MIAMI – Researchers have found a way to predict lung cancer risk using an artificial intelligence, or AI tool.
Dr. Tiago Machuca, a thoracic surgeon with Jackson Memorial Hospital, said researchers used low dose chest CT scans to develop and train the deep learning tool called “Sybil.”
The AI model had a 94 percent accuracy of correctly classifying people with or without lung cancer within one year of screening, and up to 81 percent accuracy within six years of screening.
“I think it’s going to be extremely helpful to identify those patients that have a positive finding on screening, however, the finding is extremely low risk based on this algorithm and, on the other hand, I think what is another important development is that on this study, even those CT scans that didn’t have any nodules, they were able to have some prediction power over the course of six years if that patient was going to develop lung cancer or not,” he said. “So I think on one hand identifying those patients that are low risk but have a positive finding, and those that are at high risk even though their CT scan six years before didn’t show anything, I think is going to be extremely interesting.”
Machuca said the hope is that by reducing the rate of “false positives,” the AI model could lead to fewer biopsies and follow-ups.
He added that it’s vital to improve the lung cancer screening rate because it’s the second most common cancer among both men and women and the deadliest cancer, taking more lives than prostate, breast and colon cancer combined.
Also in today’s health news, a new study is shedding light on the impact stress has on our brains.
According to research published in Jama Open Network, people with high stress levels are 37 percent more likely to experience lower cognitive function.
Stress was found to affect memory, concentration and the ability to learn new things.
Black participants in the study reported higher levels of stress overall .
The link between stress and cognitive function was relatively consistent across all age groups.