The Value of High-Resolution Ultrasound Combined with Shear-Wave Elastography under Artificial Intelligence Algorithm in Quantitative Evaluation of Skin Thickness in Localized Scleroderma

Abstract

The aim of this study was to explore the value of high-resolution ultrasound combined with shear-wave elastography (SWE) in measuring skin thickness in patients with localized scleroderma (LS). Fifty patients with LS diagnosed by pathology in the hospital were selected as the research object, with a total of 96 lesions. Healthy people (50 cases) in the same period were selected as the control group. The skin thickness of the abdomen, chest, and left finger of the two groups was compared. The traditional nonlocal means (NLM) algorithm was improved by changing the Euclidean distance and introducing a cosine function, which was applied to the ultrasonic imaging intelligent diagnosis of patients with localized scleroderma. SWE imaging was evaluated, and the results demonstrated that LS lesion edema stage accounted for 7.29%, hardening stage occupied 43.75%, and the proportion of atrophy stage reached 48.96%. When the size of shell was 1 mm, maximum elastic modulus (Emax) was 0.984, mean of elastic modulus (Emean) was 0.926, and electro-static discharge (Esd) was 0.965. When the size of shell was 2 mm, the elastic moduli around lesions were as follows: Emax was 0.998, Emean was 0.968, and Esd was 0.997. By comparing the skin thickness of the abdomen, chest, and left finger, it was found that there was a significant difference between the LS group and the control group ( $P<0.05$ ). When the shell was 2 mm, the effect of sensitivity specificity on SWE imaging was better than that when the shell was 1 mm. In summary, the improved NLM algorithm showed excellent denoising effects on the ultrasonic images of LS patients. Besides, it could assist clinicians in ultrasonic imaging diagnosis for LS patients and effectively improve the diagnostic accuracy of diseases.