How to integrate AI into cardiac imaging workflows
The growing number of artificial intelligence (AI) algorithms available for clinical use is creating new opportunities in cardiac imaging. However, hospitals need to treat the implementation of these algorithms like the launch of a new clinical service instead of just another piece of software.
Ron Blankstein, MD, director of cardiac computed tomography at Brigham and Women's Hospital and professor of medicine and radiology at Harvard Medical School, discussed the challenges of integrating AI into daily practice at SCCT2026, the Society of Cardiovascular Computed Tomography (SCCT) annual meeting in an interview with Cardiovascular Business.
Cardiology ranks second, behind radiology, for the number of U.S. Food and Drug Administration (FDA)-cleared clinical AI algorithms. This is due to both the high level of clinical evidence showing that AI can change outcomes and the specialty's ability to secure reimbursements. Two of the most mature AI applications in cardiology involve fractional flow reserve derived from CT (FFR-CT) and coronary plaque analysis. Both now have Level I CPT codes supporting reimbursement. These payment codes are rare among the vast number of FDA-cleared algorithms.
But reimbursement and clinical evidence alone do not guarantee successful implementation. Blankstein said it requires a team effort to turn that vision into a reality.
“Unlike a software that you just download and run when you want to integrate a new AI service in your hospital, that is a brand new workflow, and it's not easy to integrate,” Blankstein said.
The implementation process can involve cardiologists, radiologists, information technology teams, legal departments, compliance committees and hospital AI oversight groups. In many cases, cardiac CT data must be transmitted to an outside vendor for analysis, with the results then returned to the institution for physician review and reporting. Blankstein said hospitals need to develop a workflow that determines who orders the AI analysis, how the order is entered, how images are transferred and how results are incorporated into clinical reports.
A clinical champion is critical to coordinating the effort. That person, typically either a cardiologist or radiologist, has the responsibility of bringing stakeholders together and ensuring the technology addresses an actual clinical need.
“It's very important for any hospital that offers cardiac CT to think about what's the best way to integrate this into their existing workflows to deliver solutions that will help patient care and that will be adopted,” Blankstein said.
Before investing in an algorithm, hospitals also need to evaluate the evidence supporting it. Blankstein said that includes assessing validation against appropriate reference standards, accuracy, prognostic value, impact on clinical management and, ultimately, whether the technology improves patient outcomes.
FFR-CT and coronary plaque analysis have emerged as success stories partly because they have accumulated more robust clinical evidence than many other applications. Studies have demonstrated not only diagnostic accuracy, but also prognostic value and potential impacts on patient management. FFR-CT also has guideline support, including a Class 2A recommendation in chest pain guidance, further supporting its clinical adoption.
The prevalence of cardiovascular disease also creates a strong need for better risk assessment, Blankstein said. Despite existing risk scores and diagnostic technologies, clinicians continue to seek more precise ways to characterize and monitor coronary disease.
However, hospitals also must develop a business case. AI services carry costs related to software, vendor analysis and additional clinical workflow, meaning institutions need to determine how the service will be paid for and ensure the program is financially sustainable.
Ultimately, successful AI implementation requires more than purchasing an algorithm. Hospitals need to identify the clinical problem first, establish evidence that the technology can solve it, build an operational workflow, secure physician and administrative buy-in and designate a clinical leader to guide the process.
AI may be increasingly sophisticated, Blankstein said, but its value in cardiac imaging will depend on how effectively hospitals incorporate it into everyday patient care.
Dave Fornell has covered healthcare for more than 17 years, with a focus in cardiology and radiology. Fornell is a 5-time winner of a Jesse H. Neal Award, the most prestigious editorial honors in the field of specialized journalism. The wins included best technical content, best use of social media and best COVID-19 coverage. Fornell was also a three-time Neal finalist for best range of work by a single author. He produces more than 100 editorial videos each year, most of them interviews with key opinion leaders in medicine. He also writes technical articles, covers key trends, conducts video hospital site visits, and is very involved with social media. E-mail: [email protected]