Wenting Luo, Jia Liu, Jie Wang, Yun Yuan, Jianxiu Lian, Wei Li, Shouyi Wei, Xiaofeng Qin, Wei Ma, Jianxing Qiu
In dystrophinopathy, multiparametric CMR enables structural and functional assessment of myocardial remodeling. Segmental native T1, particularly in the basal and mid-inferolateral segments, provides complementary information on regional myocardial tissue abnormalities. In contrast, CS accounts for most of the discrimination of disease severity. These findings suggest that the distinctive contribution of CMR may lie in regional myocardial tissue characterization rather than conventional functional assessment.
BACKGROUND: Cardiac involvement in dystrophinopathy progresses from diffuse myocardial remodeling to overt systolic failure. Multiparametric cardiac magnetic resonance (CMR) offers simultaneous insights into myocardial mechanics and tissue characterization, yet the optimal diagnostic parameters for tracking disease severity remain to be fully characterized.
OBJECTIVE: To objectively evaluate the descriptive and discriminative performance of integrating strain and quantitative T1 mapping parameters for identifying myocardial fibrosis burden and stratifying systolic dysfunction in patients with dystrophinopathy.
MATERIALS AND METHODS: This retrospective study analyzed 55 patients with genetically confirmed dystrophinopathy and 30 healthy controls who underwent CMR between January 2020 and December 2024. Patients were sequentially stratified by late gadolinium enhancement (LGE) status (Group A: LGE-negative, n = 16; Group B: LGE-positive, n = 39), and further sub-stratified by left ventricular ejection fraction (LVEF) (Group C: LVEF ≥ 50%, n = 23; Group D: LVEF < 50%, n = 16). Least absolute shrinkage and selection operator (LASSO) regression with 10-fold cross-validation was applied for non-zero feature selection, followed by the construction of multivariable Firth penalized logistic regression models. Model comparisons were made using the non-parametric DeLong test.
RESULTS: Compared with Group A, Group B exhibited significant differences in LVEF (p < .001), mid-ventricular circumferential strain (CS_Mid, p < .001), and native T1 in segment 11 (T1map_seg11, mid-inferolateral, p = .004). In single-parameter ROC analysis, CS_Mid, LVEF, and T1map_seg11 exhibited an area under the curve (AUC) of 0.842 [95% confidence interval (CI): 0.735-0.949], 0.828 (95% CI: 0.714-0.942), and 0.715 (95% CI: 0.549-0.880), respectively. The combined model of CS_Mid and T1map_seg11 achieved the AUC of 0.888 (95% CI: 0.793-0.983). Compared Group C with Group D, CS_Global and T1map_seg5 models yielded AUCs of 0.921 (95% CI: 0.822-1.000) and 0.823 (95% CI: 0.694-0.953). The combined model (CS_Global + T1map_seg5) achieved the AUC of 0.951 (95% CI, 0.864-1.000, DeLong p = .567 vs T1map_seg5, DeLong p = .101 vs. T1map_seg5).
CONCLUSIONS: In dystrophinopathy, multiparametric CMR enables structural and functional assessment of myocardial remodeling. Segmental native T1, particularly in the basal and mid-inferolateral segments, provides complementary information on regional myocardial tissue abnormalities. In contrast, CS accounts for most of the discrimination of disease severity. These findings suggest that the distinctive contribution of CMR may lie in regional myocardial tissue characterization rather than conventional functional assessment.