A Ternary Classification Model Based on Ultrasound Images for Non-Invasive Staging of Hepatic Fibrosis In Patients with Autoimmune Liver Disease
Abstract
Purpose
Liver biopsy is the gold standard for evaluating liver fibrosis, yet it is invasive and subject to sampling error and interobserver variability. We therefore developed a B-mode ultrasound-based deep learning model for noninvasive staging of liver fibrosis in patients with autoimmune liver disease (AILD).
Methods
We retrospectively enrolled 245 consecutive patients with AILD and randomly assigned them to the training set (60%), validation set (20%), and internal testing set (20%). Additionally, 61 biopsy-confirmed AILD patients from another hospital were recruited as an external testing set. A deep learning model was constructed using the ResNet34 network architecture based on two-dimensional B-mode ultrasound images to evaluate its performance in liver fibrosis staging. Model performance was assessed using metrics such as macro and micro area under the curve (AUC). Calibration curves and decision curves were employed to evaluate model goodness of fit and clinical utility, and class activation mapping was used for model interpretation.
Results
The model demonstrated robust performance across different datasets. In the internal and external test sets, the macro-average AUCs were 0.812 (0.692–0.901) and 0.801 (0.688–0.902), respectively, while the micro-average AUCs were 0.819 (0.717–0.900) and 0.847 (0.761–0.911), respectively. The calibration and decision curves indicated favorable goodness-of-fit and clinical utility, and the class activation maps revealed the model’s decision-making rationale, enhancing interpretability.
Conclusion
The ultrasound-based deep learning model shows promise for noninvasive staging of liver fibrosis in patients with autoimmune liver disease, enabling clinicians to refine treatment strategies and enhance the timeliness and precision of interventions.