Analysis of the value of artificial intelligence combined with musculoskeletal ultrasound in the differential diagnosis of pain rehabilitation of scapulohumeral periarthritis

Abstract

To explore the value of artificial intelligence combined with musculoskeletal ultrasound in the differential diagnosis of pain rehabilitation of scapulohumeral periarthritis. A total of 165 patients with periarthritis of the shoulder, who were admitted to our hospital from January 2020 to January 2022, were selected. Konica SONIMAGE HS1 PLUS color Doppler ultrasound diagnostic instrument was used to detect muscles and bones of patients with scapulohumeral periarthritis. This study proposed an intelligent clustering analysis algorithm with musculoskeletal ultrasound parameters. The neural network was trained on a GeForce RTX 3060 with the Adam W optimizer, a batch size of 12, and an initial learning rate of 5E-4. Each batch of 2 types of trained samples was input into the network in a specific proportion. Pain was assessed on a 10-point visual analogue scale. On the affected side of scapulohumeral periarthritis, the mild pain group showed thickening of the shoulder posterior capsule (2.02 ± 0.72) mm with clear edges. In the moderate pain group, the thickness of the shoulder posterior capsule gradually narrowed (1.01 ± 0.38) mm and became even thinner than that of the unaffected side, with irregular and blurred edges. In the severe pain group, the thickness of the shoulder posterior capsule largely returned to normal (1.21 ± 0.42) mm, and the edge was clear. Multivariate logistic regression showed that in addition to musculoskeletal ultrasound parameters, the length of service, work nature, and the busy degree of work of patients with periarthritis of the shoulder were significant factors that influenced shoulder periarthritis pain (P < .05). The performance of the proposed intelligent auscultation algorithm was further examined in a clinical setting, in which we used 165 clinical musculoskeletal ultrasound samples (including 81 positive and 84 negative samples) as a test set. The accuracy, sensitivity, and specificity were 0.833, 0.872, and 0.801, respectively. Musculoskeletal ultrasound combined with artificial intelligence algorithm is a new diagnostic and staging tool for scapulohumeral periarthritis.