Risk factors for 1-year allograft loss in pediatric heart transplant patients using machine learning: An analysis of the pediatric heart transplant society database.
machine learning
pediatric heart transplant
Journal
Pediatric transplantation
ISSN: 1399-3046
Titre abrégé: Pediatr Transplant
Pays: Denmark
ID NLM: 9802574
Informations de publication
Date de publication:
Dec 2023
Dec 2023
Historique:
revised:
25
07
2023
received:
14
04
2023
accepted:
05
09
2023
medline:
27
11
2023
pubmed:
19
9
2023
entrez:
19
9
2023
Statut:
ppublish
Résumé
Pediatric heart transplant patients are at greatest risk of allograft loss in the first year. We assessed whether machine learning could improve 1-year risk assessment using the Pediatric Heart Transplant Society database. Patients transplanted from 2010 to 2019 were included. The primary outcome was 1-year graft loss free survival. We developed a prediction model using cross-validation, by comparing Cox regression, gradient boosting, and random forests. The modeling strategy with the best discrimination and calibration was applied to fit a final prediction model. We used Shapley additive explanation (SHAP) values to perform variable selection and to estimate effect sizes and importance of individual variables when interpreting the final prediction model. Cumulative incidence of graft loss or mortality was 7.6%. Random forests had favorable discrimination and calibration compared to Cox proportional hazards with a C-statistic (95% confidence interval [CI]) of 0.74 (0.72, 0.76) versus 0.71 (0.69, 0.73), and closer alignment between predicted and observed risk. SHAP values computed using the final prediction model indicated that the diagnosis of congenital heart disease (CHD) increased 1 year predicted risk of graft loss by 1.7 (i.e., from 7.6% to 9.3%), need for mechanical circulatory support increased predicted risk by 2, and single ventricle CHD increased predicted risk by 1.9. These three predictors, respectively, were also estimated to be the most important among the 15 predictors in the final model. Risk prediction models used to facilitate patient selection for pediatric heart transplant can be improved without loss of interpretability using machine learning.
Sections du résumé
BACKGROUND
BACKGROUND
Pediatric heart transplant patients are at greatest risk of allograft loss in the first year. We assessed whether machine learning could improve 1-year risk assessment using the Pediatric Heart Transplant Society database.
METHODS
METHODS
Patients transplanted from 2010 to 2019 were included. The primary outcome was 1-year graft loss free survival. We developed a prediction model using cross-validation, by comparing Cox regression, gradient boosting, and random forests. The modeling strategy with the best discrimination and calibration was applied to fit a final prediction model. We used Shapley additive explanation (SHAP) values to perform variable selection and to estimate effect sizes and importance of individual variables when interpreting the final prediction model.
RESULTS
RESULTS
Cumulative incidence of graft loss or mortality was 7.6%. Random forests had favorable discrimination and calibration compared to Cox proportional hazards with a C-statistic (95% confidence interval [CI]) of 0.74 (0.72, 0.76) versus 0.71 (0.69, 0.73), and closer alignment between predicted and observed risk. SHAP values computed using the final prediction model indicated that the diagnosis of congenital heart disease (CHD) increased 1 year predicted risk of graft loss by 1.7 (i.e., from 7.6% to 9.3%), need for mechanical circulatory support increased predicted risk by 2, and single ventricle CHD increased predicted risk by 1.9. These three predictors, respectively, were also estimated to be the most important among the 15 predictors in the final model.
CONCLUSIONS
CONCLUSIONS
Risk prediction models used to facilitate patient selection for pediatric heart transplant can be improved without loss of interpretability using machine learning.
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
e14612Subventions
Organisme : Enduring Hearts
Organisme : Pediatric Heart Transplant Society
Informations de copyright
© 2023 Wiley Periodicals LLC.
Références
Dipchand AI, Kirk R, Mahle WT, et al. Ten year of pediatric heart transplantation: a report from the Pediatric Heart Transplant Study. Pediatr Transplant. 2013;17:99-111.
Singh TP, Almond CS, Piercey G, Gauvreau K. Current outcomes in US children with cardiomyopathy listed for heart transplantation. Circ Heart Fail. 2012;5:594-601.
Pietra BA, Kantor PF, Bartlett HL, et al. Early predictors of survival to and after heart transplantation in children with dilated cardiomyopathy. Circulation. 2013;126:1079-1086.
Kirk R, Edwards LB, Aurora P, et al. Registry of the International Society for Heart and Lung Transplantation: twelfth official pediatric heart transplantation report-2009. J Heart Lung Transplant. 2009;28:993-1006.
Schumacher KR, Almond C, Singh TP, et al. Predicting graft loss by 1 year in pediatric heart transplantation candidates: an analysis of the pediatric heart transplant study database. Circulation. 2015;131(10):890-898.
Almond CS, Gauvreau K, Canter CE, Rajagopal SK, Piercey GE, Singh TP. A risk-prediction model for in-hospital mortality after heart transplantation in US children. Am J Transplant. 2012;12(5):1240-1248.
Weiss ES, Allen JG, Arnaoutakis GJ, et al. Creation of a quantitative recipient risk index for mortality prediction after cardiac transplantation (IMPACT). Ann Thorac Surg. 2011;92:914-922.
SRTR risk adjustment model documentation: posttransplant outcomes. Scientific Registry of Transplant Recipients. Available at https://www.srtr.org/reports-tools/risk-adjustment-models-postrtransplant-outcomes
Dag A, Oztekin A, Yucel A, Bulur S, Megahed FM. Predicting heart transplantation outcomes through data analytics. Decis Supp Syst. 2017;94:42-52.
Dag A, Topuz K, Oztekin A, Bulur S, Megahed FM. A probabilistic data-driven framework for scoring the preoperative recipient-donor heart transplant survival. Decis Support Syst. 2016;86:1-12.
Delen D, Oztekin A, Kong ZJ. A machine learning-based approach to prognostic analysis of thoracic transplantations. Artif Intell Med. 2010;49:33-42.
Medved D, Nusques P, Nilsson J. Selection of an optimal feature set to predict heart transplantation outcomes. Conf Proc IEEE Eng Med Biol Soc. 2016;2016:3290-3293.
Nilsson J, Ohlsson M, Höglund P, Ekmehag B, Koul B, Andersson B. The international heart transplant survival algorithm (IHTSA): a new model to improve organ sharing and survival. PLoS ONE. 2015;10:e0118644.
Oztekin A, Delen D, Kong ZJ. Predicting the graft survival for heart-lung transplantation patients: an integrated data mining methodology. Int J Med Inform. 2009;78:e84-e96.
Yoon J, Zame WR, Banerjee A, Cadeiras M, Alaa AM, van der Schaar M. Personalized survival predictions for cardiac transplantation via tree predictors. PLoS ONE. 2017;13:e0194985.
Jeewa A, Manlhiot C, Kantor PF, Mital S, McCrindle BW, Dipchand AI. Risk factors for mortality or delisting of patients from the pediatric heart transplant waiting list. J Thorac Cardiovasc Surg. 2014;147:462-468.
Vanderlaan RD, Manlhiot C, Edwards LB, Conway J, McCrindle BW, Dipchand AI. Risk factors for specific causes of death following pediatric heart transplant: an analysis of the registry of the International Society of Heart and Lung Transplantation. Pediatr Transplant. 2015;19:896-905.
Friedman JH. Greedy function approximation: a gradient boosting machine. Ann Stat. 2001;29:1189-1232.
Breiman L, Friedman J, Stone CJ, Olshen RA. Classification and Regression Trees. CRC press; 1984.
Chen T, Guestrin C. Xgboost: a scalable tree boosting system. In: Proceedings of the 22nd ACM Sigkdd International Conference on Knowledge Discovery and Data Mining; 2016:785-794.
Breiman L. Random forests. Mach Learn. 2001;45:5-32.
Breiman L. Bagging predictors. Mach Learn. 1996;24:123-140.
Jaeger BC, Welden S, Lenoir K, et al. Accelerated and interpretable oblique random survival forests. J Comput Graph Stat. 2023;1-24. doi:10.1080/10618600.2023.2231048
Ishwaran H, Kogalur UB, Blackstone EH, Lauer MS. Random survival forests. Ann Appl Stat. 2008;2:841-860.
Jaeger BC, Long DL, Long DM, et al. Oblique random survival forests. Ann Appl Stat. 2019;13:1847-1883.
Josse J, Prost N, Scornet E, Varoquaux G. On the consistency of supervised learning with missing values. arXiv Preprint. arXiv:1902.06931; 2019.
Steyerberg EW, Vickers AJ, Cook NR, et al. Assessing the performance of prediction models: a framework for some traditional and novel measures. Epidemiology (Cambridge, Mass.). 2010;21:128.
Steyerberg EW, Vergouwe Y. Towards better clinical prediction models: seven steps for development and an ABCD for validation. Eur Heart J. 2014;35:1925-1931.
Blanche P, Dartigues J-F, Jacqmin-Gadda H. Estimating and comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Stat Med. 2013;32:5381-5397.
Austin PC, Putter H, Giardiello D, van Klaveren D. Graphical calibration curves and the integrated calibration index (ICI) for competing risk models. Diagn Progn Res. 2022;6:2.
Kattan MW, Gerds TA. The index of prediction accuracy: an intuitive measure useful for evaluating risk prediction models. Diagn Progn Res. 2018;2:7.
Molnar C. Interpretable Machine Learning. Lulu.com; 2020.
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing; 2020. https://www.R-project.org/
Jaeger B. Table.glue: Make and Apply Customized Rounding Specifications for Tables. https://github.com/bcjaeger/table.glue
Landau WM. The drake R package: a pipeline toolkit for reproducibility and high-performance computing. J Open Source Softw. 2018;3:550. doi:10.21105/joss.00550
Kuhn M, Wickham H. Tidymodels: A Collection of Packages for Modeling and Machine Learning Using tidyverse Principles. 2020 https://www.tidymodels.org
Wickham H, Averick M, Bryan J, et al. Welcome to the tidyverse. J Open Source Softw. 2019;4:1686.