To evaluate the diagnostic yield of ultrasonography (US)-guided core needle biopsy (CNB) in the diagnosis of soft tissue tumors (STTs) and to analyze the failure factors.

139 patients with STTs that underwent both US-guided CNB and surgical resection were collected retrospectively. Compared with the histopathological results of surgical resection, the biopsy failure was defined as the following conditions: indefinitive diagnosis, including insufficient samples and unknown subtypes with correct biological potential classification; wrong diagnosis, including wrong biological potential classification and wrong subtypes with correct biological potential classification. Univariate and multivariate analyses from the perspectives of histopathological, demographic and US features together with biopsy procedures were performed to determine risk factors for diagnostic failure.

The diagnostic yield of US-guided CNB for STTs in our study was 78.4%, but when only considering the correct biological potential classification of STTs, the diagnostic yield was 80.6%. The multivariate analysis showed that adipocytic tumors (odds ratio (OR) = 10.195, 95% confidence interval (CI): 1.062 - 97.861, p = 0.044), vascular tumors (OR = 41.710, 95% CI: 3.126 - 556.581, p = 0.005) and indeterminate US diagnosis (OR = 8.641, 95% CI: 1.852 - 40.303, p = 0.006) were correlated with the diagnostic failure. The grade III vascular density (OR = 0.019, 95% CI: 0.001 - 0.273, p = 0.007) enabled a higher diagnostic accuracy.

US-guided CNB can be an effective modality for the diagnosis of STTs. The diagnostic yield can be increased when the tumor vascular density was grade III in Color Doppler US, but can be decreased in adipocytic tumors, vascular tumors and masses with indeterminate US diagnosis.