Machine Learning Helps Refine a New Approach for Early Colorectal Cancer Screening
In a recent study funded by NCI’s Division of Cancer Biology, researchers used machine learning (ML) to uncover a screening tool for colorectal cancer (CRC), based on a new biomarker source.
That source, called “field carcinogenesis,” measures precancerous changes in cells throughout the colon lining, a site that’s easily accessible without the need for a more invasive colonoscopy.
Read the full report, “Early Screening of Colorectal Cancer Using Feature Engineering with Artificial Intelligence-Enhanced Analysis,” in Nature.
Armed with this new biomarker source, the researchers turned to ML to help interpret nanoscale microscopic images (i.e., csPWS microscopy images). With this boost from ML technology, they could accurately capture changes in the three-dimensional structure of the genome that plays an important role in gene expression—changes that are key drivers in the development and progression of cancer.
According to corresponding author Dr. Vadim Backman of the Center for Physical Genomics at Northwestern University, “With this approach, we’re able to identify very early on if a person is harboring cells indicative of cancer development.” He added, “In the future, this type of early warning screening could be an important part of routine care, helping clinicians detect CRC and other types of cancer when it’s most treatable.”