Oct. 22, 2019

Artificial intelligence (AI) is a branch of computer science that attempts to emulate human problem-solving skills. Also called cognitive computing, AI includes concepts such as machine learning — including deep learning and natural language processing — which are especially relevant to health care.

Mayo Clinic views AI as a set of techniques and technologies that serve to augment, rather than replace, human intelligence. AI excels in discerning patterns in complex data sets that may not be easily detected using human analysis. Patterns that are nuanced and highly multifactorial and those buried in an extremely large data set are often difficult for any individual human to navigate or detect.

The value of AI in medicine comes from its ability to automate time-consuming processes, tasks that require highly tuned, but very specific skills. In some cases, the data and conclusions drawn from these processes can yield medical insights that might not otherwise be accessible.

In this Q & A, Shounak Majumder, M.D., discusses this tool's potential impact within health care and ongoing research, clinical practice and education efforts related to the field of gastroenterology and hepatology. Dr. Majumder is a gastroenterologist at Mayo Clinic's campus in Rochester, Minnesota, whose research focuses on pancreatic diseases.

How will the unique practice model at Mayo Clinic support and guide the use of AI?

Although everyone has access to AI tools, the question is what do you do with it? One of Mayo Clinic's s strength is our decades of clinical experience in caring for thousands of very complex patients, and access to what I call deep phenotyping. We know what symptoms these patients presented with, what the pathology and imaging data showed, and what the eventual outcomes were. The depth of clinical information our expert clinicians have carefully and thoughtfully recorded and our inherent strengths in innovation and application of advanced technologies are both unique to this institution. These qualities will become Mayo's strength in using AI tools.

In keeping with our institutional focus — which prioritizes the needs of the patient — our leadership recently formed an Artificial Intelligence in Gastroenterology Leadership (AGILE) team. The AGILE team, which includes Mayo Clinic gastroenterologists William A. Faubion, M.D., Cadman L. Leggett, M.D., and me, realizes that our patients are our biggest asset. We are privileged and excited to have this opportunity to learn from the data that they have generated, and we hope to use it to empower our patients and to improve their care in the future.

What preliminary goals involving AI have you outlined?

We recognize that AI is a tool that has the potential to transform health care delivery and outcomes. Working closely with a team of data scientists, analysts and disease content experts in the division, the AGILE team's goal is to identify existing knowledge and practice gaps within our specialty that can be bridged using currently available AI tools. We also hope to develop innovative AI applications that can transform health care delivery in the near future.

What potential applications do you see for AI within the field of gastroenterology and hepatology specifically?

Some of the early efforts in applying AI tools to GI will be focused on improving diagnosis and outcomes in common, yet complex GI diseases such as liver cirrhosis and inflammatory bowel disease. Other potential applications could include practice optimization for endoscopic procedures, the development of new image analysis tools, and innovation focused on early detection of pancreatic and other GI cancers.

We are also collaborating with Mayo Clinic's Center for Clinical and Translational Science to identify new pathways for incorporating AI learning modules into the training we provide for residents, fellows and junior faculty.

What are some of the challenges associated with defining a role for AI within health care?

It's important to remember that we are still in the very early stages of development in determining how best to integrate AI into health care in general. The fact that the margin for error in health care is small means that the way in which we integrate this tool into research and clinical arenas will require careful thinking from a design standpoint and a clear vision of how to use it to improve the lives of our patients.