April 12, 2023 – Doctors are using artificial intelligence to spice up the fight against lung cancer.
A latest AI tool called Sybil can detect early signs of the disease years sooner than doctors would detect them through a conventional CT scan, say researchers on the Mass General Cancer Center in Boston and the Massachusetts Institute of Technology in Cambridge.
In one in all her studies, Sybil was in a position to appropriately predict lung cancer in 86 to 94% of cases inside one yr.
Lung cancer is the third most typical cancer within the United States, in keeping with the CDC. According to the American Cancer Society, it's the leading reason for cancer deaths, with over 127,000 deaths annually.
Now researchers say Sybil can analyze a CT image for abnormal growths or dangerous patterns, the researcher said. The results were published within the Journal of Clinical Oncology.
“The naked eye cannot see everything,” said Dr. Lecia Sequist, program director of the Cancer Early Detection and Diagnostics Clinic at Massachusetts General Hospital on NBC.com“The AI we developed looks at the scan in a completely different way than a human radiologist.”
Low-dose computed tomography (LDCT) has been shown to cut back lung cancer mortality by as much as 24%. However, lung cancer is usually not discovered until it's already too advanced.
“The lung cancer rate continues to rise among people who have never smoked or have not smoked for years. This suggests that there are many risk factors that contribute to lung cancer risk, some of which are currently unknown,” Sequist said in a Massachusetts General Hospital Press release.
“Instead of evaluating individual environmental or genetic risk factors, we developed a tool that can use images to examine collective biology and make predictions about cancer risk.”
The U.S. Preventive Service Task Force recommends annual LDCTs for some long-term smokers over 50 or those that have quit smoking inside 15 years, the hospital said. But fewer than 10 percent of eligible patients are screened annually.