Mayo Clinic Unveils AI Study, New Incubator for Innovative Start-Ups

Researchers used an AI tool to detect heart problems in the voice recordings of patients, while the health system unveiled a new program for new companies looking to develop AI platforms.

Researchers at the Mayo Clinic have developed an AI tool that can screen a patient’s voice recording for evidence of coronary artery disease (CAD).

According to a study published in Mayo Clinic Proceedings, researchers led by Jaskanwal Deep Singh Sara, MD, a cardiology fellow at the Mayo Clinic, developed a smartphone app for AI voice analysis that is able to accurately predict which patients have clogged arteries, which can cause heart attacks. The study builds on earlier work that had identified vocal biomarker components in voice samples.

“Telemedicine is non-invasive, cost-effective and efficient and has become increasingly important during the pandemic,” Sara said in a press release issued by the American Cardiology Association. “We’re not suggesting that voice analysis technology would replace doctors or replace existing methods of health care delivery, but we think there’s a huge opportunity for voice technology to act as an adjunct to existing strategies. Providing a voice sample is very intuitive and even enjoyable for patients, and it could become a scalable means for us to enhance patient management.”

Researchers found that a patient with a high voice biomarker score was 2.6 times more likely to suffer major problems associated with CAD and three times more likely to show evidence of plaque buildup. It reportedly is the first study of its kind to use voice analysis technology to predict CAD outcomes.

According to the ACA press release, the Vocalis Health algorithm has been trained to analyze more than 80 features of voice recordings, including frequency, amplitude, pitch, and cadence, based on a training set of more than 10,000 voice samples collected in Israel. Researchers had identified six features that were highly correlated with CAD and combined them into a single score, expressed as a number between -1 and 1 for each individual. One-third of patients were categorized as having a high score and two-thirds had a low score.

Sara said researchers haven’t determined why certain vocal features indicate a prevalence for CAD, but they’re looking closely at the autonomic nervous system, which regulates the voice box, many parts of the cardiovascular system and other bodily functions that aren’t under conscious control.

In addition, the study focused on English speaking patients in the Midwest, using software developed in Israel. More studies will be needed to determine if the platform works in other languages and dialects, and whether the platform can be scaled to address other chronic conditions.

The news follows a report just last week that researchers at Cedars-Sinai have developed an AI tool that can analyze plaque buildup in a patient’s arteries to predict if he or she will likely have a heart attack within five years.

And it comes alongside the Mayo Clinic’s launch of an incubator for AI platforms, joining a growing number of health systems looking to develop innovative home-grown services.

Mayo Clinic Platform_Accelerate recently launched its first 20-week program, with four start-ups in the first class. The companies will work with Mayo Clinic researchers as well as experts from Google and Epic to develop, test and eventually market new technologies.

"Health tech startups are critical contributors to the cycle of innovation," John Halamka, MD, president of the health system’s Mayo Clinic Platform initiative, said in a recent press release. "We are excited to collaborate with these innovators to solve some of the most complex problems in medicine today."

With interest in AI soaring – and some critics saying the hype has far outpaced practical applications – health systems are looking to cut through the static by developing their own platforms. They’re using incubators to attract promising start-ups, then guiding the companies’ growth with the goal of fine-tuning their products before they go to market.

"We are helping participants take a crucial step in their growth trajectory by providing startups with a disciplined focus on model validation and clinical readiness to show product value," Eric Harnisch, vice president of partner programs for Mayo Clinic Platform, said in the press release. "The program is integral to our Mayo Clinic Platform mission to enable new knowledge, new solutions and new technologies that improve patients' lives worldwide."

The four startups chosen for the Mayo Clinic program are:

• Cliexa, based in Denver, which “aims to transform patient-centered data into actionable insights for people with cardiovascular conditions and multiple chronic diseases such as diabetes;”
• Quadrant Health, based in New York, which “will analyze electronic health record and patient messaging data to triage messages and predict patient harm before it occurs;”
• ScienceIO, based in Boston and New York, which “will develop tools for organizing data to help streamline care and reduce the administrative burden for physicians;” and
• Seer Medical, based in Melbourne, Australia, which “will use data to refine and test its home-based epilepsy diagnostics and management models, as well as look for digital biomarkers to predict seizures.”