Multi-Omics Integration and Artificial Intelligence for Dynamic Risk Stratification of Hepatocellular Carcinoma in MASLD: Toward Precision Surveillance and Personalized Immunotherapy

Abstract

Background: Hepatocellular carcinoma (HCC) is the most common primary liver malignancy and remains a major cause of cancer-related mortality worldwide. Metabolic dysfunction-associated steatotic liver disease (MASLD) has emerged as an important driver of hepatocarcinogenesis, characterized by heterogeneous disease progression and the development of HCC even in non-cirrhotic patients, thereby challenging conventional surveillance strategies.

Objectives: This review aimed to evaluate the role of multi-omics integration and artificial intelligence (AI) in the early detection and dynamic risk stratification of MASLD-related HCC, synthesize evidence regarding the association between multi-layer biomarkers, biological heterogeneity, and immunotherapy response prediction, and formulate a conceptual framework for dynamic risk stratification based on longitudinal molecular trajectories as a foundation for precision surveillance and personalized therapeutic strategies.

Methods: A structured narrative literature review using a conceptual synthesis approach was conducted. Literature searches were performed in PubMed, Scopus, ScienceDirect, and Google Scholar for studies published between 2021 and 2026. Studies investigating genomics, transcriptomics, proteomics, metabolomics, or epigenomics integrated with AI-based approaches for HCC detection, risk stratification, or therapeutic response prediction were qualitatively synthesized.

Discussion: Current evidence suggests that multi-omics integration provides a more comprehensive understanding of molecular heterogeneity and hepatocarcinogenic trajectories than single-biomarker approaches. AI facilitates the integration of multidimensional data through complex modeling, enabling dynamic risk stratification, improving early detection, and predicting immunotherapy response.

Conclusion: Multi-omics and AI have the potential to transform risk-based surveillance and personalized management of MASLD-related HCC. However, prospective multicenter validation, standardized integration strategies, and adequate healthcare infrastructure are required for broader clinical implementation.

Keyword: artificial intelligence; hepatocellular carcinoma; MASLD; multi-omics; precision hepatology

Latar Belakang: Karsinoma hepatoseluler (hepatocellular carcinoma/HCC) merupakan keganasan hati primer tersering dan masih menjadi penyebab utama mortalitas terkait kanker di seluruh dunia. Metabolic dysfunction-associated steatotic liver disease (MASLD) telah berkembang sebagai determinan penting dalam hepatokarsinogenesis dengan karakteristik progresi penyakit yang heterogen serta kemampuan berkembang menjadi HCC bahkan pada pasien tanpa sirosis, sehingga menantang strategi surveilans konvensional.  

Tujuan: Tinjauan ini bertujuan mengevaluasi peran integrasi multi-omics dan kecerdasan buatan (artificial intelligence/AI) dalam deteksi dini dan stratifikasi risiko dinamis HCC terkait MASLD, mensintesis bukti mengenai hubungan biomarker multi-lapis dengan heterogenitas biologis dan prediksi respons imunoterapi, serta merumuskan kerangka konseptual stratifikasi risiko dinamis berbasis trajektori molekuler longitudinal sebagai dasar pengembangan surveilans presisi dan strategi terapi yang dipersonalisasi.

Metode: Tinjauan literatur naratif terstruktur dengan pendekatan sintesis konseptual dilakukan melalui pencarian literatur pada PubMed, Scopus, ScienceDirect, dan Google Scholar terhadap studi yang dipublikasikan pada tahun 2021–2026. Studi yang mengevaluasi integrasi genomik, transkriptomik, proteomik, metabolomik, atau epigenomik dengan pendekatan berbasis AI untuk deteksi HCC, stratifikasi risiko, atau prediksi respons terapi dianalisis secara kualitatif.

Pembahasan: Bukti terkini menunjukkan bahwa integrasi multi-omics memberikan pemahaman yang lebih komprehensif mengenai heterogenitas molekuler dan trajektori hepatokarsinogenesis dibandingkan pendekatan biomarker tunggal. AI memungkinkan integrasi data multidimensi melalui pemodelan kompleks untuk mendukung stratifikasi risiko dinamis, meningkatkan deteksi dini, serta memprediksi respons imunoterapi.

Kesimpulan: Integrasi multi-omics dan AI berpotensi mentransformasi surveilans berbasis risiko dan tata laksana personalisasi HCC terkait MASLD. Namun, validasi prospektif multisenter, standardisasi strategi integrasi data, serta kesiapan infrastruktur kesehatan masih diperlukan untuk implementasi klinis yang lebih luas.

Kata kunci: kecerdasan buatan; karsinoma hepatoseluler; MASLD; multi-omics; hepatologi presisi