Analyzing a Picture to Prevent Cervical Cancer

Mark Schiffman, M.D., M.P.H., has always had a passion for social justice. It led him to become a public health doctor with the intent of helping people. “I hope things I’ve done are helping people I’ll never meet,” he says.

Mark found a challenge—and an opportunity to realize that hope—in the study of why human papillomavirus (HPV) is such a strong carcinogen. HPV infection causes cervical cancer, the fourth most common cancer among women globally.

Recently, Mark and a team of researchers developed an artificial intelligence (AI) approach to analyze images for cervical precancer. They tested the approach using more than 60,000 cervical images from an NCI cervical cancer screening trial. The trial started 25 years ago in an area of Costa Rica that has high rates of cervical cancer. The investigators systematically collected cervical images from women who volunteered to participate and continued to observe them for up to 18 years to learn about the course of cancer development.

The AI algorithm was able to positively identify precancerous changes that required medical attention to prevent cancer. “The computer algorithm was at least twice as accurate as the best doctors we showed the images to,” Mark recalls. “By identifying cervical abnormalities that would likely progress, it was predicting 6 to 7 years into the future who would develop a precancer and who wouldn’t.”

Mark is especially heartened that this AI technology can be used easily in low- and middle- resource settings in the United States and elsewhere. “This has the potential to improve the lives of many women worldwide,” he affirms.