Leveraging Artificial Intelligence as a Safety Net for Incidentally Identified Lung Nodules at a Tertiary Center
View abstract on PubMed
Summary
This summary is machine-generated.An artificial intelligence (AI) tool improved the detection of clinically significant lung nodules. The AI software ensured 30% of nodules received appropriate follow-up, preventing missed diagnoses and enabling timely care.
Area Of Science
- Radiology and Medical Imaging
- Artificial Intelligence in Healthcare
- Pulmonary Medicine
Background
- Artificial intelligence (AI) platforms offer potential for managing clinically significant lung nodules.
- This study evaluated a commercial AI natural language processing tool's impact on detecting indeterminate pulmonary nodules (IPNs) and ensuring guideline-consistent care.
Purpose Of The Study
- To assess the effectiveness of an AI natural language processing tool in identifying clinically significant indeterminate pulmonary nodules (IPNs).
- To determine the impact of AI on the provision of guideline-consistent care for IPNs.
Main Methods
- AI natural language processing algorithm processed CT scan radiology reports from a tertiary care center.
- Reports mentioning lung nodules were flagged and reviewed by a lung nodule expert.
- IPNs were classified as "appropriately followed" or "not appropriately followed" based on documented care within 2 weeks.
Main Results
- The AI processed 76,507 reports, identifying 389 IPNs.
- 70% of IPNs were appropriately followed, while 30% were not.
- AI identified 117 nodules without documented follow-up, leading to 43 additional appointments and 3 procedures.
Conclusions
- AI software brought 30% of clinically significant incidental pulmonary nodules to clinicians' attention that would have otherwise been missed.
- This AI-driven detection facilitated the initiation of appropriate follow-up care for potentially serious lung conditions.

