Division of Transplant, Mayo Clinic Florida, San Pablo, Jacksonville, Florida, USA.
Curr Opin Organ Transplant. 2021 Jun 1;26(3):296-301. doi: 10.1097/MOT.0000000000000875.
The field of heart transplantation is a complex practice that combines both science and art to optimize the quality and quantity of an organ transplant recipient's life span. In the current age of Transplant Medicine there are many limiting factors that prevent potentially usable organs to be transplanted in addition to the many unknown factors in assessing the risk of posttransplant complications in a proactive manner. This review focuses on the current state, and potential use, and implementation of artificial intelligence technologies in the field of heart transplantation. Furthermore, the utilization of predictive algorithms to assess donor quality, graft function, posttransplant complications and prediction of high-risk complications will be discussed. Artificial intelligence technologies in the pretransplant population is also explored.
Artificial intelligence process use has been increasing over the past decade. Early adoption in radiology and laboratory medicine have shown promise for future applications. Implementation of nascent technologies within the field of transplant medicine remains in its infancy. Cardiac and renal medicine have been recent focuses of large-scale artificial intelligence projects because of the wealth of data, the main limiting factor for producing accurate models. Understanding the true role of artificial intelligence in medicine is key - and has been divided into three areas of focus: data quality, interpretation, and clinical application. These areas allow the clinician to translate problems facing patients into algorithms utilized by data scientists to create solutions, which may provide in-depth analysis and detection of relationships not immediately clear. Although some published data has led to commercial products for cardiac, diabetic, and dermatologic applications -- widespread adoption remains limited to specialized centers.
Artificial intelligence applications with clinically relevant models in transplant medicine have the potential to optimize organ utilization, prediction of complications, and potential pretransplant management, which may mitigate the need for transplant. Further translational projects are under development at major centers, with proof of concepts demonstrating validity and safety in the clinical setting. Limiting factors of infrastructure, expertise, and data availability continue to be addressed. Ongoing efforts for commercialization and large-scale trials will provide a foundation for the development of artificial intelligence applications in transplant medicine.
心脏移植领域是一门复杂的实践学科，它将科学和艺术相结合，以优化器官移植受者的生活质量和寿命。在当今的移植医学时代，除了积极评估移植后并发症风险方面的许多未知因素外，还有许多限制因素阻止潜在可用的器官进行移植。本综述重点介绍了人工智能技术在心脏移植领域的现状、潜在用途和实施情况。此外，还讨论了利用预测算法评估供体质量、移植物功能、移植后并发症以及预测高危并发症的问题。还探讨了人工智能技术在移植前人群中的应用。
过去十年中，人工智能处理的使用量一直在增加。放射学和实验室医学的早期应用已经显示出未来应用的前景。移植医学领域新兴技术的实施仍处于起步阶段。由于数据丰富，心脏和肾脏医学是人工智能项目的最近重点，这是产生准确模型的主要限制因素。了解人工智能在医学中的真正作用是关键——并已分为三个重点关注领域：数据质量、解释和临床应用。这些领域使临床医生能够将患者面临的问题转化为数据科学家用来创建解决方案的算法，从而可以提供深入的分析和检测关系，这些关系并不立即明显。虽然一些已发表的数据已经导致了心脏、糖尿病和皮肤科应用的商业产品，但广泛采用仍然仅限于专门的中心。
具有临床相关模型的人工智能应用有可能优化器官利用、并发症预测以及潜在的移植前管理，从而可能减少对移植的需求。主要中心正在开展进一步的转化项目，概念验证在临床环境中证明了有效性和安全性。基础设施、专业知识和数据可用性等限制因素仍在得到解决。正在努力进行商业化和大规模试验，为移植医学人工智能应用的发展提供基础。
