A group of United Kingdom-led researchers has created an artificial intelligence method that can spot mutation patterns of more than 28 types of cancer on digital pathology slides.
Researchers at the European Molecular Biology Laboratory-European Bioinformatics Institute in Cambridge, trained their algorithm from thousands of images housed in the Cancer Genome Atlas, according to the study published July 27 in Nature Cancer.
The team repurposed a Google computer vision algorithm to do the job of the histopathologist—who normally examines cancer tissue under a microscope—and the cancer geneticist, who analyzes genetic changes in cancer cells. It proved capable of predicting patient survival, as well as DNA and RNA changes based on tumor tissue images.
“Clinicians use microscopy slides for cancer diagnosis all the time,” co-author Yu Fu, a postdoctoral fellow in the Gerstung Group at EMBL-EBI, said in a statement. “However, the full potential of these slides hasn't been unlocked yet. As computer vision advances, we can analyze digital images of these slides to understand what happens at a molecular level.”
Fu et al. trained their AI with more than 17,000 images, and found it could detect 167 different cancer mutations and thousands of gene activity alterations. A separate group validated these results using a similar algorithm and eight different digital cancer slides.
Combining the molecular and histopathological data illuminates a tumor’s cancer profile, the authors noted. And such information has a variety of uses, from customizing treatments to addressing workforce shortages.
“While the number of cancer cases is increasing worldwide, the number of pathologists is declining, “ added co-author, Luiza Moore, a pathologist at the Wellcome Sanger Institute and Addenbrooke's Hospital. “At the same time, we strive to move away from the 'one size fits all' approach and into personalized medicine.
A combination of digital pathology and artificial intelligence can potentially alleviate those pressures and enhance our practice and patient care,” Moore concluded in a statement.