BioFact-MoE: Biologically Factorized Mixture of Experts for Vision-Language Prognostic Modeling in Hepatocellular Carcinoma

2026-05-25

AI Summary

This study addresses the significant biological heterogeneity of hepatocellular carcinoma (HCC) and the limitations of existing vision–language prognostic models, which conflate liver function and tumor-related factors. To overcome this, the authors propose a biologically informed mixture-of-experts (MoE) framework that explicitly disentangles hepatic and tumoral phenotypes through a pathway-guided gating mechanism within a residual survival network. The model integrates multiparametric MRI and radiology reports, introducing biologically supervised factorization into vision–language prognostic modeling for the first time. Evaluated on 588 HCC patients, the approach achieves AUCs of 75.33%, 75.85%, and 73.96% at 12, 18, and 24 months, respectively, significantly outperforming baseline methods. The disentangled embeddings correlate strongly with liver function and tumor burden (p < 0.05), revealing treatment-associated survival heterogeneity.

Abstract

Hepatocellular carcinoma (HCC) is biologically heterogeneous, shaped by the interplay between hepatic functional reserve and tumor-related oncologic factors; thus, similar survival outcomes may reflect fundamentally different underlying biological processes. Prognostic modeling in HCC is informed by rich multimodal information from multiparametric MRI and radiology reports from routine clinical practice. Existing prognostic vision-language models (VLMs) learn a single entangled latent representation that blends hepatic and tumor-related factors, limiting both accuracy and biological interpretability. We present BioFact-MoE, a biologically factorized Mixture of Experts (MoE) framework that explicitly decomposes liver and tumor factors via biologically supervised experts within a residual MoE survival architecture. On a HCC cohort of N=588 patients (pretrained on 4,582 3D MRI image-report pairs), BioFact-MoE consistently improves survival prediction over all baselines across time horizons, achieving 12-, 18-, and 24-month AUCs of 75.33%, 75.85%, and 73.96%. Beyond scalar risk prediction, gated expert weights enable phenotype-aware risk stratification. Pathway-informed gating uncovers clinically meaningful treatment-associated survival heterogeneity. In held-out validation, hepatic and tumor embeddings show selective associations with liver function and tumor burden markers, respectively (p<0.05), without supervision. The code is available at https://github.com/jy-639/BioFact-MoE.

Problem

• hepatocellular carcinoma
• prognostic modeling
• biological heterogeneity
• vision-language models
• multimodal representation

Innovation

• Biologically Factorized MoE
• Vision-Language Prognostic Modeling
• Hepatocellular Carcinoma
• Interpretable Survival Prediction
• Multimodal Representation Learning