Performance of AI system for detecting breast cancers promising in setting of large population-based mammography screening program
WEDNESDAY, June 1, 2022 (HealthDay News) — An artificial intelligence (AI) system can help detect breast cancer in a population-based mammography screening program, according to a study published online March 29 in Radiology.
Marthe Larsen, from the Cancer Registry of Norway in Oslo, and colleagues compared the performance of a commercially available AI system to routine, independent double reading with consensus as performed in a population-based mammography screening program. A total of 122,969 screening examinations from 47,877 women were included; 752 screen-detected cancers (6.1 per 1,000 examinations) and 205 interval cancers (1.7 per 1,000 examinations) comprised the data set. Each examination had an AI score between 1 and 10 (with 10 indicating high risk). The performance of the AI system was assessed using thresholds 1, 2, and 3: Threshold 1 was set to an AI score of 10; threshold 2 was set to yield a selection rate similar to the consensus rate of 8.8 percent; and threshold 3 was set to yield a selection rate similar to an average individual radiologist at 5.8 percent.
The researchers found that 86.8 percent of screen-detected cancers and 44.9 percent of interval cancers were given a score of 10 by the AI system (threshold 1). Using threshold 3, 80.1 and 30.7 percent of the screen-detected and interval cancers were selected. Favorable histopathologic characteristics were seen in screen-detected cancer with AI scores not selected using the thresholds versus those selected; for interval cancer, the opposite results were observed.
“Based on our results, we expect AI to be of great value in the interpretation of screening mammograms in the future,” a coauthor said in a statement.
One author is a member of the Siemens Healthineers Advisory Board.
Abstract/Full Text (subscription or payment may be required)
Copyright © 2022 HealthDay. All rights reserved.
