Machine learning can expedite diagnosis of colorectal cancer

A machine learning tool could help physicians in predicting the onset of colorectal cancer as much as a year before the cancer progresses, according to a study published in Digestive Diseases and Sciences. 

The ColonFlag tool uses machine learning to detect colon cancer. This study evaluated the tool’s accuracy in cancer detection using blood samples and information including the patient’s gender and age.

The study included data from 900 colorectal cancer patients and 9,108 control patients. ColonFlag then analyzed patient’s blood count, gender and age to determine rick of developing colorectal cancer.

Results showed the tool was able to identify patients with a 10-fold higher risk of developing cancer at its earliest stages, from 180 to 360 days before clinical diagnosis. Additionally, the tool was more effective in identifying patients at risk of right-sided cancers.