Artificial intelligence and multi-omics driven models would be the future of intrahepatic cholangiocarcinoma prediction research

Intrahepatic cholangiocarcinoma (ICC) is a highly aggressive form of liver cancer and the incidence rate is increasing. Despite after curative surgery, postoperative prognosis remains poor for patients. While evidence supports the potential benefits of postoperative adjuvant therapy in reducing recurrence rates and enhancing survival outcomes, the specific beneficiaries of such treatment remain uncertain. Numerous prognostic models for ICC have been developed in recent years; however, their reliability and efficiency are constrained by small sample sizes, single-center research designs, and data limitations, thereby limiting their suitability for widespread clinical use.

Chen et al. (1) conducted a study in which they gathered multicenter ICC genomic data and developed an ICC prognostic methylation score (PMS) utilizing machine learning and methylation features. The primary findings of the study were:

• The PMS demonstrated strong predictive ability for overall survival (OS), with a C-index of 0.79 [95% confidence interval (CI): 0.75–0.84] in the discovery cohort and a C-index of 0.74 (95% CI: 0.68–0.80) in an independent external validation cohort. These results suggest that the PMS exhibits robust performance across different datasets.
• When analyzing the predictive performance of PMS in comparison to established models [Johns Hopkins University School of Medicine (JHUSM) line chart, Eastern Hepatobiliary Surgery Hospital (EHBSH) line chart, American Joint Committee on Cancer (AJCC) TNM staging system, and MEGNA (multifocality, extrahepatic extension, grade, node positivity, and age older than 60 years) prognostic score] for predicting OS, the findings indicate that PMS demonstrates a statistically significant and robust predictive capacity.
• Additionally, despite the relatively low utilization of adjuvant therapy and variability in treatment approaches within the cohort, the results suggest that adjuvant therapy can improve OS specifically in quartile 4 patients with PMS.
• The study investigated the mechanism underlying the result 3, revealing that PMS may be indicative of tumor biology, pathway activation, and immune cell infiltration to a certain degree. While comprehensive genome-wide methylation analysis remains costly and is not yet widely applicable in clinical settings, the intriguing findings presented by Chen et al. continue to inspire curiosity and enthusiasm.

As algorithmic capabilities progress, machine learning and other artificial intelligence (AI) algorithms are being increasingly utilized in the development of cancer prediction models. This trend is evident not only in genomics prediction models, as demonstrated by Chen et al., but also in the fields of radiomics and pathomics. Additionally, the acquisition of medical imaging and pathological data is typically more accessible compared to genomic data. A retrospective study developed a radiomics model used Flourine-18 fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT), which exhibited strong predictive capabilities for determining the differentiation grade [area under the curve (AUC) 0.78], presence of microvascular invasion (AUC 0.87), OS (AUC 0.81), and progression-free survival (AUC 0.81) in patients with ICC (2). Another multicenter study involving 345 ICC patients utilized a CT radiomics model to predict postoperative outcomes (3). The findings revealed a C-index of 0.68 for predicting postoperative recurrence-free survival (RFS), with ranging from 0.68 to 0.71 for OS. Yang et al. developed a radiomics model based on diffusion-weighted imaging (DWI), which demonstrated an AUC of 0.821 for early postoperative recurrence prediction in ICC (4). This model also enabled the stratification of ICC patient prognosis, suggesting that individuals in the high-risk group may benefit from adjuvant chemotherapy.

Furthermore, the utilization of AI and deep learning algorithms in conjunction with pathological images holds promise for predicting the prognosis of ICC. Additionally, a multi-omics approach incorporating genomics, radiomics, and pathomics, in tandem with AI algorithms, has the potential to achieve favorable outcomes and may represent the “super model” of ICC prediction models.

Acknowledgments

Funding: None.

Footnote

Provenance and Peer Review: This article was commissioned by the editorial office, Hepatobiliary Surgery and Nutrition. The article did not undergo external peer review.

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://hbsn.amegroups.com/article/view/10.21037/hbsn-24-136/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Chen X, Dong L, Chen L, et al. Epigenome-wide development and validation of a prognostic methylation score in intrahepatic cholangiocarcinoma based on machine learning strategies. Hepatobiliary Surg Nutr 2023;12:478-94. [Crossref] [PubMed]
2. Fiz F, Masci C, Costa G, et al. PET/CT-based radiomics of mass-forming intrahepatic cholangiocarcinoma improves prediction of pathology data and survival. Eur J Nucl Med Mol Imaging 2022;49:3387-400. [Crossref] [PubMed]
3. Park HJ, Park B, Park SY, et al. Preoperative prediction of postsurgical outcomes in mass-forming intrahepatic cholangiocarcinoma based on clinical, radiologic, and radiomics features. Eur Radiol 2021;31:8638-48. [Crossref] [PubMed]
4. Yang Y, Zou X, Zhou W, et al. DWI-based radiomic signature: potential role for individualized adjuvant chemotherapy in intrahepatic cholangiocarcinoma after partial hepatectomy. Insights Imaging 2022;13:37. [Crossref] [PubMed]