To explore an effective noninvasive method for differential diagnosis of liver tumors.

A retrospective study was performed to assess the performance of certain ultrasound-based parameters, in distinguishing benign from malignant with liver biopsy as the gold standard. Demographic data, clinical information and serological data were obtained from participants who were retrospectively enrolled. Receiver operating characteristic (ROC) curve analysis was conducted, including calculation of the optimal area under the ROC curve (AUC), to assess the performance of these measures in distinguishing between benign and malignant in patients with or without chronic liver disease.

Inclusion Criteria:

(1) age 18-70 years; (2) liver examination performed for all patients prior to US-guided liver biopsy or surgery; (3) tumor can be visualized in conventional US; (4) largest tumor sectional area saved in DICOM format for use in later analyses; (5) a histopathological diagnosis made either from US-guided liver biopsy or surgical resection; (6) suitable routine clinical and US data obtained within 7 days of the histopathological diagnosis.

Exclusion Criteria:

(1) unavailable pathological diagnosis; (2) liver metastatic tumors; (3) poor quality US images; and (4) a past history of surgical resection, percutaneous local ablation, radiation, transarterial and systemic therapies or liver transplantation.

A total of 200 patients whose pathological diagnoses were confirmed by US-guided liver biopsy or surgery and who underwent preoperative conventional US between January 2018 and November 2025 were enrolled in the study.