To evaluate the impact of a specific AI-based lung nodule evaluation tool on the detection of lung nodules during low-dose chest computed tomography interpretation.
Approach:
Key Findings:
AI-assisted interpretation led to a higher detection rate of Lung-RADS-positive nodules (17% vs 10%).
Overall nodule detection was greater with AI (53% vs 33%).
The number of nodules detected per examination was also higher with AI, particularly for nodules measuring 4 to 8 mm.
Follow-up imaging recommendations were more frequent with AI (15% vs 7%).
Interpretation time was similar between AI and non-AI groups (187 vs 172 seconds).
No lung cancer diagnoses were made during a median follow-up of seven months.
Interpretation:
The use of the AI tool significantly increased the detection of clinically actionable nodules without affecting interpretation times.
Limitations:
Single-center design may limit generalizability and applicability to broader populations.
Participants were asymptomatic and many were at low risk for lung cancer, which may affect the relevance of findings.
Use of a dedicated reporting interface may not reflect routine workflows in clinical practice.
Short follow-up duration limits assessment of the long-term clinical impact of increased detection.
Conclusion:
While AI improved nodule detection rates, it did not lead to a significant reduction in interpretation time, and the clinical impact of increased detection remains uncertain.
Researchers analyzed UK Biobank data to determine whether spending sedentary time in prolonged vs interrupted bouts was associated with subsequent cancer incidence and cancer mortality.
