Integrated Conventional Ultrasound, Shear-Wave Elastography, and Radiomics for Risk Stratification of Superficial-Organ and Musculoskeletal Lesions: A Prospective Clinical Study

Superficial-organ and musculoskeletal lesions are commonly evaluated by ultrasound, but benign and malignant lesions often show overlapping morphological features. Conventional ultrasound remains the first-line imaging tool, whereas shear-wave elastography and ultrasound-based radiomics may provide additional quantitative information for lesion risk stratification. This study evaluated whether an integrated model combining conventional ultrasound, shear-wave elastography, and ultrasound-based radiomics improved risk stratification for superficial-organ and musculoskeletal lesions compared with conventional ultrasound alone. This prospective study included 220 patients with thyroid, breast, superficial lymph-node, or musculoskeletal soft-tissue lesions who underwent conventional ultrasound and shear-wave elastography between 1 March 2025 and 31 December 2025. One target lesion was analyzed per patient. The final reference standard was histopathology or qualifying imaging follow-up. Conventional ultrasound features, elastography parameters, and radiomics features extracted from gray-scale ultrasound images were used to construct three models: a conventional ultrasound model, a conventional ultrasound plus elastography model, and an integrated model combining conventional ultrasound, elastography, and radiomics. The cohort was divided into training and validation sets at a 7:3 ratio. Model performance was assessed using discrimination, calibration, decision-curve analysis, and prespecified risk-group stratification. The final cohort included 149 benign and 71 malignant lesions. Malignant lesions showed higher mean stiffness, maximum stiffness, elasticity ratio, radiomics score, and integrated score than benign lesions. In the validation cohort, the integrated model achieved an AUC of 0.859, compared with 0.656 for the conventional ultrasound model, 0.803 for the elastography model, and 0.802 for the radiomics model. Using prespecified probability thresholds, the integrated model classified lesions into low-, intermediate-, and high-risk groups, with observed malignancy rates of 7.3%, 39.4%, and 90.0%, respectively. Decision-curve analysis showed positive net benefit across clinically relevant threshold probabilities. The integration of conventional ultrasound, shear-wave elastography, and ultrasound-based radiomics improved risk stratification for superficial-organ and musculoskeletal lesions. The model provided clinically interpretable risk categories and may support individualized biopsy or follow-up decisions after external validation.