UC-backed startup develops AI to diagnose coronary artery disease

Founders of Cincinnati-based Genexia developed artificial intelligence to diagnose coronary artery disease risk during a mammogram — providing women with an important additional preventative diagnostic health marker.

“Cardiovascular disease is the No. 1 killer of women worldwide, and 500,000 women in the U.S. — mothers, wives, daughters, sisters — die each year because cardiovascular disease is severely underdiagnosed,” said Dino Martis, Genexia’s CEO.

Genexia, founded by Martis, Anoop Sathyan and Kelly Cohen, aims to use explainable AI in mammograms for early detection of coronary artery disease to significantly reduce deaths and quality of life degradation for women. With 40 million annual mammograms performed in the U.S., the expected impact on women’s health will be significant.

There are a number of reasons why women often go undiagnosed for cardiovascular disease. The most prominent among them is that guidelines to diagnose and treat heart disease come from studies in the 1990s done almost exclusively on men, and even today only about 30% of study subjects are women.

This is especially meaningful because women’s symptoms present differently than men’s. Further diagnostics and interventional procedures are performed less frequently in women.

“Diagnostics focused on women’s unique physiology that can enable earlier detection is essential to improved outcomes,” Martis said.

Driven by their own personal experiences and a desire to improve lives, the University of Cincinnati Venture Lab-backed startup is developing explainable artificial intelligence to better diagnose coronary artery disease risk in women.

“Our explainable AI to diagnose coronary artery disease risk will immediately help a lot of women live long, fulfilling lives,” said Martis. Because this will be done as part of an annual or biennial mammogram where no additional images are needed and will fit seamlessly within existing workflows, the individual receives two preventative diagnoses from a mammogram at no [additional] cost resulting in true health equity and democratization.”

Prior to helping found Genexia, Cohen and Sathyan worked to develop explainable, trustworthy AI which has applicability in health, manufacturing, aerospace, financial technology and other industries.

Together, Cohen and Sathyan enrolled in UC's Venture Lab to learn how to turn their idea into a viable product.

“We don't have the training and understanding of what it takes to start up a business, and we don't have the experience,” said Cohen, director of AI Bio Lab at Digital Futures and the Brian H. Rowe Endowed Chair in aerospace engineering at UC’s College of Engineering and Applied Science. “Without the process we went through with the Venture Lab and Dino’s expertise in founding and scaling health care companies, we wouldn't have been able to reach this stage.”

UC’s Venture Lab provides a launch pad for entrepreneurs to turn their ideas into reality. The lab is located in the 1819 Innovation Hub in the Cincinnati Innovation District, providing a connection point to talent, support and funding.

The Venture Lab pre-accelerator program is available for free to all members of the university community. The seven-week program provides guidance for launching a startup and pairs aspiring entrepreneurs with executives for mentoring.

Through the Venture Lab, Cohen and Sathyan met Martis, one of the Venture Lab's entrepreneurs-in-residence.

Martis was drawn to Cohen and Sathyan because of their many years of award-winning work in AI, and Sathyan’s innovative and novel explainable AI based on Fuzzy Logic taught and championed by Cohen.

“The explainable AI we have created can be used in multiple applications,” Martis said. “What is relevant is what is the intent, interest and passion of the people who control that AI. What do the individuals behind that technology believe in? What motivates them?

“For us, improving health and bringing about health democratization and equity was what motivated us,” said Martis.

Not all AI is created equal. Explainable AI refers to the capability of an artificial intelligence system to provide transparent and human-friendly explanations for their decisions and actions. It encompasses attributes such as interpretability, transparency and accountability — ethical AI.

In high stake decision making, like in health care, the clinician’s ability to comprehend the AI reasoning and identification of potential bias allows for informed decision making and has a profound impact on safety and user acceptance.

“AI's ability to identify and compute large data available in mammogram images, determine relationships and patterns and do it with a very high degree of accuracy allows us to output a very accurate coronary artery disease risk diagnosis,” Martis said.

In January 2018, a year before he enrolled in the Venture Lab, Cohen suffered a heart attack while hosting an event for UC friends and alumni at a conference in Orlando, Florida. He had 100% blockage in the widow-maker artery, the heart's biggest artery.

Thanks to his wife's insistence that he go to the hospital and the hospital staff's efforts, Cohen was among the approximately 12% of individuals who survived such a heart attack outside a hospital.

Martis' mother also suffered from cardiovascular disease and the progressive degradation in her quality of life and death had a profound impact that motivates him and his team.

Genexia is collaborating with University Hospitals which is a leading research hospital and known for its innovative work in cardiovascular disease.

Genexia has received a lot of investor interest and once funding has been secured, the clinical trial will enroll 2,000 women, collecting data that will be used to train explainable AI models to identify breast density and diagnose coronary artery disease risk. Leonardo Kayat Bittencourt, associate professor at University Hospitals Cleveland Medical Center, said, “We chose to work with Genexia Health because of their team and expertise. We feel the collaboration will be a great fit for our data and people.”

Following the clinical trial, Genexia will seek Food and Drug Administration approval for its technology and has agreements in place for its use in University Hospitals’ 23 facilities. The diagnostic tool could potentially be commercially available as early as 2027.