Using AI in Cardiology to Soothe the Heart
Our Take:
Artificial intelligence (Ai) can maximize the analytic value of devices that allow for continuous monitoring of cardiac patients such as wearable devices and remote patient monitoring (RPM) and have the potential to dramatically impact cardiovascular care. In the words of “Artificial intelligence in cardiology: fundamentals and applications”, “[AI] is becoming integral to the day- to-day practice of cardiology, including interventional cardiology, electrophysiology and cardiac imaging. AI not only holds great promise in cardiology by improving outcomes, increasing accessibility, and enhancing the efficiency of delivery data collected by these devices but it holds the potential to reduce morbidity and create new treatment protocols. However, it also presents challenges related to data security, regulatory compliance, and integration into existing healthcare systems. Adapting to these changes will be essential for realizing the full potential of AI in cardiology.
Key Takeaways:
6.2 million adult Americans have heart failure, with prevalence projected to increase by 46% and direct medical costs escalating to $53 billion by 2030 (CDC, Journal of Managed Care & Specialty Pharmacy)
Between 2017 and 2020, almost 128M US adults had some form of cardiovascular disease with total costs of $407.3B (American Heart Association)
Despite improvements in the treatment and incidence of heart failure the 1-year mortality rate remains approximately 30%, while the 5-year mortality rises to 40% (Circulation)
Cardiovascular disease was the underlying cause of death, accounting for almost 1M deaths in the United States in 2020 (American Heart Association)
The Problem:
Several challenges exist in the integration of artificial intelligence into the future of healthcare delivery in cardiology. First and foremost is data quality and accuracy given that the quality and accuracy of data collected by wearable devices and other digital health tools can vary widely. For example, as noted in the “Artificial intelligence and heart failure: A state-of-the-art review”, “model accuracy may be compromised if optimal image quality or accurate views are not acquired. Using integrated ECG, echocardiography, and clinical data to develop ML algorithms presents the additional challenge of con-currently processing diverse data formats."
In addition, integrating digital health solutions may require significant upfront investments in technology and infrastructure. As the aforementioned, “Artificial intelligence and heart failure: A state-of-the-art review” notes " Implementing AI algorithms in clinical practice requires a comprehensive approach that goes beyond obtaining clearance. Implementing AI algorithms in clinical practice can be costly.” Consequently, healthcare systems, particularly less well funded ones may need to consider various funding options including joint-ventures and partnerships.
Like other specialties using AI, the use of AI in cardiology will be heavily regulated due to the risk of poor or inconsistent results. As pointed out by “Artificial intelligence in cardiology: Hope for the future and power for the present”, "Another important aspect is the achievement of robust regulation and quality control of AI systems. As AI is a new and rapidly evolving innovative field, it carries significant risks if underperforming and unregulated." Moreover, ensuring that devices are compliant with existing and evolving data protection laws and medical device regulations will be a significant challenge.
The Backdrop:
The integration of digital technologies such as wearable devices, mobile apps, and remote patient monitoring with AI has enabled rapid advancements in cardiac care technologies. For example, in the article “Artificial intelligence in cardiology: Hope for the future and power for the present” the authors point out that "the Apple Heart Study showed that the utilization of smartphones was effective in identifying patients with subclinical paroxysmal [atrial fibrillation] AF. Highlighting, it detected 0.5% of patients with possibly irregular pulse, 34% of which were diagnosed with AF confirmed by ECG." Clearly the new combinations of new digital technologies and AI has created new opportunities for monitoring, diagnosing, and treating cardiac conditions.
In addition, with demographics and aging populations in many countries the incidence of cardiovascular disease is projected to increase in the coming years. For example, according to PharmaNucleus the market for congestive heart failure was valued at $21B in 2021, and is projected to reach over $36B by 2030, with a CAGR of 7% per year. Hence, as pointed out in “Artificial intelligence in cardiology: Hope for the future and power for the present”, "the incorporation of ML methodology into the field of [heart failure] HF aims the early detection of those patients most at risk of developing the disease, correct classification of patients based on their personalized risk and prompt intervention which can be beneficial for patients with improvements in morbidity and mortality via early initiation of treatment and secondary care (via shifting treatment and follow up in the community and reducing hospital admissions)."
Cardiology is and will continue to be both a global and domestic health challenge. For example, according to 2023 Heart Disease and Stroke Statistics from the American Heart Association, between 2017 and 2020, almost 128M US adults had some form of cardiovascular disease with total costs of $407.3B. Applications of AI models to these disease states could help both diagnose and treat these diseases. As noted in the “Role of Artificial Intelligence and Machine Learning in Interventional Cardiology '',”one study used [support vector machines] SVM to detect potentially life-threatening ventricular arrhythmias. Public access ECG databases were used to train, test, and validate datasets, giving a test accuracy of 96.3%, sensitivity, and specificity of 96.2%. Another investigation classified non-life-threatening ECG beats using a convolutional neural network into 5 classes (nonectopic, supraventricular ectopic, ventricular ectopic, fusion, and unknown).” Moreover, as highlighted in “A review of smart sensors coupled with Internet of Things and Artificial Intelligence approach for heart failure monitoring”, “to convince clinicians, government, and funding agencies to pay for the cost of implementation of AI algorithms, it will be important to demonstrate measurable improvements in clinical outcomes, such as reduced length of hospital stays, morbidity, and mortality rates. Moreover, demonstrating a positive return on investment, such as increased revenue or cost saving can help justify the upfront cost of implementing AI algorithms and encourage investments in implementation of technology."
Implications:
Continuous monitoring and predictive analytics enable early intervention, potentially preventing cardiac emergencies and reducing the severity of conditions including CHF and hypertension. The wealth of patient data collected through digital health tools can fuel research and innovation in cardiology, leading to the development of new therapies and diagnostic methods. For example, collecting data via “Apple’s Siri, Amazon’s Alexa, and Google Assistant. Voice is more convenient and faster than typing on keyboards. [These] AI-assisted virtual assistants can process the input of multi- modal data and present them to the cardiologist in a meaningful manner.” Moreover, by giving patients access to real-time health data they can actively participate in their care. Empowering them in this way can improve adherence to treatment plans and encourage patients to make necessary lifestyle changes. However, as outlined in “Artificial intelligence in cardiology: Hope for the future and power for the present'', “an ethical platform is required for the responsible delivery of [any] AI project. This necessitates cooperation from all the team members of the multidisciplinary team, in order to maintain a culture of responsibility and execute a governance architecture that will adopt ethical practices at every point in the innovation and implementation lifecycle.”
Related Reading:
