DASI Simulations uses AI and computer vision in its technology for personalized and more accurate heart valve replacement modelling.
The 68-year-old man was admitted to the Texas hospital with serious complications from a previously implanted heart valve. Sales reps from two different heart valve manufacturers told his surgical team they couldn’t help him.
The surgeons turned to a startup — co-founded by a Georgia Tech researcher — to see if its technology could help them save his life.
DASI Simulations uses artificial intelligence (AI) and computer vision in its technology for personalized and more accurate heart valve replacement modelling. The result, the company says, is a reduction in errors and better patient outcomes, as in the case of the Texas patient.
The company’s modeling gave the Texas medical team four safe device options, as well as ways to resolve the original problem, and the patient was discharged 24 hours later.
“The doctors sent this case to our headquarters. And basically we could send them these 3D images where the doctors can see exactly what’s going to happen. Not only that, they were able to boldly go in and expand the previous valve and perform the procedure of implanting a second transcatheter heart valve in the first,” said Lakshmi Prasad Dasi, co-founder, chief technology officer and renowned scholar in heart valve engineering and cardiovascular biomechanics.
Doctors saved the man’s life, but in 2023 cardiovascular disease – the leading cause of death worldwide – will claim the lives of more than 22,000 Georgians, about 700,000 people in the US and 19 million people around the world. Even scarier, the global death toll is expected to surpass 23 million by 2030.
The company aims to improve these results by using AI to focus on the root cause of cardiovascular disease: valvular heart disease, which occurs when one of the four heart valves is damaged and impairs blood flow.
“By allowing the surgeons to run these patients through our system, the medical team can eliminate many of the complications they are concerned about. And they’ll be able to treat with a personalized plan,” said Dasi, associate chair in undergraduate studies and Rozelle Vanda Wesley professor in the Wallace H. Coulter Department of Biomedical Engineering at Emory University and Georgia Tech.
Founded in 2019, the company, which is currently in a bridging round and anticipates a Series A capital increase in the fall of 2023, has already secured approximately $4 million in investment funding, including a $600,000 non-dilutive grant. DASI Simulations has 40 customer hospitals across the US and partners with Georgia Tech’s VentureLab, which works with faculty and students to commercialize their research to secure capital investment.
The hospitals using Dasi Simulations’ technology only apply it to a group of high-risk patients. Surgeons do not want to operate on these patients because of the complexity that comes with advanced disease. Without surgical intervention, these patients die within six months to two years. Dasi Simulations technology can help surgeons master these high-risk surgeries with advanced AI-based simulations.
“The problem is that doctors are currently planning the treatment of patients, such as B. valve replacement, are extremely limited in their flexibility and adaptability to the specific anatomical features of the patient.”
All patients undergo CT scans of their hearts, followed by time-consuming measurements, said Dasi, who is also a faculty member at Georgia Tech’s Parker H. Petit Institute for Bioengineering and Bioscience. Medical imagers today use a computer mouse and use these 2D scans to measure a patient’s heart to determine which valve replacement is needed.
It’s a time-consuming and inaccurate process that’s made even more complicated now because heart valve salespeople — not the surgeons — perform the bulk of those scan measurements at most of the 800 hospitals in the US that perform structural heart procedures, Dasi said.
“So sales reps do the measurements as part of their sales service to the hospitals,” he said. “Overall, there is an increased human error of 15% to 20% variability, and a lot of time is lost when multiple people take inconsistent measurements.”
Because of this scenario, the resulting decisions about what to do with a particular patient’s case may be affected.
“The decisions that the cardiac team makes are hampered not only by the fact that measurements can be inaccurate, but also by their inability to predict risks that the patient might be exposed to with a given device choice. Consequently, there is hardly any personalization of the procedure for the patient,” said Dasi. “When these surgeons perform a surgical procedure five years later, they often find that ‘we shouldn’t have performed the surgery the way we did it.’ Now you are stuck in this difficult scenario because a different choice of device then could have made today’s surgery less risky.”
It is a scenario that leads to complications occurring at high rates and increased costs for hospitals that are not reimbursable if they occur within 30 days of patient discharge.
Risk reduction through bias elimination
DASI simulation technology, based on research at Georgia Tech, Ohio State University, Emory University and Piedmont Hospital, uses AI to create 3D models for accurate measurements based on the CT scans. It’s a process that takes a computer seconds, compared to the 30 minutes it might take a doctor or salesperson.
The company says because the 3D modeling and AI measurements are accurate, manual errors are eliminated. Because sales reps are not doing the modeling, as is currently the norm to drive sales, there is no potential for a bias towards the use of a heart valve replacement device.
The company’s technology also includes 3D predictive modeling that gives medical teams a better understanding of potential outcomes and the likelihood of complications. With the current approach, doctors make the decisions about the type of stents or valves based on the 2D scans, but they cannot necessarily predict the complications that may arise.
With DASI Simulations’ predictive technology, surgical teams don’t have to guess what the likely complications of a given valve might be and can make better decisions for the patient, Dasi said.
“The real benefit here is that they now have the science in their hands and can make data-driven scientific decisions as opposed to guesswork based on clinical intuition.”
