Journal of Hepatocellular Carcinoma (Mar 2024)
Machine Learning-Based Nomogram for Predicting Overall Survival in Elderly Patients with Cirrhotic Hepatocellular Carcinoma Undergoing Ablation Therapy
Abstract
Wenying Qiao,1– 5,* Shugui Sheng,1– 3,* Junnan Li,1– 3,* Ronghua Jin,1– 4 Caixia Hu5 1Beijing Key Laboratory of Emerging Infectious Diseases, Beijing Ditan Hospital, Capital Medical University, Beijing, People’s Republic of China; 2Beijing Institute of Infectious Diseases, Beijing Ditan Hospital, Capital Medical University, Beijing, People’s Republic of China; 3National Center for Infectious Diseases, Beijing Ditan Hospital, Capital Medical University, Beijing, People’s Republic of China; 4Changping Laboratory, Beijing, People’s Republic of China; 5Interventional Therapy Center for Oncology, Beijing You’an Hospital, Capital Medical University, Beijing, People’s Republic of China*These authors contributed equally to this workCorrespondence: Ronghua Jin, Beijing Key Laboratory of Emerging Infectious Diseases, BeijingDitan Hospital, Capital Medical University, Beijing, People’s Republic of China, Email [email protected] Caixia Hu, Interventional Therapy Center for Oncology, Beijing You’an Hospital, Capital Medical University, Beijing, People’s Republic of China, Email [email protected]: The aim of the study is to identify and evaluate multifaceted factors impacting the survival of elderly cirrhotic HCC patients following ablation therapy, with the goal of constructing a nomogram to predict their 3-, 5-, and 8-year overall survival (OS).Patients and Methods: A retrospective analysis was conducted on 736 elderly cirrhotic HCC patients who underwent ablation therapy between 2014 and 2022. LASSO regression, random survival forest (RSF), and multivariate Cox analyses were employed to identify independent prognostic factors for OS, followed by the development and validation of a predictive nomogram. Harrell’s concordance index (C-index), calibration plot and decision curve analysis (DCA) were used to assess the performance of the nomogram. The nomogram was finally utilized to stratify patients into low-, intermediate-, and high-risk groups, aiming to assess its efficacy in precisely discerning individuals with diverse overall survival outcomes.Results: Alcohol drinking, tumor number, globulin (Glob) and prealbumin (Palb) were identified and integrated to establish a novel prognostic nomogram. The nomogram exhibited strong discriminative ability with C-indices of 0.723 (training cohort) and 0.693 (validation cohort), along with significant Area Under the Curve (AUC) values for 3-year, 5-year, and 8-year OS in both cohorts (0.758, 0.770, and 0.811 for training cohort; 0.744, 0.699 and 0.737 for validation cohort). Calibration plots substantiated its consistency, while DCA curves corroborated its clinical utility. The nomogram further demonstrated exceptional effectiveness in discerning distinct risk populations, highlighting its robust applicability for prognostic stratification.Conclusion: Our study successfully developed and validated a robust nomogram model based on four key clinical parameters for predicting 3-, 5- and 8-year OS among elderly cirrhotic HCC patients following ablation therapy. The nomogram exhibited a remarkable capability in identifying high-risk patients, furnishing clinicians with invaluable insights for postoperative surveillance and tailored therapeutic interventions.Keywords: hepatocellular carcinoma, HCC, nomogram, LASSO regression, random survival forest, RSF, overall survival, OS
