Data-driven optimization of therapy for heart failure

摘要 这个协作项目融合了工程学、人工智能、计算机建模、 生理学和临床心脏病学，为心力衰竭患者探索新的治疗策略。这个 Moonshot Goal是一个模拟框架，可以预测患者在治疗期间心脏将如何生长和重塑 潜在的治疗干预。一旦使用患者数据验证了该框架，就可以部署它 比较不同治疗方法的预测结果。然后，临床医生可以使用这些预测来指导 他们选择的治疗方法。 该项目旨在推进计算心脏病学，并使该领域更接近随机临床试验 它测试了接受模型优化治疗的患者是否比接受模型优化治疗的患者有更好的结果 接受标准的临床护理。 多学科研究团队由3名科学家组成(Ken Campbell博士；Jonathan Wink博士；Lik- 2名心脏病专家(Emma Birks，医学/博士；Gaurang Vaidya，医学博士)。他们的技能加在一起 范围从分子生物物理学，到计算机建模和工程，再到临床护理和心脏 辅助设备。 该计划有三个目标： 1)开发PyMyoVent作为实现压力感受器反射控制和心肌生长的试验床。 2)使用MyoFE创建和验证患者特定的双心室有限元模型 生长和功能重塑。 3)部署个性化的MyoFE模型，以预测心力衰竭患者的最佳治疗方案。 该计划是高度创新的奖励，并智能地利用收集的临床数据作为正常护理的一部分 来自肯塔基大学研究登记的100名患者。这些数据将包括 通过无线传输的压力信号来自已在心脏周围植入心脏MEMS设备的患者 肺动脉。基本贡献包括创建由控制的有限元模型 压力感受器反射，生长和适应生理信号，包括肌丝压力和细胞 能源使用。

ABSTRACT This collaborative project integrates concepts from engineering, artificial intelligence, computer modeling, physiology, and clinical cardiology to explore new therapeutic strategies for patients who have heart failure. The moonshot goal is a simulation framework that can predict how a patient's heart will grow and remodel during a potential therapeutic intervention. Once the framework has been validated with patient data, it could be deployed to compare the outcomes predicted for different treatments. A clinician could then use the predictions to guide their choice of therapy. This project seeks to advance computational cardiology and move the field closer to a randomized clinical trial that tests whether patients treated with model-optimized therapies have better outcomes than patients who received standard clinical care. The multidisciplinary research team consists of 3 scientists (Ken Campbell, PhD; Jonathan Wenk, PhD; Lik- Chuan Lee, PhD) and 2 cardiologists (Emma Birks, MD/PhD; Gaurang Vaidya, MD). Together, their skillsets range from molecular biophysics, through computer modeling and engineering, to clinical care and Ventricular Assist Devices. The plan has 3 Aims: 1) Develop PyMyoVent as a testbed for implementing baroreflex control and myocardial growth. 2) Use MyoFE to create and validate patient-specific biventricular finite element models that incorporate growth and functional remodeling. 3) Deploy personalized MyoFE models to predict optimal therapies for patients who have heart failure. The plan is highly innovative reward and makes intelligent use of clinical data collected as part of normal care from 100 patients who are enrolled in a research registry at the University of Kentucky. These data will include pressure signals transmitted wirelessly from patients who have had a CardioMEMS device inserted around their pulmonary artery. Fundamental contributions include the creation of finite element models that are controlled by a baroreflex and grow and adapt in response to physiological signals including myofilament stress and cellular energy use.