ideasIdea
AI malignancy scores to end repeat scans and biopsies for benign lung nodules
Most lung nodules on CT are harmless but trigger years of follow-up scans. A validated AI score could discharge low-risk nodules immediately.
Nodule AI improves discrimination over the Brock model. Propose an RCT where AI-low-risk nodules (under 1% malignancy) are discharged at baseline versus standard follow-up, with missed cancer at three years as safety and scans and biopsies avoided as efficacy.
Hypothesis
AI-based discharge removes at least 40% of follow-up CTs and 25% of benign biopsies with missed cancers under 0.5%.
Rationale
Follow-up imaging burden is a major cost and anxiety source in screening and incidental nodule management.
What would test it
Run a multicentre RCT of 5,000 nodules.
Maturity
early clinical
Who has to act
clinic
Cost to try
Medium ($1M to $50M)
Years to first evidence
4
Bottlenecks it attacks
- Overdiagnosis and false alarms · Finding more cancer is not the same as saving lives. Screening also finds cancers that would never have hurt anyone, and treats them.
- AI that is built but not validated or deployed · Thousands of cancer AI models are published; a handful are in clinical use, and fewer have shown they help patients.