A Holistic Clustering Methodology for Liver Transplantation Survival

Area of Science:

• Transplantation research
• Medical data analysis
• Patient stratification

Background:

• Liver transplants are common but patient heterogeneity pre-transplant is understudied.
• Limited research predicts post-transplant survival based on diverse patient characteristics.

Purpose of the Study:

• To identify patient clusters that predict mortality in liver transplant recipients.
• To explain population heterogeneity using a holistic approach.

Main Methods:

• Retrospective cohort study of 344 adult liver transplant recipients (2008-2014).
• Utilized comorbidity severity scores across 11 body systems, primary disease, demographics, and MELD score.
• Employed logistic regression, hierarchical clustering, Lasso-penalized regression, and Kaplan-Meier analysis.

Main Results:

• Five distinct patient clusters were identified.
• Clusters varied in health status, including circulatory problems, age, primary disease severity, and musculoskeletal/endocrine issues.
• Significant differences in mortality were observed between clusters (p < .001).

Conclusions:

• Developed a novel methodology for analyzing high-dimensional patient data in liver transplantation.
• A holistic approach, including psychosocial factors, can improve liver transplant care and research.
• Identified patient clusters are predictive of post-transplant survival.