โก Quick Summary
This study developed a machine learning-informed nomogram for predicting 90-day graft failure in heart transplantation, utilizing data from over 25,200 patients. The model demonstrated a consistent AUC of 0.67, providing a valuable tool for optimizing donor-recipient matching and personalizing post-transplant management.
๐ Key Details
- ๐ Dataset: UNOS registry (2008-2020; n=25,200) and external validation cohort (Wuhan Union Hospital; 2018-2023; n=563)
- ๐งฉ Features used: 32 donor-recipient variables
- โ๏ธ Technology: XGBoost and Random Forest models
- ๐ Performance: AUC 0.67, HR 2.42 for high-risk group
๐ Key Takeaways
- ๐ก Machine learning was employed to identify robust predictors of 90-day graft failure.
- ๐ The nomogram incorporates eight key predictors, including recipient age and donor BMI.
- ๐ External validation confirmed the model’s generalizability across different populations.
- ๐ Explainable AI techniques were used to elucidate non-linear interactions among predictors.
- ๐ฅ The tool aims to bridge the gap between machine learning insights and clinical practice.
- ๐ High-risk patients identified by the nomogram had a 2.4-fold increased hazard of graft failure.
- ๐บ๏ธ Potential impact on addressing geographic disparities in heart transplant outcomes.
๐ Background
Heart transplantation remains a critical intervention for end-stage heart disease, yet early graft failure within the first 90 days post-surgery is a leading cause of mortality. Traditional risk scores often fall short in capturing the complex biological interactions that influence outcomes, necessitating a more sophisticated approach to donor-recipient matching.
๐๏ธ Study
This multicentre study utilized the UNOS registry to analyze data from over 25,200 heart transplant recipients. By employing advanced machine learning techniques, specifically XGBoost and Random Forest, the researchers aimed to identify key predictors of graft failure and develop a clinically interpretable nomogram that could be applied across diverse populations.
๐ Results
The final nomogram included eight significant predictors, such as recipient age, prior cardiac surgery, and donor BMI. The model achieved an AUC of 0.67, indicating moderate discrimination ability. Notably, patients classified in the high-risk group exhibited a 2.4-fold increased hazard of graft failure, underscoring the model’s potential clinical utility.
๐ Impact and Implications
The introduction of this machine learning-informed nomogram represents a significant advancement in the field of heart transplantation. By translating complex ML insights into a user-friendly tool, clinicians can enhance donor-recipient matching and tailor post-transplant management strategies. This approach not only improves patient outcomes but also addresses disparities in transplant success rates across different geographic regions.
๐ฎ Conclusion
This study highlights the transformative potential of machine learning in clinical practice, particularly in the realm of heart transplantation. The development of a nomogram for predicting 90-day graft failure bridges a crucial gap in risk assessment, paving the way for more personalized and effective patient care. Continued research and validation of such tools will be essential in optimizing transplant outcomes and improving overall healthcare delivery.
๐ฌ Your comments
What are your thoughts on the integration of machine learning in heart transplantation? We invite you to share your insights and engage in a discussion! ๐ฌ Leave your comments below or connect with us on social media:
Bridging machine learning and clinical practice: a multicentre nomogram for 90-day graft failure risk stratification in heart transplantation.
Abstract
BACKGROUND: Early graft failure within 90 postoperative days is the leading cause of mortality after heart transplantation. Existing risk scores, based on linear regression, often struggle to capture the complex, multifactorial biological interactions necessary for personalised donor-recipient matching. This study utilised explainable machine learning (ML) to identify robust predictors of 90-day graft failure and developed a clinically interpretable, ML-informed nomogram designed specifically for cross-population generalisability.
METHODS: Using the UNOS registry (2008-2020; n=25โ200), XGBoost/Random Forest models identified 90-day graft failure predictors from 32 donor-recipient variables. Explainable AI (SHapley Additive exPlanations) analysis revealed key predictors and their non-linear interactions, which were translated into a clinically applicable nomogram. External validation was performed on a large, single-centre Chinese cohort (Wuhan Union Hospital ; 2018-2023; n=563), assessing performance via area under the curve (AUC), calibration and decision curve analysis (DCA).
FINDINGS: The final model incorporated eight predictors: recipient factors (prior cardiac surgery, age, bilirubin, body mass index (BMI)), donor factors (age, gender, BMI) and cold ischaemia time. The XGBoost-derived nomogram demonstrated consistent discrimination (AUC 0.67, 95%โCI 0.64 to 0.70) and calibration. Patients stratified into the high-risk group (top quantile by nomogram score) had a 2.4-fold increased hazard of graft failure (HR 2.42, 95%โCI 2.11 to 2.78). DCA confirmed the model’s clinical utility across a wide range of risk thresholds (0.0-0.4). External validation in the Chinese cohort affirmed its generalisability (AUC 0.67).
CONCLUSION: This study introduces an ML-informed nomogram for 90-day graft failure, validated across USA and Chinese populations. By translating ML insights into a clinically interpretable tool using routinely available pretransplant variables, it bridges a key translational gap in transplant risk prediction. This tool can aid in optimising donor-recipient matching and personalising post-transplant management, with the potential to help address geographic disparities in heart transplant outcomes.
Author: [‘Yim WY’, ‘Li Y’, ‘Hou J’, ‘Chen Y’, ‘Xiong T’, ‘Li C’, ‘Lai J’, ‘Peng Y’, ‘Geng B’, ‘Wu Y’, ‘Tong F’, ‘Wang Y’, ‘Dong N’]
Journal: Open Heart
Citation: Yim WY, et al. Bridging machine learning and clinical practice: a multicentre nomogram for 90-day graft failure risk stratification in heart transplantation. Bridging machine learning and clinical practice: a multicentre nomogram for 90-day graft failure risk stratification in heart transplantation. 2026; 13:(unknown pages). doi: 10.1136/openhrt-2025-003790