Andersen, Jesper B.

This review explores how multimodal foundation models (MFMs) are transforming biliary tract cancer (BTC) research. BTCs are aggressive malignancies with poor prognosis, presenting unique challenges due to difficult diagnostic methods, molecular complexity, and rarity. Importantly, intrahepatic cholangiocarcinoma (iCCA), perihilar cholangiocarcinoma (pCCA), and distal bile duct cholangiocarcinoma (dCCA) represent fundamentally distinct clinical entities, with iCCA presenting as mass-forming lesions amenable to biopsy and targeted therapies, while pCCA manifests as infiltrative bile duct lesions with challenging diagnosis and primarily palliative management approaches. MFMs offer potential to advance research by integrating radiological images, histopathology, multi-omics profiles, and clinical data into unified computational frameworks, with applications tailored to these distinct BTC subtypes. Key applications include enhanced biomarker discovery that identifies previously unrecognizable cross-modal patterns, potential for improving currently limited diagnostic accuracy—though validation in BTC-specific cohorts remains essential—accelerated drug repurposing, and advanced patient stratification for personalized treatment. Despite promising results, challenges such as data scarcity, high computational demands, and clinical workflow integration remain to be addressed. Future research should focus on standardized data protocols, architectural innovations, and prospective validation studies. The integration of artificial intelligence (AI)-based methodologies offers new solutions for these historically challenging malignancies. However, current evidence for BTC-specific applications remains largely theoretical, with most studies limited to proof-of-concept designs or related cancer types. Comprehensive clinical validation studies and prospective trials demonstrating patient benefit are essential prerequisites for clinical implementation. The timeline for evidence-based clinical adoption likely extends 7–10 years, contingent on successful completion of validation studies addressing current evidence gaps.

The integration of artificial intelligence (AI) in medical imaging has revolutionized diagnostic capabilities, yet the black-box nature of deep learning models poses significant challenges for clinical adoption. Current explainable AI (XAI) approaches, including SHAP, LIME, and Grad-CAM, predominantly focus on post hoc explanations that may inadvertently undermine clinical decision-making by providing misleading confidence in AI outputs. This paper presents a systematic review and meta-analysis of 67 studies (covering 23 radiology, 19 pathology, and 25 ophthalmology applications) evaluating XAI fidelity, stability, and performance trade-offs across medical imaging modalities. Our meta-analysis of 847 initially identified studies reveals that LIME achieves superior fidelity (0.81, 95% CI: 0.78–0.84) compared to SHAP (0.38, 95% CI: 0.35–0.41) and Grad-CAM (0.54, 95% CI: 0.51–0.57) across all modalities. Post hoc explanations demonstrated poor stability under noise perturbation, with SHAP showing 53% degradation in ophthalmology applications (ρ = 0.42 at 10% noise) compared to 11% in radiology (ρ = 0.89). We demonstrate a consistent 5–7% AUC performance penalty for interpretable models but identify modality-specific stability patterns suggesting that tailored XAI approaches are necessary. Based on these empirical findings, we propose a comprehensive three-pillar accountability framework that prioritizes transparency in model development, interpretability in architecture design, and a cautious deployment of post hoc explanations with explicit uncertainty quantification. This approach offers a pathway toward genuinely accountable AI systems that enhance rather than compromise clinical decision-making quality and patient safety.

Intrahepatic cholangiocarcinoma remains a highly heterogeneous malignancy that has eluded effective patient stratification to date. The extent to which such heterogeneity can be influenced by individual driver mutations remains to be evaluated. Here, we analyzed genomic (whole‐exome sequencing, targeted exome sequencing) and epigenomic data from 496 patients and used the three most recurrently mutated genes to stratify patients ( IDH , KRAS , TP53 , “undetermined”). Using this molecular dissection approach, each subgroup was determined to possess unique mutational signature preferences, comutation profiles, and enriched pathways. High‐throughput drug repositioning in seven patient‐matched cell lines, chosen to reflect the genetic alterations specific for each patient group, confirmed in silico predictions of subgroup‐specific vulnerabilities linked to enriched pathways. Intriguingly, patients lacking all three mutations (“undetermined”) harbored the most extensive structural alterations, while isocitrate dehydrogenase mutant tumors displayed the most extensive DNA methylome dysregulation, consistent with previous findings. Conclusion : Stratification of intrahepatic cholangiocarcinoma patients based on occurrence of mutations in three classifier genes ( IDH , KRAS , TP53 ) revealed unique oncogenic programs (mutational, structural, epimutational) that influence pharmacologic response in drug repositioning protocols; this genome dissection approach highlights the potential of individual mutations to induce extensive molecular heterogeneity and could facilitate advancement of therapeutic response in this dismal disease. (H epatology 2018).

Abstract Cholangiocarcinomas are rare but aggressive cancers, with an increasing incidence and a dismal 5-year survival rate of less than 5%. Mutations in the genes encoding isocitrate dehydrogenase, IDH1 and IDH2, have been reported in ∼70% of gliomas, ∼20% of myeloid leukemias, ∼56% of chondrosarcomas, and ∼10% of melanomas. We discovered IDH1 and IDH2 mutations in 29 of 305 (9%) intrahepatic cholangiocarcinomas. Tumors with mutations in IDH1 or IDH2 had lower 5-hydroxymethylcytosine and higher 5-methylcytosine levels, as well as increased dimethylation of histone H3K79. Mutations in IDH1 or IDH2 were independently associated with longer time to recurrence of cholangiocarcinoma in multivariate analysis (p = 0.046). We identified 2,267 genes that were significantly hypermethylated in tumors with mutations in IDH1 or IDH2. Hypermethylated CpG sites were significantly enriched in CpG shores and in the 1500 bp upstream of annotated transcription start sites, suggesting a global regulation of transcriptional potential. Gene expression profiling of 7 IDH1/2-mutant and 20 IDH1/2-wildtype tumors revealed increased expression of citric acid cycle and oxidative phosphorylation enzymes, as well as decreased expression of cytoskeleton organization genes. Integrated analyses of DNA methylation and gene expression identified interleukin-6 and several of its downstream targets as candidates for silencing by DNA methylation. Conclusions: Cholangiocarcinomas with mutations in IDH1 or IDH2 represent a molecular subclass with distinct signaling pathway alterations. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1058. doi:1538-7445.AM2012-1058