Machine Learning-Based Survival Analysis for Patients Receiving Lenvatinib for Unresectable Hepatocellular Carcinoma

Chien-Hung Lu,1,* Ching-Wen Chang,2,3,* San-Chi Chen,4– 6 Wei-Yu Kao,1,2,7– 9 Wei-Yi Ting,10 Chia-Hsun Lu,11 Kai-I Chuang,12 Cheng-Fu Ni,12 Yao-Yu Hsieh,13,14 Ming-Shun Wu,7,8,15 Chien-Wei Su,4,16,17 Chun-Chao Chang,1,7,8 Chih-Horng Wu18– 20 1Division of Gastroenterology and Hepatology, Department of Internal Medicine, Taipei Medical University Hospital, Taipei, Taiwan; 2Graduate Institute of Metabolism and Obesity Sciences, Taipei Medical University, Taipei, Taiwan; 3Laboratory of Human Carcinogenesis, National Cancer Institute, Bethesda, MD, USA; 4Division of Medical Oncology, Department of Oncology, Taipei Veterans General Hospital, Taipei, Taiwan; 5School of Medicine, College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan; 6Institute of Clinical Medicine, College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan; 7Division of Gastroenterology and Hepatology, Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan; 8TMU Research Center for Digestive Medicine, Taipei Medical University, Taipei, Taiwan; 9Taipei Cancer Center, Taipei Medical University, Taipei, Taiwan; 10Department of Medical Imaging, Shuang Ho Hospital, Taipei Medical University, New Taipei City, Taiwan; 11Department of Medical Imaging, Wan Fang Hospital, Taipei Medical University, Taipei, Taiwan; 12Department of Medical Imaging, Taipei Medical University Hospital, Taipei, Taiwan; 13Division of Hematology and Oncology, Shuang Ho Hospital, Taipei Medical University, New Taipei City, Taiwan; 14Division of Hematology and Oncology, Department of Internal Medicine, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan; 15Division of Gastroenterology and Hepatology, Department of Internal Medicine, Wan Fang Hospital, Taipei Medical University, Taipei, Taiwan; 16Division of Gastroenterology and Hepatology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan; 17Division of General Medicine, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan; 18Department of Medical Imaging and Radiology, National Taiwan University Hospital and College of Medicine, Taipei, Taiwan; 19Center of Minimal-Invasive Interventional Radiology, National Taiwan University Hospital, Taipei, Taiwan; 20Hepatitis Research Center, National Taiwan University Hospital, Taipei, Taiwan*These authors contributed equally to this workCorrespondence: Chih-Horng Wu, Department of Medical Imaging and Radiology, National Taiwan University Hospital and College of Medicine, No. 7 Chung-Shan South Road, Taipei City, 100, Taiwan, Tel +886 2 23123456 ext. 62570, Email [email protected]: Lenvatinib is an effective treatment for patients with intermediate- to advanced-stage unresectable hepatocellular carcinoma (HCC). However, tumor response and survival outcomes vary widely. Traditional machine learning (ML) models have been developed to predict treatment response or survival status at discrete time points. However, an overall prediction of overall survival (OS) and progression-free survival (PFS) incorporating censored survival data is lacking. We aimed to conduct a comprehensive survival analysis of OS and PFS by using ML-based survival models.Patients and Methods: This multicenter, retrospective study included patients with unresectable HCC receiving lenvatinib across five healthcare centers. Demographic data, laboratory results, tumor characteristics, and survival outcomes were collected. Five ML-based survival models were developed and compared using Harrell’s concordance index (C-index). The predicted risk scores were used to stratify patients into low-, intermediate-, and high-risk groups and validated in the test set.Results: 205 patients were included for training and validation. Among the five ML models, the GBM-Cox model achieved the highest C-indices for both OS (0.617) and PFS (0.645) prediction. The predicted risk scores stratified the patients into low-, intermediate-, and high-risk groups for OS (median, 18.7 vs 13.6 vs 8.8 months; p = 0.004) and PFS (median, 8.2 vs 4.0 vs 3.7; p = 0.017). The most influential prognostic factors included albumin-bilirubin (ALBI) score, alanine aminotransferase, and age for OS, and macrovascular invasion, ALBI score, and alpha-fetoprotein for PFS.Conclusion: ML-based survival models successfully stratified patients into low-, intermediate-, and high-risk groups for OS and PFS. Key features included ALBI score and alanine aminotransferase for OS, and macrovascular invasion and ALBI score for PFS. These models have the potential to guide clinicians’ treatment decisions and provide prognostic evaluations. Future prospective studies with larger cohorts, as well as integration of imaging biomarkers are warranted to optimize these predictive models. Keywords: survival models, cox proportional hazards model, gradient boosting machine, random survival forest, overall survival, progression-free survival