Artificial Intelligence in Hepatocellular Carcinoma: Current Applications, Clinical Performance, and Barriers to...

Area of Science:

• Oncology
• Medical Imaging
• Artificial Intelligence

Background:

• Hepatocellular carcinoma (HCC) presents significant global mortality challenges due to complex risk profiles, surveillance limitations, diagnostic uncertainties, and difficulties in personalized prognosis.
• Existing management strategies for HCC are hindered by heterogeneity in patient risk, suboptimal surveillance effectiveness, diagnostic ambiguities in chronic liver disease, and challenges in tailoring post-treatment prognoses.

Purpose of the Study:

• To critically evaluate the applications of artificial intelligence (AI) across the entire spectrum of hepatocellular carcinoma (HCC) patient care.
• To assess the intended clinical roles, reported performance metrics, maturity of evidence, and implementation barriers for AI tools in HCC management.

Main Methods:

• Conducted a narrative review, structuring findings around key clinical decision points in the HCC care continuum.
• Distinguished between emerging proof-of-concept AI tools and those with stronger potential for clinical translation.
• Analyzed AI applications in risk stratification, surveillance, imaging-based diagnosis, pathology, treatment response prediction, and prognostication.

Main Results:

• AI demonstrates consistent promise in identifying high-risk HCC patients, enhancing lesion detection and characterization across various imaging modalities (ultrasound, CT, MRI).
• AI shows potential in assisting histopathologic classification and predicting outcomes like microvascular invasion, recurrence, survival, and response to locoregional therapies.
• Significant limitations exist, including retrospective and lesion-enriched study designs, lack of robust external validation for prognostic models, and variability in data and protocols.

Conclusions:

• AI offers potential near-term value in specific HCC clinical scenarios but most current systems require further development before routine adoption.
• The field of AI in HCC is transitional, with future impact dependent on improved algorithms, defined use cases, prospective validation, and equitable implementation strategies.
• Addressing uneven evidence, ensuring robust validation, and developing clear implementation pathways are crucial for realizing AI's full potential in HCC care.