When cells in the body become diseased, their signals and molecules change, sometimes long before symptoms emerge. New technology out of the University of Arizona aims to detect these metabolic changes with machine learning to hopefully catch diseases sooner and expedite the healing process.

A new startup company co-founded by two UA researchers uses artificial intelligence to identify these “fingerprints of disease,” possibly before the issue is detectable through other means. Ruslan Rafikov and Olga Rafikova, both associate professors in the UA College of Medicine, launched MetFora with help from Tech Launch Arizona, the UA office that commercializes inventions from university research.

While the research that led to MetFora originally focused on the lungs, this technology has the potential to detect diseases in a wide variety of organs.

“The idea came from our animal research on pulmonary hypertension. We found that if we induce pulmonary hypertension in animals, before they produce any physiological effects in the lungs, they are changing their metabolic profile very quickly. It can be after only three days. That gave us the idea that we check these same changes in a patient’s blood,” Rafikov said. “This is important because people usually struggle to get a diagnosis. They can spend years from the first symptoms to get a correct diagnosis. If we can detect it as soon as possible, it will be a more impactful treatment.”

While human researchers may manually monitor the molecules, Rafikov describes it as a very complex process involving more than a dozen metabolites.

“It’s so complex that in our finding, training AI is more important than finding a way to train a physician to find these tiny metabolites,” said Rafikov, who is an affiliate of the American Thoracic Society and the American Association for the Advancement of Science. “Each disease affects different organs, and within the organs, each disease can affect different types of cells. And once the cells are affected, they change their fine-tuned metabolic fluxes in and out of the cell. Using this knowledge, we think we can determine the exact problem in each organ.”

The process involves a blood draw that is then tested with mass spectrometry. However, because of the complicated nature of the test, it would not be able to be conducted in-hospital, and the blood would need to be shipped to MetFora’s lab for testing through an AI statistical analysis.

“The model for most diagnostics is some sort of kit or instrument that the hospital has, but this would not be that. The most natural way to bring it to market is through a lab developed test,” said MetFora CEO Martin Fuchs. “The test time isn’t long at all. It’s really about getting the sample to us in our lab and getting the results back. And we see that taking about two to four days.”

Fuchs was introduced to the researchers through Arizona FORGE, an office under the same umbrella as Tech Launch Arizona that helps foster entrepreneurship across the university campus.

“One of the things that caught my eye is that it has applicability to a broad range of diseases,” said Fuchs, who has launched two other technology companies. “We really see this as something everyone can get during their annual doctor’s visit. Because it’s a simple blood draw, I feel like it could become a standard of care.”

Although Tech Launch Arizona does not create the startups themselves, they help researchers through the process and work to negotiate licenses.

Doug Hockstad, assistant vice president for TLA, says the university system has 250 to 300 inventions disclosed every year. For researchers and their inventions, TLA assigns licensing managers to help them through the process of commercialization.

“They do a market analysis to see what companies are operating in a particular space, and a patent landscape analysis to see what patents are out there in the same area. That information is taken back to the inventor, and we more or less jointly come up with a plan on whether or not we’re going to target this invention to an existing company or target this to create a startup company around the technology,” Hockstad said. “Technologies coming out of universities are at a very early stage. Often, much too early for an existing company to be interested. So with a lot of technologies, the right way to go is to create a startup around it.”

Hockstad says the market analysis will sometimes discover that there is no market for the product, or that it’s too heavily patented already.

“We try not to pick and choose ‘winners.’ At the stage these are at, it’s very hard to foresee what is going to be successful or not,” Hockstad said.

TLA has helped organize more than 100 startups since 2013, with more than 19 startups in fiscal year 2020 alone. According to a report from the McGuire Center for Entrepreneurship at the UA’s Eller College of Management, startups associated with TLA have generated more than $25 million in state and local taxes and more than 5,000 new jobs.

“We do everything from so ware to curriculum to therapeutic compounds to medical devices,” Hockstad said. “It runs the gamut.”

Most recently, MetFora was one of four medical technology companies that participated in the startup competition Venture Madness in Phoenix on March 2 and 3.

“We are very grateful to TLA for the support. They helped with patents and regulatory analysis and mentoring. We’ve even had seasoned executives from area companies help us through this process,” Fuchs said. “They really provided the impetus to get the company started.”