Combined AI-assisted and expert-led teaching boosts radiology residents' pulmonary nodule diagnosis

Combined AI-assisted and expert-led teaching improved radiology residents' diagnostic scores for pulmonary nodules on CT more than either alone, a small RCT found. Post-test theoretical: 91.8 vs 82.6 (P<0.001); practical: 93.5 vs 81.9. Higher confidence at 1 month.

• Single-center RCT: 36 radiology residents randomized to traditional, AI-assisted, or combined teaching for pulmonary nodule assessment.
• Combined group also had longer reading times (P=0.01) and more complete feature descriptions (P<0.001); diagnostic confidence remained higher at 1 month (P=0.002).
• Exploratory study with a small sample; not externally validated, limiting generalizability.

RadPigeon summaries are original and for information only. They are not clinical advice.