Molecular MRI 기반의 동맥경화증 및 혈관염의 시각화 평가

Abstract

This study evaluates the potential of molecular magnetic resonance imaging (mMRI) with the specific contrast agent lmNMC for visualizing atherosclerotic plaques and vascular inflammation in early cardiovascular disease diagnosis. Using an animal model with induced atherosclerosis, MRI scans were obtained before and after the contrast agent injection, followed by analysis of the differential images. Signal intensity analysis showed a significant increase from 58.51 to 66.36 after contrast injection, while Discrete Wavelet Transform (DWT) analysis revealed prominent high-frequency components in the horizontal, vertical, and diagonal directions in the differential images, highlighting the histological boundaries of lesions. Gray-level co-occurrence matrix (GLCM) texture analysis showed an increase in contrast (0.0846 -> 0.1021), a decrease in energy (0.2256 -> 0.2019), and an increase in correlation (0.9659 -> 0.9708), indicating increased tissue heterogeneity and structural consistency. Furthermore, a Neural Network Regression model was used to learn the nonlinear relationship between DWT features extracted from differential images and lesion segmentation. The model achieved high prediction accuracy with R= 0.93244, RMSE = 28.60, and R-2 = 0.87, and demonstrated excellent performance with Precision = 1.0, Recall = 0.9844, and F1 Score = 0.9921. This study demonstrates that lmNMC-based mMRI provides a robust method for the quantitative visualization and early diagnosis of atherosclerotic lesions, offering significant potential as a key tool in precision medical imaging analysis.