Predicting the complexity of minimally invasive liver resection for hepatocellular carcinoma using machine learning.

IF 2.7 3区医学 Q2 GASTROENTEROLOGY & HEPATOLOGY

Hpb Pub Date : 2025-03-04 DOI:10.1016/j.hpb.2025.02.014

Giovanni Catalano, Laura Alaimo, Odysseas P Chatzipanagiotou, Andrea Ruzzenente, Francesca Ratti, Luca Aldrighetti, Hugo P Marques, François Cauchy, Vincent Lam, George A Poultsides, Tom Hugh, Irinel Popescu, Sorin Alexandrescu, Guillaume Martel, Minoru Kitago, Itaru Endo, Ana Gleisner, Feng Shen, Timothy M Pawlik

{"title":"Predicting the complexity of minimally invasive liver resection for hepatocellular carcinoma using machine learning.","authors":"Giovanni Catalano, Laura Alaimo, Odysseas P Chatzipanagiotou, Andrea Ruzzenente, Francesca Ratti, Luca Aldrighetti, Hugo P Marques, François Cauchy, Vincent Lam, George A Poultsides, Tom Hugh, Irinel Popescu, Sorin Alexandrescu, Guillaume Martel, Minoru Kitago, Itaru Endo, Ana Gleisner, Feng Shen, Timothy M Pawlik","doi":"10.1016/j.hpb.2025.02.014","DOIUrl":null,"url":null,"abstract":"Background: Despite technical advancements, minimally invasive liver surgery (MILS) for hepatocellular carcinoma (HCC) remains challenging. Nonetheless, effective tools to assess MILS complexity are still lacking. Machine learning (ML) models could improve the accuracy of such tools.Methods: Patients who underwent curative-intent MILS for HCC were identified using an international database. An XGBoost ML model was developed to predict surgical complexity using clinical and radiological characteristics.Results: Among 845 patients, 186 (22.0 %) were classified as high-risk patients. In this subgroup, median Charlson Comorbidity Index (CCI) (5.0, IQR 3.0-7.0 vs. 2.0, IQR 2.0-5.0, p < 0.001) and tumor burden score (TBS) (median 4.12, IQR 3.0-5.1 vs. 4.22, IQR 3.2-7.1, p < 0.001) were higher. The model was able to effectively predict complexity of surgery in both the training and testing cohorts with high discriminating power (ROC-AUC: 0.86, 95%CI 0.82-0.89 vs. 0.73, 95%CI 0.65-0.81). The most influential variables were CCI, TBS, BMI, extent of resection, and sex. Patients predicted to have a complex surgery were more likely to develop severe complications (OR 4.77, 95%CI 1.82-13.9, p = 0.002). An easy-to-use calculator was developed.Conclusion: Preoperative ML-prediction of complex MILS for HCC may improve preoperative planning, resource allocation, and patient outcomes.","PeriodicalId":13229,"journal":{"name":"Hpb","volume":" ","pages":""},"PeriodicalIF":2.7000,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Hpb","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1016/j.hpb.2025.02.014","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"GASTROENTEROLOGY & HEPATOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Background: Despite technical advancements, minimally invasive liver surgery (MILS) for hepatocellular carcinoma (HCC) remains challenging. Nonetheless, effective tools to assess MILS complexity are still lacking. Machine learning (ML) models could improve the accuracy of such tools.

Methods: Patients who underwent curative-intent MILS for HCC were identified using an international database. An XGBoost ML model was developed to predict surgical complexity using clinical and radiological characteristics.

Results: Among 845 patients, 186 (22.0 %) were classified as high-risk patients. In this subgroup, median Charlson Comorbidity Index (CCI) (5.0, IQR 3.0-7.0 vs. 2.0, IQR 2.0-5.0, p < 0.001) and tumor burden score (TBS) (median 4.12, IQR 3.0-5.1 vs. 4.22, IQR 3.2-7.1, p < 0.001) were higher. The model was able to effectively predict complexity of surgery in both the training and testing cohorts with high discriminating power (ROC-AUC: 0.86, 95%CI 0.82-0.89 vs. 0.73, 95%CI 0.65-0.81). The most influential variables were CCI, TBS, BMI, extent of resection, and sex. Patients predicted to have a complex surgery were more likely to develop severe complications (OR 4.77, 95%CI 1.82-13.9, p = 0.002). An easy-to-use calculator was developed.

Conclusion: Preoperative ML-prediction of complex MILS for HCC may improve preoperative planning, resource allocation, and patient outcomes.

查看原文本刊更多论文

应用机器学习预测肝癌微创肝切除术的复杂性。

背景：尽管技术进步，微创肝手术（MILS）治疗肝细胞癌（HCC）仍然具有挑战性。尽管如此，评估MILS复杂性的有效工具仍然缺乏。机器学习（ML）模型可以提高这些工具的准确性。方法：使用国际数据库确定接受HCC治疗意向mls的患者。开发了XGBoost ML模型，利用临床和放射学特征预测手术复杂性。结果：845例患者中186例（22.0%）为高危患者。在该亚组中，Charlson合并症指数中位数（CCI）为5.0，IQR为3.0-7.0 vs. 2.0， IQR为2.0-5.0,p。结论：术前预测HCC的复杂MILS可改善术前计划、资源分配和患者预后。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Hpb GASTROENTEROLOGY & HEPATOLOGY-SURGERY

CiteScore

5.60

自引率

3.40%

发文量

244

审稿时长

57 days

期刊介绍： HPB is an international forum for clinical, scientific and educational communication. Twelve issues a year bring the reader leading articles, expert reviews, original articles, images, editorials, and reader correspondence encompassing all aspects of benign and malignant hepatobiliary disease and its management. HPB features relevant aspects of clinical and translational research and practice. Specific areas of interest include HPB diseases encountered globally by clinical practitioners in this specialist field of gastrointestinal surgery. The journal addresses the challenges faced in the management of cancer involving the liver, biliary system and pancreas. While surgical oncology represents a large part of HPB practice, submission of manuscripts relating to liver and pancreas transplantation, the treatment of benign conditions such as acute and chronic pancreatitis, and those relating to hepatobiliary infection and inflammation are also welcomed. There will be a focus on developing a multidisciplinary approach to diagnosis and treatment with endoscopic and laparoscopic approaches, radiological interventions and surgical techniques being strongly represented. HPB welcomes submission of manuscripts in all these areas and in scientific focused research that has clear clinical relevance to HPB surgical practice. HPB aims to help its readers - surgeons, physicians, radiologists and basic scientists - to develop their knowledge and practice. HPB will be of interest to specialists involved in the management of hepatobiliary and pancreatic disease however will also inform those working in related fields. Abstracted and Indexed in: MEDLINE® EMBASE PubMed Science Citation Index Expanded Academic Search (EBSCO) HPB is owned by the International Hepato-Pancreato-Biliary Association (IHPBA) and is also the official Journal of the American Hepato-Pancreato-Biliary Association (AHPBA), the Asian-Pacific Hepato Pancreatic Biliary Association (A-PHPBA) and the European-African Hepato-Pancreatic Biliary Association (E-AHPBA).