Texture analysis of T1- and T2-weighted images identifies myofiber atrophy and grip strength decline in streptozotocin-induced type 1 diabetic sarcopenia rats

Abstract

Background: Type 1 diabetes mellitus is associated with accelerated skeletal muscle aging and sarcopenia, a condition characterized by muscle mass and function loss. Early and noninvasive evaluation of muscle microstructural damage is critical for managing sarcopenia in diabetes. This study evaluated the potential of MRI texture analysis as a noninvasive imaging tool to assess myofiber size and grip strength alterations in a rat model of diabetic sarcopenia.

Materials and methods: Twenty SD rats were randomly and evenly allocated to the control (CON) and diabetic sarcopenia (DS) groups. Blood glucose, body weight, and forelimb grip strength were measured weekly. In the eighth week, axial T₁WI and T₂WI scans were performed to extract 16 texture features of the gastrocnemius muscle. The size of the fibers, including the cross-sectional area, perimeter, and minFeretDiam, was calculated from the HE-stained images. Differences between the two groups were analyzed via t tests or Mann‒Whitney U tests. Receiver operating characteristic (ROC) analysis was conducted to evaluate the diagnostic performance of MRI texture features. Associations between MRI texture features, grip strength, and myofiber size were assessed via Spearman correlation analysis.

Results: The DS group presented significant reductions in body weight, grip strength, and myofiber size (p < 0.001). Seven texture features demonstrated high repeatability. Among them, T₁WI_Entropy and T₂WI_Entropy were significantly lower, whereas T₁WI_Mean, T₂WI_Mean, and T₂WI_Variance were greater in the DS group than in the control group (p < 0.05 to p < 0.001). ROC analysis revealed that the MRI texture features performed well in differentiating the DS and CON groups (AUC 0.830 ~ 1.000). Strong correlations were found between MRI texture parameters and muscle strength and myofiber size.

Conclusion: Muscle texture analysis based on T₁WI and T₂WI effectively differentiated type 1 diabetic sarcopenic rats from normal controls and revealed strong associations with muscle strength and myofiber size. These findings indicate that MRI texture parameters serve as potential biomarkers for diagnosing muscle damage in type 1 diabetic sarcopenia patients, suggesting that MRI texture parameters are promising noninvasive tools for early detection and monitoring of myofiber atrophy in diabetic sarcopenia.

Keywords: Magnetic resonance imaging; Muscle strength; Myofiber atrophy; Sarcopenia; Skeletal muscle; Texture analysis.

Fig. 1

The cross-sectional regions of interest (ROIs) of the right gastrocnemius muscle in normal control (CON) and diabetic sarcopenia (DS) group rats, manually drawn on axial T₁-weighted imaging (T₁WI) and T₂-weighted imaging (T₂WI)

Fig. 2

Body weight (A), muscle strength (B), and cross-sectional area (CSA) of the gastrocnemius muscle on MRI (C) in the diabetic sarcopenia (DS) group and control (CON) group at week 8. The data are presented as the means ± standard deviations. *** p < 0.001

Fig. 3

Type 1 diabetes-related sarcopenia leads to muscle fiber atrophy in rats. A Hematoxylin and eosin (HE) staining (20 ×) of the right gastrocnemius muscle and muscle fiber cross-section segmentation (Seg) images in the diabetic sarcopenia (DS) group and control (CON) group. Myofiber cross-sectional area (CSA) (B), perimeter (C), and minimum Feret diameter (minFeretDiam) (D) at week 8. The data are presented as the means ± standard deviations. *** p < 0.001. Scale bar = 50 μm

Fig. 4

Heatmap of repeatable muscle texture features in the diabetic sarcopenia (DS) and control (CON) groups

Fig. 5

Correlation plots showing Spearman correlation values between MRI muscle texture features rat body measurements, and skeletal muscle histological characteristics. CSA = cross-sectional area