Which small round lung foci are dangerous, which are not? How can artificial intelligence (AI) help? Part 2
At last year’s RSNA in Chicago from December 1 to 6, 2019, we learned about a remedy for the problems described in Part 1.
In a study of the North American National Lung Screening Trial (NLST) under the direction of Dr. Pritam Mukherjee, CT scans of more than 1,000 patients were evaluated, half of whom developed lung cancer over time. A deep learning algorithm was sent over three consecutive CT studies of each patient. The AI learned in which image characteristics, which escape the human eye, small malignant lung round foci of about 4 mm differed over time from benign tumors. The results are impressive. Once this application is commercialized and FDA and CE-approved, we will contact the lead investigator to make this AI service available to our customers and a wide range of other medical experts.