Featured News - Current News - Archived News - News Categories
UB spinoff QAS.AI is developing software – trained on past cases – to inform decision-making during surgical procedures
By the University at Buffalo
A company founded by University at Buffalo researchers is looking to employ artificial intelligence (AI) to improve treatment of vascular lesions in the brain, starting with intracranial aneurysms.
Called QAS.AI, the startup is developing software for the operating room. The program is trained on data from past surgeries and uses live video data analysis to forecast how likely a selected treatment method is to succeed in a new patient – all while a procedure is going on.
The goal is to provide real-time information to aid in decision-making during operations on intracranial aneurysms, which are bulges in arteries that can sometimes leak or burst. Ruptured aneurysms can lead to problems that can include stroke, coma or death.
QAS.AI has received $256,000 from the U.S. National Science Foundation’s Small Business Technology Transfer (STTR) program to support development of its software.
“We want to see patients get the best health care,” says Kelsey Sommer, QAS.AI chief operating officer and a UB Ph.D. candidate in biomedical engineering. “The main potential benefit is to reduce the rate of re-treatment. Re-treatment, involving another surgery, can come with so many complications and adds additional treatment cost. If we can improve outcomes from the initial surgery, the hope is that patients won’t have to come back in.”
“The neurosurgeon is always going to be the driver. Our AI will provide valuable data and predictions to help the surgeon make informed decisions,” says Ciprian Ionita, Ph.D., QAS.AI CEO and a UB assistant professor of biomedical engineering and neurosurgery. The software is based on research done in his lab, and UB has filed a patent application covering the technology.
In the future, the team hopes to expand the software’s capabilities to target treatments for other vascular diseases, such as atherosclerosis and ischemia, Ionita says.
AI Trained Using Outcomes from Past Cases
QAS is short for Quantitative Angiographic Systems. The team’s AI-driven software is designed to be used during surgery to detect complications such as areas of inadequate blood flow in the brain while also providing a forecast of a treatment’s possible outcome.
After surgeons place coils, stents or a combination of these devices in the region of an aneurysm, the program analyzes a variety of factors to predict the treatment’s probability of success.
The forecast is based on a patient’s current condition – encompassing details like blood flow in areas of interest – and what the AI has learned from reviewing thousands of past cases whose outcomes are known. If the software reports a high probability of failure, surgeons might opt to add additional devices before ending the procedure to improve the chance of success.
“What we are developing is an AI-based software that acts as a companion to medical imaging,” Sommer explains. “During a procedure, the surgical team conducts imaging to view the structure of blood vessels as the procedure is going on.”
“Our software uses this imaging data to make predictions during surgery,” Ionita says. “It guides you in deciding the treatment type. Are you going to use a stent or a coil, or both? The neurosurgeon has the ability to readjust the treatment in real-time, during the intervention.”
Building on Years of UB Biomedical Engineering Research
QAS.AI is located in incubator space at UB’s New York State Center of Excellence in Bioinformatics and Life Sciences. The company’s work builds on years of research by faculty and students in UB’s department of biomedical engineering, part of the UB School of Engineering and Applied Sciences and the Jacobs School of Medicine and Biomedical Sciences at UB.
Ionita, who conceived of the software, has engaged biomedical engineering students in a variety of projects to advance the technology. Most recently, Ph.D. candidate Mohammad Mahdi Shiraz Bhurwani served as the company’s lead AI scientist in summer 2021, working on software development as part of his doctoral research in the department.
UB’s relationship with QAS.AI also extends beyond biomedical engineering:
√ UB faculty are among company leaders. Ionita is CEO. Jason Davies, M.D., Ph.D., assistant professor of neurosurgery and biomedical informatics, is chief medical officer for QAS.AI. The chief financial officer is Vincent Tutino, Ph.D., a research assistant professor with a joint appointment in UB’s departments of neurosurgery, pathology and anatomical sciences, mechanical and aerospace engineering, and biomedical engineering.
√ UB neurosurgeons have partnered on R&D, providing feedback on the software and facilitating partnerships with Kaleida Health’s Gates Vascular Institute. Three are on the QAS.AI clinical advisory board: Adnan Siddiqui, M.D., Ph.D., UB professor of neurosurgery and radiology; Elad Levy, M.D., L. Nelson Hopkins, MD Chair of the department of neurosurgery and professor of radiology; and Kenneth Snyder, M.D., Ph.D., UB associate professor of neurosurgery, radiology and neurology. They and Davies are all affiliated with UB Neurosurgery, and Davies, Siddiqui and Snyder with the UB Canon Stroke and Vascular Research Center (of which Ionita is also a member).
√ QAS.AI has received funding and support through programs led by UB’s Business and Entrepreneur Partnerships (BEP) team. Ionita’s UB team was awarded $100,000 from UB’s Buffalo Innovation Accelerator Fund to develop an early version of the software, and QAS.AI will receive funding from the UB Center for Advanced Technology in Big Data and Health Sciences (UB CAT) to support additional R&D. The company also took part in the NSF Innovation Corps (I-Corps) site program at UB and a national version of the program, both of which help researchers evaluate a technology or idea for the marketplace.
“Our collaboration with UB’s Incubator @ CBLS has been extensive and extremely beneficial to us,” Sommer says. “Through mentoring programs and funding opportunities, we were able to develop a business plan and solution which addresses real clinical needs. These collaborations included: mentoring for business development, understanding the marketplace, mentoring for (grant) submission, and providing support for our IP.”