Deep learning and machine learning in image-based hepatocellular carcinoma detection: a systematic review and
Jingchao Wang1, Jingjie Bai1, Yanyang Wang1
1Department of ultrasound, Hebei Medical University Third Hospital, Shijiazhuang, China.
Abdominal Radiology (New York)
|June 30, 2026
Summary
Artificial intelligence (AI) shows high accuracy in detecting hepatocellular carcinoma (HCC) from medical images, outperforming human clinicians. Further prospective studies are needed for robust validation in clinical practice.
Area of Science:
- Medical Imaging Analysis
- Artificial Intelligence in Oncology
- Hepatocellular Carcinoma Diagnostics
Background:
- Early detection and management are crucial for hepatocellular carcinoma (HCC) patient outcomes.
- Artificial intelligence (AI) is increasingly utilized to improve the precision and efficiency of diagnostic imaging for liver tumors.
Purpose of the Study:
- To systematically review and meta-analyze the diagnostic performance of AI algorithms for HCC detection.
- To compare the diagnostic accuracy of AI algorithms against human clinicians in identifying HCC.
Main Methods:
- A comprehensive literature search was performed across major databases (PubMed, Embase, etc.) in November 2024.
- Diagnostic performance data from 41 AI studies were pooled using meta-analysis, with comparisons made to 11 studies of human clinicians.
- Subgroup analyses examined performance based on classification task, validation type, AI approach (DL vs. ML), and imaging modality (US, CT, MRI).
Main Results:
- AI algorithms demonstrated high pooled diagnostic accuracy: sensitivity 87%, specificity 91%, AUC 0.95.
- AI performance metrics were numerically superior to those of human clinicians (SEN 78%, SPC 91%, AUC 0.91).
- Consistent high performance was observed across various subgroups, though significant heterogeneity was noted.
Conclusions:
- AI algorithms exhibit high diagnostic accuracy for image-based HCC detection, with potential to exceed human performance.
- Prospective studies with rigorous external validation and larger sample sizes are essential to confirm AI's role in HCC detection.