A new software tool may help doctors eliminate mistakes when judging whether a spot that turns up on a smoker's lung scan is cancerous or not, researchers say.
The clinical risk assessment method described in the New England Journal of Medicine helped correctly decide nine times out of 10 whether a spot or lesion was benign or malignant.
Computed tomography (CT) scans can save lives, but they are imperfect and can also lead to unnecessary surgery as much as 25 per cent of the time, research has shown.
"Now, we have evidence that our model and risk calculator can accurately predict which abnormalities that show up on a first CT require further follow up, such as a repeat CT scan, a biopsy, or surgery, and which ones do not," said co-principal investigator Stephen Lam.
"This is extremely good news for everyone - from the people who are high risk for developing lung cancer to the radiologists, respirologists and thoracic surgeons who detect and treat it," said Lam, a professor of medicine at the University of British Columbia.
News that makes sense
Your trusted source for staying up-to-date with the world around you. Get free daily news updates and analysis, straight to your inbox.
The prediction model includes a risk calculator that assesses age, sex, family history, emphysema, location of the nodule and other characteristics.
"Reducing the number of needless tests and increasing rapid, intensive diagnostic workups in individuals with high-risk nodules are major goals of the model," said Martin Tammemagi, an epidemiologist at Brock University who developed it.
Researchers tested the tool in a population of nearly 3000 people, including current and former smokers aged 50 to 75 who had undergone low-dose CT screening.
Researchers found that bigger nodules did not always mean cancer, and that cancers were more often found in the upper parts of the lung than the lower lobes.
The risk analysis model helped correctly determine whether the nodule was cancerous or not 94 per cent of the time, which the researchers described as "excellent predictive accuracy."
Furthermore, it helped diagnose tricky small nodules that are at most 10 millimetres in size 90 per cent of the time.
"Previous prediction models for lung nodules were hospital-based or clinic-based and showed a high prevalence of lung cancer - 23 to 75 per cent, as compared with 5.5 per cent in our study," said the article.
"Our models are coupled with risk calculators, which make possible the rapid and easy calculation of lung-cancer risk given the characteristics of the person and the nodules."
