Artificial intelligence (AI) technology is revolutionizing the entire spectrum of healthcare, from clinical specialties to administrative functions — but is still viewed with caution by cardiology departments.
During a July Becker's Hospital Review webinar sponsored by Merative, medical imaging leaders discussed the vast potential and value-add of AI for cardiology and what it will take for physicians to overcome lingering doubts about AI solutions. Panelists were:
- Steven Feinstein, MD, professor of medicine in cardiology
- Hila Goldman-Aslan, CEO and co-founder, DiA Imaging Analysis
- Rachael Terseck, global training and education manager, Circle Cardiovascular Imaging
- Marwan Sati, PhD, distinguished engineer and AI development executive, Merative
Four key takeaways were:
1. Cardiologists aren´t fully embracing AI because of concerns about its limitations and accuracy. For echocardiographers, who study the heart's left ventricular ejection fraction, questions center mainly on whether AI can quantify that metric while establishing intraobserver and interobserver consistency. For cardiovascular imaging specialists, concerns revolve around whether AI can capture measurements about flow and myocardial tissue motion. AI's level of accuracy is also a top priority for cardiovascular physicians and technicians. "The way I see AI is as a really good second opinion," Dr. Feinstein said. "Maybe as I get more comfortable with it and the software improves, I'll be more reliant upon it."
2. Despite skepticism, AI can bring measurable value to cardiology departments. Lingering doubts about AI's accuracy and applicability notwithstanding, AI technology has the potential to alleviate sonographers' workload by automating and expediting some measurements they perform manually. If AI can demonstrate reproducibility of results, it can reduce the uncertainty that derives from sonographer-to-sonographer or physician-to-physician variability, thereby making the echo lab workflow more efficient.
Another use case for AI is combing through high volumes of patient data and flagging patients whose condition or disease is not progressing as intended with treatment. "That is one good example of how AI can help in complex environments, where there's too much information for any single person to absorb," Dr. Sati said.
Dr. Sati added that AI is also expected to improve disease detection through predictive analytics that can enable physicians to intervene early and prevent patients from developing certain conditions that would lead to a diagnosis. "That [implies] huge savings in not only the person's livelihood, but in hospital costs."
3. Bringing AI into cardiology organizations can lend a competitive edge. The advantages of AI — once the technology is advanced enough to be deployed without reservations — include:
- Automating, quantifying and standardizing measurements that are currently done manually, visually or qualitatively
- Level-setting diagnostic performance
- Implementing AI into the scanning process
- Attracting younger generations of physicians who already use AI in different platforms and want similar innovation at work
- Providing the assurance some patients seek that their healthcare provider is using the latest AI technology
"Implementing AI into the scanning process, shortening scan time and freeing up scanner space is definitely competitive. From an organization perspective, you're freeing up resources for your hospital, which is key in implementing AI in the future," Ms. Terseck said.
4. When selecting an AI solution, look for certain characteristics. Given the highly sensitive nature of cardiovascular conditions and the resource demands and workflows of cardiology departments, it is important that organizations use the right criteria when evaluating and selecting AI solutions. According to the panelists, AI tools should be rapid, accurate, preferably not cloud-based due to security considerations, vendor-neutral, simple to apply, site- or device-specific (i.e., available on a sonographer's workstation) and accessible off-site.
"With these new technologies out there, we are still just scratching the surface. There is so much to do going forward," Ms. Goldman-Aslan said.
To watch the webinar on-demand, click here.