Artificial intelligence analysis of atrial remodeling evolution in patients with atrial fibrillation: Towards optimal ablation strategies

心房颤动患者心房重塑演变的人工智能分析：寻求最佳消融策略

• 批准号: 10559270
• 负责人: Eugene Kholmovski
• 金额: $ 79.86万
• 依托单位: JOHNS HOPKINS UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-01-20 至 2027-12-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10559270
• 关键词: Ablation; American; Arrhythmia; Artificial Intelligence; Atrial Fibrillation; Cardiac ablation; Catheters; Clinical; Clinical Data; Computer Models; Development; Diagnosis; Disease; Electrophysiology (science); Environment; Evolution; Fibrosis; Health Care Costs; Heart Atrium; Image; Knowledge acquisition; Lead; Learning; Link; Longitudinal Studies; Machine Learning; Maps; Medical; Modeling; Morbidity - disease rate; Morphology; Patients; Pharmaceutical Preparations; Population; Prevention; Procedures; Psychological reinforcement; Public Health; Pulmonary veins; Recommendation; Recurrence; Refractory; Research; Role; Shapes; Structure; Testing; Time; Treatment Efficacy; cardiac magnetic resonance imaging; clinically significant; contrast enhanced; deep learning; deep reinforcement learning; economic impact; functional disability; global health; health goals; heart rhythm; improved; indexing; individual patient; insight; mortality; novel; novel strategies; patient stratification; personalized strategies; prospective; risk minimization; stem; success; treatment optimization

PROJECT SUMMARY Atrial fibrillation (AF) is the most prevalent sustained cardiac arrhythmia, leading to morbidity and mortality in 1-2% of the population and contributing significantly to global health care costs. For patients in whom AF can- not be treated by drugs, the recommended therapy is catheter-based ablation to isolate arrhythmia triggers and eliminate the substrate for arrhythmia perpetuation. The success rate of catheter ablation in rhythm controlled AF patients is 50-75%, and is worse in patients with persistent AF. The mechanisms by which baseline and post-ablation atrial remodeling, including atrial distension, functional impairment, and fibrosis, contribute to AF recurrence following catheter ablation, are not well understood and the underling factors have not been charac- terized. Understanding atrial remodeling in drug-refractory AF patients and discovering new personalized strategies for successful AF ablation and prevention of AF recurrence is a quest of paramount clinical significance. There is an urgent need to develop new approaches to ablation that account mechanistically for the remodeling of the atrial substrate post-procedure, and thereby improve the efficacy of the therapy and eliminate repeat procedures. The overall objective of this application is to use novel combination of imaging, artificial intelligence (AI), electroanatomical mapping, and mechanistic computational modeling to understand the causes for AF recurrence in drug-refractory AF patients and to develop a new paradigm for personalized ablation that eliminates repeat procedures. Leveraging our advancements in the acquisition of high-quality atrial im- ages, our expertise in AI and particularly deep learning, and our ability to efficiently generate personalized com- putational atrial models, we propose to characterize baseline atrial remodeling in shape, structure and function as well as its progression post-procedure. Using the obtained insights, we will develop a comprehensive abla- tion strategy where AF ablation targets will be determined by reinforcement learning based on the mechanistic knowledge acquired in the proposed studies. The project will culminate in a pilot prospective patient study that will test the new ablation strategy. Successful execution of the project will pave the way for a paradigm shift in the clinical procedure of AF ablation and in the quest to eliminate repeat procedures in drug-refractory AF patients, resulting in a dramatic improvement in the efficacy of the therapy. Importantly, completion of this project will be major leap forward in the integration of imaging, AI, and computational modeling in the diagnosis and treatment of heart rhythm disorders.

项目概要 心房颤动 (AF) 是最常见的持续性心律失常，导致患者发病率和死亡率 1-2% 的人口，并为全球医疗保健费用做出了重大贡献。 如果不能通过药物治疗，推荐的治疗方法是导管消融术，以隔离心律失常触发因素和 消除心律失常持续的基础 心律可控的导管消融成功率。 AF 患者为 50-75%，并且持续性 AF 患者的情况更糟。 消融后心房重塑，包括心房扩张、功能障碍和纤维化，导致 AF 导管消融后复发的问题尚未得到充分了解，并且其潜在因素尚未明确。 了解药物难治性 AF 患者的心房重塑并发现新的个性化治疗 成功房颤消融和预防房颤复发的策略是临床关键的探索 迫切需要开发新的消融方法来机械地解释这一点。 术后心房基质重塑，从而提高治疗效果并消除 重复程序。 该应用程序的总体目标是使用成像、人工智能的新颖组合 (AI)、电解剖图和机械计算模型来了解原因 药物难治性 AF 患者的 AF 复发并开发个性化消融的新范例 利用我们在获取高质量心房造影方面的进步，消除了重复手术。 年龄，我们在人工智能（特别是深度学习）方面的专业知识，以及我们有效生成个性化内容的能力 假设心房模型，我们建议在形状、结构和功能方面表征基线心房重塑 以及手术后的进展，我们将利用获得的见解制定全面的 abla- AF 消融目标将通过基于机制的强化学习来确定 在拟议研究中获得的知识将最终形成一项试点前瞻性患者研究。 将测试新的消融策略，该项目的成功执行将为范式转变铺平道路。 房颤消融的临床程序并寻求消除药物难治性房颤患者的重复手术， 重要的是，该项目的完成将显着提高治疗效果。 影像、人工智能和计算模型在诊断和治疗中融合的重大飞跃 心律失常。