Vascular and Cardiac Collaboration Essential…So Is AI

LONDON — Take the opportunity to assess long-term cardiac health during the preoperative assessment of patients with peripheral artery disease (PAD) and bring in cross-specialty expertise earlier. This is what speakers emphasized during a session that brought together vascular surgeons with cardiac specialists onstage at the European Society of Cardiology (ESC) Congress 2024.

They also encouraged the use of artificial intelligence (AI) to improve patient outcomes and expedite workflow by identifying and prioritizing treatment of the highest risk cardiovascular cases and enabling more precisely planned interventions that might even be digitally rehearsed.

Session co-moderator, Alison Halliday, professor of vascular surgery at the University of Oxford, Oxford, England, said, “We need to look at the patient as a whole and work out what the risks are going to be and determine the optimal order of events and expertise. So whether and when to bring in a vascular surgeon, cardiologist, radiologist, or whoever is best suited to ensure timely and effective treatment, considering both short- and longer-term outcomes.”

“Working together more will reduce morbidity and mortality, so people need to collaborate both in regular multidisciplinary teams and within the emergency setting too,” she added.

Preoperative Checks to Optimize Cardiac and Vascular Health

Victor Aboyans, preventive and vascular cardiologist at the University Hospital of Limoges, Limoges, France, spoke about the importance of optimizing cardiac disease treatment before major vascular surgery, stressing there were longer-term cardiac benefits to be had because “heart failure is frequent and underreported in patients with PAD.”

Preoperative PAD revascularization checks by the anesthesiologist provide an optimal window for an in-depth look at a patient’s longer-term cardiac health, he said. “The anesthesiologist assesses the patient’s risk of a perioperative event, but this is also a great opportunity for looking at the patient’s long-term cardiovascular health because the post-revascularization period can be punctuated by cardiovascular events.”

Most major adverse limb events occur within the first 6 months post-revascularization, but with time this diminishes, said Aboyans, explaining that “the risk of heart events or stroke continues, so the preoperative assessment is an opportunity not only to assess perioperative risk but to optimize treatment for the long run.”

He stressed that this was particularly important in patients with PAD. Many of them, he explained, are not able to put their body under much stress due to their limb issues, which limit their movement. So simply asking if they have shortness of breath or pain in the chest is not a good reflection of their health.

He added that some simple tests prior to the PAD procedure, like an ECG or biomarker measurements, may show issues related to coronary disease or elevated markers of heart failure.

“We need to remember that 10%-20% of PAD patients have some heart failure, subclinical or clinical, suggesting that involvement of the cardiologist is warranted. If we carry out these long-term cardiac risk checks earlier on and administer preventative treatment, then we can prevent rehospitalization for these conditions.”

Aboyans added that patients that require revascularization have severe vascular disease and should be closely examined by a cardiologist because they often have cardiology conditions too. This, he said, reflects the new PAD guidelines released at the ESC 2024 and will improve patient outcomes.

AI Is the Next Partner in Collaborative Cardiovascular Care

Philippe H. Kolh, cardiovascular surgeon and chief information officer at the CHU of Liège, Liège, Belgium, discussed incorporating AI into the management of aortic disease.

“In cardiovascular medicine, we need multilevel collaboration between surgeons, whether cardiac or vascular, cardiologists, radiologists, diabetologists, and podiatrists, for example,” he said, adding that “[i]t’s so important to put the concept of the heart team or of the leg team front and foremost.”

But there is another important element to collaboration, he said, turning his focus to AI, effectively the latest addition to the multidisciplinary team.

For the past 35 years, we have strongly adhered to the use of evidence-based medicine, giving high regard to randomized controlled trials (RCTs) and meta-analyses, he noted. However, a large proportion of patients are excluded from RCTs due to strict study criteria, which means drugs approved through these processes might not be optimal for every patient.

But now, natural language processing can be used to analyze a patient’s history, including their environmental, genomic, proteomic, metabolomic, and proteomic data. This could really help in clinical decision-making, he said.

Looking at aortic aneurysm as an example, Kolh explained that clinical decisions are principally based on the aortic diameter. “If the main diameter of the aneurysm reaches 50 mm in a female or 55 mm in a male, then this fits the criteria for treatment by EVAR [endovascular aneurysm repair] or open surgery. However, there are some patients who are below the 50-55 mm diameter cutoff who might actually be at risk of rupture too because of high shear stress or other conditions,” he said. “In such a case, we need other hemodynamic data to see if this particular patient with 50-55 mm aneurysm is at risk.”

“Conversely, we may also have patients who reach the 50-55 mm criteria or even 60 mm diameter, but they are not at such high risk and they can wait longer for the procedure,” said Kolh. He stressed that this is where AI can help in identifying new markers, improving medical imaging analysis, developing new predictive models, and improving diagnoses and surgical planning.

With respect to precise planning for endovascular procedures, AI can help with the fitting of a specific prosthesis through an AI-generated model of the patient’s aorta, Kolh explained.

Further, as an intraoperative assistant, AI can look at thousands of image slices of the aorta, whereas the human eye can only look at a small portion of the vessel at a time, he continued. “AI has the capacity to merge these thousands of slices to create a 3D model of the aorta, providing different views and aspects that refine the treatment of a dissection or the positioning and deployment of a prosthesis with respect to the renal arteries, aortic valve, and coronaries. It can also be a help for a trainee surgeon preoperatively through a form of digital rehearsal.”

Some endovascular surgeons are starting to use AI in this way, but in the excitement around the innovation, Kolh noted that, “Right now we need to remember AI is there to guide, and it should help the surgeon and the patients too…But it is the surgeon who remains in charge.”

Halliday, Aboyans, and Kolh had no relevant financial disclosures to declare.