Abstract
PURPOSE: Dermatofibrosarcoma protuberans (DFSP) and dermatofibroma (DF) are cutaneous lesions with overlapping clinical features, often requiring histopathological confirmation. This study aims to evaluate and compare the diagnostic utility of high-frequency ultrasound (HFUS) and ultra-high-frequency ultrasound (UHFUS) in distinguishing these two entities over 15-year period. METHODS: A retrospective analysis was conducted on 334 patients (127 DFSP, 207 DF) with pathologically confirmed diagnoses. HFUS or UHFUS was used to assess lesion characteristics, including demographics, location, size, morphology, echogenicity, homogeneity, posterior acoustic features, and vascularity. Univariate and multivariate logistic regression analyses were performed to identify significant predictors. RESULTS: DFSP patients were significantly older than DF patients (40.99 years vs 34.00 years; P < 0.001). DFSP lesions were predominantly on the trunk, while DF was more common on the extremities (P < 0.001). DFSP lesions were significantly larger (mean 43.02 mm vs 10.34 mm; P < 0.001), and exhibited more aggressive sonographic features, including tentacle-like borders, internal hyperechoic areas, peripheral hyperechoic rims, mixed echogenicity, irregular shape, ill-defined margins, internal heterogeneity, and frequent posterior enhancement (all P < 0.005). DFSP also showed higher vascularity with random, peripheral, or arborizing patterns and higher Adler grades (all P < 0.001). Multivariate analysis identified tumor location (extremities favoring DF), size, ultrasound pattern (tentacle-like border pattern, internal hyperechoic area, peripheral hyperechoic rim, and mixed echogenicity pattern favoring DFSP) as independent predictors. CONCLUSION: HFUS and UHFUS demonstrates strong diagnostic utility in differentiating DFSP from DF based on key clinical and sonographic features. These findings support the use of HFUS and UHFUS as a valuable non-invasive tool for preoperative diagnosis. Future studies should validate these criteria in multi-center settings and exploring artificial intelligence integration to further enhance diagnostic accuracy and standardization.