Radiology. 2026 Aug;320(2):e250278. doi: 10.1148/radiol.250278.
ABSTRACT
Background Differentiating benign from malignant soft tissue tumors (STTs) remains challenging, and conventional US alone is often insufficient. This study evaluated whether elastography and microvascular imaging improve diagnostic accuracy. Purpose To assess the ability of US techniques-B-mode US, shear-wave elastography (SWE), and superb microvascular imaging (SMI)-to discriminate benign from nonbenign STTs. Materials and Methods This tertiary referral center retrospective diagnostic accuracy study included consecutive patients with STTs referred for biopsy between February 2022 and July 2023. B-mode US features were recorded, including largest diameter (LD), deepest diameter (DD), DD-to-LD ratio, and SWE and SMI values. Histologic examination served as the reference standard. Receiver operating characteristic curves were used to evaluate predictions from single-variable and multivariable logistic regression models, including age and imaging-derived variables. The area under the curve (AUC) was used to summarize performance. Results A total of 167 lesions from 167 patients (mean age, 59.3 years ± 13.8 [SD]; 88 women) were included. Mean SWE and SMI values were higher in nonbenign than in benign STTs (SWE value, 4.00 m/sec vs 2.65 m/sec; SMI value, 31.5% vs 6.2%; both P < .001). DD and DD-to-LD ratio also differed between benign and nonbenign lesions (mean DD, 2.55 cm vs 4.57 cm; DD-to-LD ratio, 0.37 vs 0.51; both P < .001). Receiver operating characteristic analysis revealed AUCs of 0.82 (95% CI: 0.73, 0.92) for SMI and 0.75 (95% CI: 0.66, 0.85) for SWE. SMI sensitivity and specificity were 71% (22 of 31; 95% CI: 55, 87) and 87% (94 of 108; 95% CI: 81, 93), respectively; SWE sensitivity and specificity were 75% (24 of 32; 95% CI: 60, 90) and 71% (89 of 126; 95% CI: 63, 79). Combining LDs and DDs with SWE and SMI improved AUC to 0.87. Conclusion SWE and SMI improved diagnostic accuracy in evaluating STTs. © RSNA, 2026 Supplemental material is available for this article. See also the editorial by Lee and Stewart in this issue.
PMID:42640188 | DOI:10.1148/radiol.250278
