conclusions: The preliminary results confirm that the four-dimensional cardiac mesh model based on CMR imaging constructed in this study could integrate and evaluate multiple parameters of cardiac structure and function in patients with cardiomyopathy. It has the potential to become an intelligent imaging diagnostic tool for cardiac diseases.