A team of researchers from the Tulane University have discovered that AI can accurately diagnose Colorectal cancer. Artificial intelligence can effectively analyze tissue scans just as well or even better than pathologists.
The collaborative study combined efforts from Tulane and other universities with the objective of determining the capabilities of artificial intelligence in using it as a tool to aid pathologists. Increasing prevalence of cancer among vast population has increased the pressure on the diagnostic teams and pathologists. Researchers share that on a daily basis, pathologists evaluate and label at least thousands of histopathology images for identifying if a patient has cancer. Average workload of pathologists has drastically scaled up increasing the risk of unintended misdiagnoses.
Hong-Wen Deng, PhD, director of Biomedical Informatics Center at Tulane speaks highly of the revolutionary study. Professor Deng, also the director of Genomics at Tulane University School of Medicine added that the team has leveraged AI for Colorectal cancer diagnosis. With the help of artificial intelligence, the researchers have been able to detect Colorectal cancer in a cost-effective way. It could potentially reduce the excessive workload on pathologists.
During the study, the team examined more than 13,000 images of Colorectal cancer collected from 8,803 subjects. Researchers also gathered images from 13 independent cancer centers across Germany, the United States and China. The team used the images selected by the technicians to design a machine learning program. This program efficiently allows a computer device to recognize the images that indicate Colorectal cancer.