A complementary patch-based histogram analysis for quantifying muscle tissue in ultrasound imaging

Introduction: Ultrasound imaging is widely used for muscle assessment due to its non-invasive nature and real-time imaging capabilities. Histogram-based echotexture analysis has proven to be a valuable tool for quantifying muscle tissue composition using features such as echointensity (EI) and echovariation (EV), especially in neuromuscular and neurological disorders. However, variability in region of interest (ROI) selection and image processing methods can significantly affect the extracted echotexture features. This study presents a novel histogram-based analysis approach to investigate the effects of subdividing a single ROI into different patch sizes for muscle tissue assessment. By focusing on the EI and EV of the gastrocnemius medialis in a stroke patient, this research aims to refine quantitative ultrasound analysis to improve clinical applicability. Case presentation: One stroke patient was randomly selected from a previously collected dataset to perform a complementary analysis. The initially selected grey-scale ROI was extracted and divided into patches of different sizes (10×10, 20×20, 30×30, 40×40 and 50×50 pixels). The EI and EV of each patch were calculated, and their distributions were analyzed using descriptive statistics and correlation methods. Results: The EV values for patches of sizes 10×10, 20×20, 30×30, 40×40, and 50×50 were 26.48, 24.58, 20.44, 16.78, and 10.38, respectively, which deviated significantly from the original ROI value of 45.54. In contrast, the EI values remained around 81 across all patch sizes, indicating that varying patch sizes did not affect EI. Conclusions: Patch-based histogram analysis offers a complementary method for assessing muscle texture in ultrasound. While EI appears to be robust to ROI subdivision, EV shows variability, raising concerns about its reliability when small patches are used. Standardized methods and future research with larger datasets are needed to optimize echotexture analysis and ensure reproducibility in clinical practice.