Liver and Bile
J Hepatol. 2025;83(2):426-439
Contrast-enhanced ultrasound-based AI model for multi-classification of focal liver lesions
Background and aims: Accurate multi-classification is a prerequisite for appropriate management of focal liver lesions (FLLs). Ultrasound is the most common imaging examination but lacks accuracy. Contrast-enhanced ultrasound (CEUS) offers better performance but is highly dependent on operator experience. Therefore, the authors aimed to develop a CEUS-based artificial intelligence (AI) model for FLL multi-classification and evaluate its performance in multicenter clinical tests.
Methods: From January 2017 to December 2023, CEUS videos, immunohistochemical biomarkers and clinical information on solid FLLs > 1 cm in adults were collected from 52 centers to build and test the model. The model was developed to classify FLLs into six types: hepatocellular carcinoma, hepatic metastasis, intrahepatic cholangiocarcinoma, hepatic hemangioma, hepatic abscess and others. First, Module-Disease, Module-Biomarker and Module-Clinic were built in training set A and a validation set. Then, three modules were aggregated as Model-DCB in training set B and an internal test set. Model-DCB performance was compared with CEUS and MRI radiologists in three external test sets.
Results: In total, 3,725 FLLs from 52 centers were divided into training set A (n = 2,088), the validation set (n = 592), training set B (n = 234), the internal test set (n = 110), and external test sets A (n = 113), B (n = 276) and C (n = 312). In external test sets A, B and C, Model-DCB achieved significantly better performance (accuracy from 0.85 to 0.86) than junior CEUS radiologists (0.59–0.73), and comparable performance to senior CEUS radiologists (0.79–0.85) and senior MRI radiologists (0.82–0.86). In multiple subgroup analyses on demographic characteristics, tumor characteristics and ultrasound devices, its accuracy ranged from 0.79 to 0.92.
Conclusions: CEUS-based Model-DCB provides accurate multi-classification of FLLs. It holds promise for a wide range of populations, especially those in remote areas who have difficulty accessing MRI.