UI Health Informatics PhD Student Eric Pahl and Hawkeye Dalton Shaull use Machine Learning to Save Lives
By Elizabeth Leake (UI3 Communications Manager)
Organ and tissue transplant candidates usually endure an excruciatingly long wait as they battle conditions that worsen with time. At some point, if they’re lucky, they receive a call which triggers a frenetic rush to beat the clock, and a swell of optimism. They enter into the process with the understanding that there’s a possibility their bodies will reject the tissue, that anti-rejection drugs will make them sicker, or they won’t survive the procedure. It’s a physical and emotional rollercoaster ride that they gladly endure for the prospect of extending time on Earth with loved ones.
According to the United Network for Organ Sharing (UNOS), there are 113,668 candidates currently waiting for organ and tissue donations, but only ~1,588 donors. The American Transplant Foundation estimates that ~20 people on the wait-list die every day. UI Health Informatics PhD Student Eric Pahl is intimately familiar with the stats; his family’s history of organ disease and failure fueled a passion for the development of smarter tech to improve access, affordability and quality of organ replacement therapies.
Dalton Shaull was a UI Hawkeye linebacker in 2013 when he suffered a career-ending traffic accident that left his left arm paralyzed. With a nerve transplant at the Mayo Clinic, he has since fully recovered.
Shared experience, passion and vision brought Pahl and Shaull together as undergrads, and in 2016, they co-founded a startup called OmniLife, Inc.
As OmniLife’s chief technical officer, Pahl uses artificial intelligence (AI) and machine learning to predict organ transplant outcomes. “Once organs become available for transplant, there are many critical decisions that must be made in a short period of time,” said Pahl. “The decision to proceed or pass can mean life or death for the prospective recipient,” he added.
They raised more than $3 million from state, federal (National Institutes of Health), and private investments. With the help of 20 student interns and ten full-time software engineers (graduates of UI and Iowa State University), they developed and deployed commercially-available software products that save lives by improving transplantation in Iowa, Illinois, Ohio, Kentucky, Utah, Indiana, and Michigan. Additional states are in queue.
A mobile communication tool called TXP ChatTM is used by transplant teams. OmniLife supplemented TXP ChatTM with Ask AlanTM, a clinical decision support chat-bot assistant that offers personalized and real-time data analysis about the organ that’s available for transplant, and predicts whether or not it’s likely to be a success.
With the help of an interdisciplinary team of mentors, including primary adviser Hans Johnson (Associate Professor of Electrical and Computer Engineering); Informaticist Nick Street (Tippie College of Business; Management Sciences), and Transplant Surgeon Alan Reed (UI Hospitals and Clinics), Pahl sustained his focus on the task of using AI to compensate for missing data, and to facilitate variables that change rapidly during chaotic clinical situations. Data that are unique to the donor, recipient, clinical team, location, time, and situation at hand are pulled from messages, users, connected APIs and file uploads in TXP ChatTM. Missing data are interpolated and provided to the surgical teams. “We use these data to assemble a customized organ offer template that is unique to each clinical professional,” said Pahl who added, “It gives them the most relevant, exact and timely information needed to streamline and shorten the decision-making process.”
As a member of the UI Interdisciplinary Graduate Program in Informatics (IGPI), Pahl and his team could access the high-performance computing resource, Argon, and appreciate the additional speed that comes with its graphics processing units (GPUs). “We utilized the nationally-acclaimed experts who are affiliated with the Iowa Informatics Initiative (UI3),” he said, and added, “the beautiful UI3 space on the fifth floor of the College of Public Health Building with dedicated workstations, fast network, lab space, meeting/collaboration rooms, and kitchenette offered our team a convenient place to meet any time of day or night,” he said.
“Argon GPUs are used to run experiments that identify which variables are more predictive of organ transplant success or failure at various post-op time intervals,” said Pahl. “We were able to halve the time from initial organ offer to transplant,” he said. Over the past three years, median offer to transplant times were 16, 19 and 21 hours, respectively. But during their intervention year, they observed a median time of 10 hours. “Fewer hours on ice are critical to organ transplant success; this time-savings was a significant improvement that wouldn’t be possible without AI,” he said.
Pahl’s team demonstrated the utility of AI/machine learning compared with the status quo (using a Cox piece-wise linear regression) for predicting organ transplant success or failure. Typically, more than one in five transplant-candidate kidneys are discarded. With AI-enabled tools, Pahl’s team observed a 26 percent increase in the number of viable transplants.
When asked what advice Pahl would give someone who wants to form a biotech startup, he said, “Make sure the PROBLEM is what you’re passionate about, not a particular solution; you’ll throw away and investigate many different solutions on your journey.” He emphasized the importance of choosing your team wisely, adding, “Either be all-in, or find someone to drive who is.” He recommends engaging with the wealth of resources and services available to help budding entrepreneurs at UI, including UI3, the John Pappajohn Entrepreneurial Center (JPEC), and the Iowa Economic Development Authority (IEDA), and reading the book, “The 7 Habits of Highly Effective People.”
As for plans upon graduating in December, Pahl intends to, “continue leading the technical direction of our company and recruit many more talented informaticists and engineers who will join our life-saving mission,” he said. “People interested in joining our company are invited to apply online at OmniLife Careers,” he added.