Texture analysis of cardiovascular MRI native T1 mapping in patients with Duchenne muscular dystrophy.

Abstract

Background: Duchenne muscular dystrophy (DMD) patients are monitored periodically for cardiac involvement, including cardiac MRI with gadolinium-based contrast agents (GBCA). Texture analysis (TA) offers an alternative approach to assess late gadolinium enhancement (LGE) without relying on GBCA administration, impacting DMD patients' care. The study aimed to evaluate the prognostic value of selected TA features in the LGE assessment of DMD patients.

Results: We developed a pipeline to extract TA features of native T1 parametric mapping and evaluated their prognostic value in assessing LGE in DMD patients. For this evaluation, five independent TA features were selected using Boruta to identify relevant features based on their importance, least absolute shrinkage and selection operator (LASSO) to reduce the number of features, and hierarchical clustering to target multicollinearity and identify independent features. Afterward, logistic regression was used to determine the features with better discrimination ability. The independent feature inverse difference moment normalized (IDMN), which measures the pixel values homogeneity in the myocardium, achieved the highest accuracy in classifying LGE (0.857 (0.572-0.982)) and also was significantly associated with changes in the likelihood of LGE in a subgroup of patients with three yearly examinations (estimate: 23.35 (8.7), p-value = 0.008). Data are presented as mean (SD) or median (IQR) for normally and non-normally distributed continuous variables and numbers (percentages) for categorical ones. Variables were compared with the Welch t-test, Wilcoxon rank-sum, and Chi-square tests. A P-value < 0.05 was considered statistically significant.

Conclusion: IDMN leverages the information native T1 parametric mapping provides, as it can detect changes in the pixel values of LGE images of DMD patients that may reflect myocardial alterations, serving as a supporting tool to reduce GBCA use in their cardiac MRI examinations.