Predictive Analytics Model Helps Determine Colon Cancer Risk

Regenstrief Institute researchers created a new prediction model that determines an individual’s risk for colon cancer and precancerous polyps.

The work was led by Thomas F. Imperiale, MD, Regenstrief Institute and US Department of Veterans Affairs research scientist and Indiana University School of Medicine professor.

According to researchers, the prediction tool could be beneficial in countries such as Canada, the United Kingdom, Italy, Netherlands, Australia, and others where colonoscopies are not routinely used for screening the average risk for individuals.

The new weighted models are bases on scoring a range of factors such as age, gender, smoking history, cohabitation history, nonsteroidal anti-inflammatory drug (e.g., ibuprofen) and aspirin use, and average physical activity. Then, the tools weigh the importance of each factor to determine an individual’s personalized risk.

“We need to more precisely measure the risk of individuals and populations, so good, scientifically-based decisions about when and how to be screened for colorectal cancer can be made by patients, clinicians and public health officials,” Imperiale, who is also a practicing gastroenterologist who has researched colon cancer risks and screening for two decades, said in a press release.

“Our composite, weighted scoring methods enable simple risk stratification to determine an individual’s risk, for example, of a precancerous polyp with certain ominous features or of colon cancer, both indicating a need for colonoscopy.”

Imperiale and his colleagues developed two tools. One applies to patients who have had limited endoscopy screening of a part of the colon with sigmoidoscopy. The other can be used on patients without any tests involving visualization of the colon.

For various reasons, such as expense and the need for patient preparation and anesthesia, many countries do not embrace routine screening colonoscopies and prefer using sigmoidoscopy.

Sigmoidoscopy is a more straightforward procedure that only examines the lower portion of the colon and is done every five to ten years.  According to researchers, tool samples are inexpensive, easy to use, and repeated annually or biennially.

Predictive and data-based tools for identifying risk in patients are invaluable for determining who should get a colonoscopy and who may be well-screened without a colonoscopy. The World Health Organization (WHO) indicates that in 2020, colorectal cancer was the third most common cancer in the world and the second most common cause of death.

The study, which enrolled patients between 2004 and 2011, was supported by the National Cancer Institute, Walther Cancer Foundation, and Indiana University Melvin and Bren Simon Cancer Center.