Personalized decision making for coronary artery disease treatment using offline reinforcement learning

Abstract Choosing optimal revascularization strategies for patients with obstructive coronary artery disease (CAD) remains a clinical challenge. While randomized controlled trials offer population-level insights, gaps remain regarding personalized decision-making for individual patients. We applied...

Full description

Saved in:
Bibliographic Details
Main Authors: Peyman Ghasemi, Matthew Greenberg, Danielle A. Southern, Bing Li, James A. White, Joon Lee
Format: Article
Language:English
Published: Nature Portfolio 2025-02-01
Series:npj Digital Medicine
Online Access:https://doi.org/10.1038/s41746-025-01498-1
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850067171963568128
author Peyman Ghasemi
Matthew Greenberg
Danielle A. Southern
Bing Li
James A. White
Joon Lee
author_facet Peyman Ghasemi
Matthew Greenberg
Danielle A. Southern
Bing Li
James A. White
Joon Lee
author_sort Peyman Ghasemi
collection DOAJ
description Abstract Choosing optimal revascularization strategies for patients with obstructive coronary artery disease (CAD) remains a clinical challenge. While randomized controlled trials offer population-level insights, gaps remain regarding personalized decision-making for individual patients. We applied off-policy reinforcement learning (RL) to a composite data model from 41,328 unique patients with angiography-confirmed obstructive CAD. In an offline setting, we estimated optimal treatment policies and evaluated these policies using weighted importance sampling. Our findings indicate that RL-guided therapy decisions outperformed physician-based decision making, with RL policies achieving up to 32% improvement in expected rewards based on composite major cardiovascular events outcomes. Additionally, we introduced methods to ensure that RL CAD treatment policies remain compatible with locally achievable clinical practice models, presenting an interpretable RL policy with a limited number of states. Overall, this novel RL-based clinical decision support tool, RL4CAD, demonstrates potential to optimize care in patients with obstructive CAD referred for invasive coronary angiography.
format Article
id doaj-art-0aea2c773ea644df931e3735c1efd114
institution DOAJ
issn 2398-6352
language English
publishDate 2025-02-01
publisher Nature Portfolio
record_format Article
series npj Digital Medicine
spelling doaj-art-0aea2c773ea644df931e3735c1efd1142025-08-20T02:48:27ZengNature Portfolionpj Digital Medicine2398-63522025-02-018111510.1038/s41746-025-01498-1Personalized decision making for coronary artery disease treatment using offline reinforcement learningPeyman Ghasemi0Matthew Greenberg1Danielle A. Southern2Bing Li3James A. White4Joon Lee5Data Intelligence for Health Lab, Cumming School of Medicine, University of CalgaryDepartment of Mathematics and Statistics, Faculty of Science, University of CalgaryCentre for Health Informatics, Cumming School of Medicine, University of CalgaryCentre for Health Informatics, Cumming School of Medicine, University of CalgaryDepartment of Cardiac Sciences, Cumming School of Medicine, University of CalgaryData Intelligence for Health Lab, Cumming School of Medicine, University of CalgaryAbstract Choosing optimal revascularization strategies for patients with obstructive coronary artery disease (CAD) remains a clinical challenge. While randomized controlled trials offer population-level insights, gaps remain regarding personalized decision-making for individual patients. We applied off-policy reinforcement learning (RL) to a composite data model from 41,328 unique patients with angiography-confirmed obstructive CAD. In an offline setting, we estimated optimal treatment policies and evaluated these policies using weighted importance sampling. Our findings indicate that RL-guided therapy decisions outperformed physician-based decision making, with RL policies achieving up to 32% improvement in expected rewards based on composite major cardiovascular events outcomes. Additionally, we introduced methods to ensure that RL CAD treatment policies remain compatible with locally achievable clinical practice models, presenting an interpretable RL policy with a limited number of states. Overall, this novel RL-based clinical decision support tool, RL4CAD, demonstrates potential to optimize care in patients with obstructive CAD referred for invasive coronary angiography.https://doi.org/10.1038/s41746-025-01498-1
spellingShingle Peyman Ghasemi
Matthew Greenberg
Danielle A. Southern
Bing Li
James A. White
Joon Lee
Personalized decision making for coronary artery disease treatment using offline reinforcement learning
npj Digital Medicine
title Personalized decision making for coronary artery disease treatment using offline reinforcement learning
title_full Personalized decision making for coronary artery disease treatment using offline reinforcement learning
title_fullStr Personalized decision making for coronary artery disease treatment using offline reinforcement learning
title_full_unstemmed Personalized decision making for coronary artery disease treatment using offline reinforcement learning
title_short Personalized decision making for coronary artery disease treatment using offline reinforcement learning
title_sort personalized decision making for coronary artery disease treatment using offline reinforcement learning
url https://doi.org/10.1038/s41746-025-01498-1
work_keys_str_mv AT peymanghasemi personalizeddecisionmakingforcoronaryarterydiseasetreatmentusingofflinereinforcementlearning
AT matthewgreenberg personalizeddecisionmakingforcoronaryarterydiseasetreatmentusingofflinereinforcementlearning
AT danielleasouthern personalizeddecisionmakingforcoronaryarterydiseasetreatmentusingofflinereinforcementlearning
AT bingli personalizeddecisionmakingforcoronaryarterydiseasetreatmentusingofflinereinforcementlearning
AT jamesawhite personalizeddecisionmakingforcoronaryarterydiseasetreatmentusingofflinereinforcementlearning
AT joonlee personalizeddecisionmakingforcoronaryarterydiseasetreatmentusingofflinereinforcementlearning