- Design
- Retrospective analysis within a randomised screening trial (NELSON)
- Population
- 213 histologically confirmed screen-detected lung cancers
- Primary outcome
- Adenocarcinoma vs other subtypes by CT features
- Effect
- Pleural attachment OR 6.67 (95% CI 1.20 to 33.33); size OR 0.92 per mm; AUC 0.75
An analysis of 213 lung cancers from the NELSON screening trial (European Radiology, 7 October) related CT features to histological subtype. Two-thirds were adenocarcinomas.
Adenocarcinomas were more often early stage, smaller (14.6 vs 17.3 mm) and in the right upper lobe. On adjusted analysis, pleural attachment was associated with higher odds of adenocarcinoma (OR 6.67), larger size with lower odds, and lower and middle lobe location with lower odds. A longer volume doubling time was also associated with adenocarcinoma. The model discriminated moderately (AUC 0.75) with high sensitivity but low specificity.
This helps explain why some screen-detected nodules grow slowly yet prove malignant: slow growth is typical of adenocarcinoma, not reassurance. It does not support predicting histology from imaging, and biopsy remains the basis for subtype.
- Do not take slow growth of a screen-detected nodule as proof of benignity; adenocarcinomas tended to grow slowly.
- Note pleural attachment and upper-lobe location in reports, as both were more common in adenocarcinoma.
- Rely on tissue, not imaging features, to determine lung cancer subtype.
- Follow Lung-RADS or the local screening protocol for nodule management.
Why it matters
It reinforces that volume doubling time alone can falsely reassure when the likely cancer is an adenocarcinoma.
Don't overread it
The model's specificity was only 29%, so these features cannot predict histology in an individual patient.
The statistics, in plain English
An odds ratio of 6.67 for pleural attachment looks large, but its interval (1.20 to 33.33) is very wide because of small numbers. An AUC of 0.75 is moderate discrimination.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for radiology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free