Kim SH. Letter to the Editor: Methodological considerations for computed tomography-based radiomics models predicting early recurrence in hepatocellular carcinoma. World J Gastroenterol 2026; 32(31): 117415 [DOI: 10.3748/wjg.117415]
Corresponding Author of This Article
Seoung Hoon Kim, MD, PhD, Organ Transplantation Center, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang 10408, South Korea. kshlj@hanmail.net
Research Domain of This Article
Gastroenterology & Hepatology
Article-Type of This Article
letter
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Baishideng Publishing Group Inc, 7041 Koll Center Parkway, Suite 160, Pleasanton, CA 94566, USA
Share the Article
Kim SH. Letter to the Editor: Methodological considerations for computed tomography-based radiomics models predicting early recurrence in hepatocellular carcinoma. World J Gastroenterol 2026; 32(31): 117415 [DOI: 10.3748/wjg.117415]
World J Gastroenterol. Aug 21, 2026; 32(31): 117415 Published online Aug 21, 2026. doi: 10.3748/wjg.117415
Letter to the Editor: Methodological considerations for computed tomography-based radiomics models predicting early recurrence in hepatocellular carcinoma
Seoung Hoon Kim
Seoung Hoon Kim, Organ Transplantation Center, National Cancer Center, Goyang 10408, South Korea
Author contributions: Kim SH conceived and designed the study, collected and analyzed the data, performed the analysis and wrote the paper.
AI contribution statement: No ChatGPT, Grammarly, DeepL, or other AI-assisted writing tools were used in the preparation of this manuscript. No part of the Main Text, including the Abstract and the body of the manuscript, was generated by AI. Google Translate was used only for minor language support and translation of text drafted by the author. The scientific content, interpretation, and conclusions were entirely developed by the author. This manuscript does not contain any figures or images. Accordingly, no images were generated by AI. No AI tool was used for study design, data analysis, or interpretation of results. The author takes full responsibility for the content of the manuscript.
Conflict-of-interest statement: The author reports no relevant conflicts of interest for this article.
Corresponding author: Seoung Hoon Kim, MD, PhD, Organ Transplantation Center, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang 10408, South Korea. kshlj@hanmail.net
Received: December 8, 2025 Revised: January 22, 2026 Accepted: March 5, 2026 Published online: August 21, 2026 Processing time: 238 Days and 15.8 Hours
Abstract
Computed tomography-based radiomics is emerging as a promising approach for postoperative risk stratification in hepatocellular carcinoma. In response to the study by Qian et al published in the World Journal of Gastroenterology, which integrates multiphase computed tomography radiomics with clinical variables to predict early recurrence (ER) after hepatectomy in cirrhotic patients, I outline several methodological considerations to enhance interpretability, reproducibility, and generalizability. Clarification of the ER endpoint and alignment with time-to-event modeling frameworks would improve prognostic validity. Reporting of full preprocessing parameters, adherence to Image Biomarker Standardization Initiative recommendations, and explicit intraclass correlation coefficient thresholds are essential for reproducible radiomics pipelines. The distinction between intra-examination “delta-radiomics” and true longitudinal delta-radiomics warrants clearer terminology. Further, risks of feature-selection bias underscore the need for nested validation, while external or temporal validation would strengthen model generalizability. Finally, greater emphasis on biological interpretability - particularly regarding peritumoral enhancement, capsule integrity, and gamma-glutamyl transferase, which relate to microvascular invasion and tumor aggressiveness - could facilitate clinical translation. Taken together, these considerations aim to support robust radiomics model development and promote reproducible, clinically actionable tools for predicting ER in hepatocellular carcinoma.
Core Tip: Radiomics-based models for predicting early recurrence of hepatocellular carcinoma hold substantial clinical promise, but their utility depends on transparent reporting and methodological rigor. This letter highlights key considerations - including clear endpoint definition, adherence to Image Biomarker Standardization Initiative-standardized preprocessing, proper handling of feature selection and validation, and distinction between intra-examination and longitudinal delta-radiomics - to strengthen reproducibility and generalizability. Emphasizing biologically interpretable imaging markers and encouraging external validation can further support the translation of computed tomography-based radiomics into robust, clinically meaningful tools for postoperative hepatocellular carcinoma risk stratification.
