AI and advanced data metrics are fake a new era of cancer research

A new project by a team of researchers across the country analyzes how digitalised health data, artificial intelligence models, and other recent technological advances have changed the way cancer is diagnosed, studied and treated.

Findings from a team led by Dr. Spyridon Bakas of Indiana University School of Medicine can be found in Cancer Research.

“Informatics and AI are used in every part of the clinical data extracted from cancer patients, from medical imaging to electronic health records, lab results, to promoting blood tests, molecular profiling and diagnosis,” says Bakas, associate professor and department chief of the Department of Pathology, Department of Pathology.

AI models are trained to identify cancer cells in tissues that humans may not be able to perceive with the naked eye, but also scan data repositories that scan relevant information at a much faster rate than human researchers.

Finally, AI can identify patterns by analyzing huge amounts of data in ways previously impossible.

However, oncologists and human researchers are still needed.

“Although AI can be used to aid in faster triaging of known cases, human oncologists are essential to approving and accessing rare cases that are not fully represented in training AI models,” Bakas said.

The article calls for the development of standardized methodologies for these informatics procedures, and develop software tools that expand databases to improve model training and enable researchers to access information.

The Bakas has been joined by almost 20 co-authors from universities and research institutes across the country, including the University of Maryland, Johns Hopkins University and Harvard University.

“Cancer informatics is a growing field with the rapid emergence of increasingly high resolution, high-dimensional, multi-dimensional datasets,” said Dr. Kathleen Noller, lead author and postdoc at the Institute of Genomic Sciences at UMD School of Medicine. “This review illuminates several areas of cutting-edge methods in cancer biology, and its shortcomings and improvements. Ultimately, these techniques will help in predicting and caring for accuracy in cancer patients.”