New research shows that adding a deep-learning AI tool to chest CT interpretation can significantly improve radiologists’ ability to detect early-stage lung cancer while maintaining accuracy.
The study found that AI assistance increased detection sensitivity by 24.3% without significantly increasing false positives. It also made readings faster, reducing the average scan interpretation time by about 13%, from 133 seconds to roughly 116 seconds.
The biggest benefits were seen in lung cancer screening, especially for very small nodules and tumors under 10 mm that are often missed. These improvements were consistent across different types of nodules, screening situations, and levels of radiologist experience.