The quest to better detect cancer has made a potentially huge strides. A study out of Yokohama, Japan, has potentially harnessed artificial intelligence to help detect colorectal cancer even before benign tumors become malignant.
Here’s how the researchers used artificial intelligence to spot cancer: The AI program observed a colorectal polyp magnified by 500 times in order to spot its variations. It then cross-referenced those variations against a database of over 30,000 images of pre-cancerous and cancerous cells that were used to train the machine-learning program. With that knowledge base, the AI was able to make a prediction in under a second. Think of it as Facebook’s face detection system, except for cancerous growths. This was the first-ever use and study of AI for this specific purpose.
The results were impressive — an accuracy of 86 percent — as the numbers were obtained by assessing patients whose colorectal polyp pathology was already diagnosed:
Overall, 306 polyps were assessed real-time by using the AI-assisted system, providing a sensitivity of 94 percent specificity of 79 percent, accuracy of 86 percent, and positive and negative predictive values of 79 percent and 93 percent respectively, in identifying neoplastic changes.
The study was led by Dr. Yuichi Mori of Showa University and presented on Sunday in Barcelona at the United European Gastroenterology conference.
This is particularly significant because colorectal cancer is the second deadliest form of cancer, right behind lung cancer. Colorectal cancer is so deadly because, in its later stages, the cancerous cells can exit the thin tissue of the colon, rectum, and intestine, and enter directly into the bloodstream. Once this occurs, cancer can spread much quicker throughout the body. Treatments typically involve surgical intervention with chemotherapy.
Growths begin in the rectum or colon as benign polyps, known as adenomas, and progress to become cancerous tumors. Adenomas are very common, especially after the age of 50.0
Early detection of colorectal cancer is crucial in boosting survival rates — the cancer is even considered “highly treatable” when discovered early enough.
Dr. Mori’s study has the potential to increase survival rates even further — and he’s looking to start using this technology as soon as possible.
“We believe these results are acceptable for clinical application and our immediate goal is to obtain regulatory approval for the diagnostic system,” he said in a statement.