+++
CHAPTER SUMMARY AND CENTRAL ILLUSTRATION
++
Chapter Summary
This chapter provides an overview of the current and potential applications for artificial intelligence (AI) in cardiovascular care. AI techniques can mimic human thought processes, learning capacity, and knowledge storage. Current AI applications do not have independent reasoning or abstraction; instead they are focused on specific tasks such as automation, risk prediction, and pattern recognition. Most current AI techniques involve some form of machine learning methods. AI holds substantial promise for cardiovascular care, given the variety of ever-growing digital data becoming available. Because pattern recognition is a particularly good application for AI, cardiovascular imaging is a high-yield area for its use. The use of AI in risk prediction has also grown rapidly, as has the interpretation of digital health data (eg, from wearable or nonwearable biosensors). However, the evidence base for AI solutions for cardiovascular care so far has been limited, and deployment of AI in routine cardiovascular clinical practice has been minimal to date. Current challenges must be successfully navigated before wide adoption of AI takes place in our field, with careful development and adoption in clinical practice (see Fuster and Hurst’s Central Illustration). With this approach, AI has the potential to advance the concepts of both precision health and population health management. Ultimately, AI is best developed and deployed as “augmented intelligence”—that is, to enhance human intelligence—in support of cardiovascular clinicians and health systems in achieving higher quality of care and improved health outcomes.
++
++
Artificial intelligence (AI) applications in health care carry great promise. If successfully developed and deployed, AI has the potential to improve efficiency and reduce waste in health-care delivery; provide novel insights on patient risk and disease trajectory, improving diagnosis and treatment decisions; and improve patient health engagement and outcomes.1–7 AI can be applied to large and diverse data sets, such as omics, imaging, or electronic health records (EHRs), and in conjunction with other emerging technologies such as wearable and nonwearable biosensors, voice or optical technologies, robotics, or any combination. These represent just a partial list of potential data sources and technologies for AI in cardiovascular (CV) care.
++
Despite all of this potential, the evidence base for AI solutions for CV care has been limited to date and there is minimal deployment of AI in routine CV clinical practice. Accordingly, the goals of this chapter are to discuss why now for AI and CV care; provide an overview of several AI methods; focus on two key areas for application with specific examples reflecting a spectrum of CV conditions and modalities; and discuss key challenges and mitigation strategies for AI to fulfill its potential for CV care. We conclude with the concept of augmented intelligence as a likely success factor for AI in CV care.