Automated Reading May Improve the Safety of Triennial Lung Cancer Screening Intervals for Low-Risk Subjects

R.E. Ledda et al. examined the feasibility of extending lung cancer screening intervals in low-risk individuals using artificial intelligence (AI) software, AVIEW, for automated reading of baseline low-dose computed tomography (LDCT) scans. The study re-evaluated 3,450 participants from the BioMILD trial with negative baseline LDCT scans. Automated AI classified 12.3% as LungRADS 3 and 5% as LungRADS 4. Three lung cancers were identified in these reclassified groups (0.5%), all detected 25–28 months after baseline. Although stage and type of detected cancers suggested minimal clinical benefits from shorter intervals, some LungRADS 4 cases might have benefited from earlier follow-ups. The study concluded that AI-driven LDCT interpretation reduces false negatives and supports personalized screening intervals for low-risk groups, improving efficiency and resource allocation. AVIEW demonstrated its potential as a complementary tool to radiologists, optimizing lung cancer screening by safely extending intervals for certain populations