A Clinlabomics-Based Machine Learning Model Accurately Differentiates Atypical Hepatocellular Carcinoma from Atypical
Qing-Qing Luo1, Ding-Fan Guo1, Qiao-Nan Li1
1Department of Gastroenterology, Jiangxi Provincial Key Laboratory of Digestive Diseases, Jiangxi Clinical Research Center for Gastroenterology, Digestive Disease Hospital, The First Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, Jiangxi, People's Republic of China.
Journal of Hepatocellular Carcinoma
|May 13, 2026
Summary
This study developed a clinlabomics diagnostic model to differentiate atypical hepatocellular carcinoma (aHCC) from atypical benign focal hepatic lesions (aBFHL). The random forest model achieved high accuracy, proving valuable for early-stage and AFP-negative aHCC cases.
Area of Science:
- Hepatology
- Medical Imaging
- Machine Learning in Medicine
Background:
- Differentiating atypical hepatocellular carcinoma (aHCC) from atypical benign focal hepatic lesions (aBFHL) on radiological images presents a diagnostic challenge.
- Existing diagnostic methods may lack sufficient accuracy, particularly for subtle or early-stage cases.
Purpose of the Study:
- To develop and validate a novel diagnostic model using clinlabomics data for the differential diagnosis of aHCC and aBFHL.
- To leverage machine learning algorithms to improve diagnostic accuracy and clinical utility.
Main Methods:
- Retrospective collection of clinlabomic data from 466 pathologically diagnosed patients (252 aHCC, 214 aBFHL).
- Development of diagnostic models using three top algorithms (Random Forest, Support Vector Machine, Linear Discriminant Analysis) and six key features.
- Validation using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis; interpretability assessed via SHapley Additive exPlanations.
Main Results:
- The Random Forest (RF) model demonstrated optimal performance with an Area Under the ROC Curve (AUC) of 0.954 and 92.5% diagnostic accuracy on the testing set.
- The RF model showed superior performance for early-stage, small, and AFP-negative aHCCs, with AUCs ranging from 0.976 to 0.982.
- The model exhibited good calibration, clinical utility, interpretability, and was realized via an online calculator.
Conclusions:
- A clinlabomics-based diagnostic model, particularly the RF model, is effective for differentiating aHCC from aBFHL.
- The model shows significant value in diagnosing challenging cases, including early-stage, small, and AFP-negative aHCC.
- This approach offers a promising tool to enhance diagnostic accuracy in liver lesion characterization.
