https://www.ncbi.nlm.nih.gov/pubmed/30571349
From Machine Learning to Artificial Intelligence Applications in Cardiac Care.
Author information
- 1
- NewYork-Presbyterian, Columbia University Irving Medical Center (D.T.).
- 2
- University of Arkansas for Medical Sciences, Little Rock (C.P.).
Abstract
Artificial intelligence offers the potential for transformational advancement in cardiovascular care delivery, yet practical applications of this technology have yet to be embedded in clinical workflows and systems. Recent advances in machine learning algorithms and accessibility to big data sources have created the ability for software to solve highly specialized problems outside of health care, such as autonomous driving, speech recognition, and game playing (chess and Go), at superhuman efficiency previously not thought possible. To date, high-order cognitive problems in cardiovascular research such as differential diagnosis, treatment options, and clinical risk stratification have been difficult to address at scale with artificial intelligence. The practical application of artificial intelligence in the underlying operational processes in the delivery of cardiac care may be more amenable where adoption has great potential to fundamentally transform care delivery while maintaining the core quality and service that our patients demand. In this article, we provide an overview on how these artificial intelligence platforms can be implemented to improve the operational delivery of care for patients with cardiovascular disease.
KEYWORDS:
artificial intelligence; echocardiography; model; tomography
- PMID:
- 30571349
- DOI:
- 10.1161/CIRCULATIONAHA.118.031734
심장내과 영역도
인공지능의 엄습에 대비하려는 듯한 움직임
크게는 심장질환 진단의 영상학적 분야
그리고 입원환자 진료의 분야에서의
인공지능 적용분야에 대한 소개.
Figure. Applied artificial intelligence in cardiac care. (A) Machine learning for improved postprocessing of cardiac MRI. Image shown here is a 4-dimensional (4D) flow cardiac MRI illustrating blood flow through the heart and pulmonary arteries. Deep learning can be used to automate postprocessing of 4D flow cardiac MRI studies, normally performed manually at dedicated 3-dimensional laboratories. 4D flow cardiac MRI image courtesy of Arterys, Inc. (B) Machine learning for automated acquisition of focused cardiac ultrasound. Image shown here depicts automatic acquisition of parasternal long-axis transthoracic echocardiography view. Deep learning networks can guide untrained providers to automate acquisition and diagnosis of standard echocardiographic views in focused cardiac ultrasound. EchoGPS image courtesy of Bay Labs, Inc. (C) Machine learning for inpatient cardiac care delivery. Sample views of real-time patient status electronic board and mobile nudge care alerts. Real-time machine learning predictions around length of stay, severity of illness, discharge readiness, and care delays are used for automation/coordination of hospital operations and care pathways. Clinical team feedback to real-time alerts are used as labels to further train machine learning algorithms for improved performance. Images courtesy of Qventus, Inc.
201812_Circ_ From Machine Learning to Artificial Intelligence Applications in Cardiac Care.pdf
