Potential use of radiomics analysis of cine-mode cardiac MRI to detect post-infarction lesions in the left ventricular myocardium

Abstract

BACKGROUND: Size and localization, as well as clear differentiation between intact tissue and infarcted area, are important for clinical diagnosis and precision medicine. This work is based on the study of radiomics features capable of differentiating areas of infarcted and distant tissue from the infarct area using data from non-contrast cine cardiac MRI imaging. AIM: Evaluation of the possibilities and informativeness of radiomics analysis in detection of postinfarction areas of the left ventricular myocardium in patients with ischemic cardiomyopathy by contrast-free cine cardiac MRI images. MATHERIALS AND METHODS: The results of contrast enhancement cardiac MRI with of 33 patients undergoing surgical treatment for ischemic cardiomyopathy were analyzed. Texture analysis was performed on 66 sections of cine cardiac MRI images, and 105 texture characteristics were determined for each of them. Cardiac MRI was performed according to the standard technique on a Vantage Titan (Toshiba) 1.5 T. Texture analysis was performed using 3D slicer-version 5.2.2 software, SlicerRadiomics RESULTS: Collinearity diagrams of the features were constructed. We also identified the features with zero importance and established the importance of the features using the gradient boosting algorithm. In addition, the cumulative importance of the features as a function of their total number was estimated. Using the method of identifying features with low importance, we identified the parameters with the lowest importance that do not affect the indicated total level. Using the method of identifying features with a single value, we did not find features with a single unique value. In addition, ROC curve for Lasso logistic regression (Se =57.14% Sp=71.43%, AUC=0.76) was generated based on the results of the analysis. The main result of this study is the identification of radiomics features able to characterize the areas corresponding to postinfarction cardiosclerosis and intact LV wall on the basis of cine cardiac MRI images. CONCLUSION: This study has shown that the use of radiomics analysis on contrast-free cine cardiac MRI images is a promising approach to identify areas corresponding to myocardial infarction and intact wall and could potentially be used as to identify areas of postinfarction cardiosclerosis in patients with ischemic cardiomyopathy without the use of contrast agents.