To better identify candidates for lung cancer screening and better predict their lung cancer risk, a team of researchers from University College London in the U.K. developed a machine learning model using three predictors: data on a subject’s age, number of cigarettes smoked each day and duration of smoking.

To help build the model, the researchers used data on 216,714 smokers from the U.K. Biobank cohort and 26,616 smokers from the U.S. National Lung Screening trial.

The machine learning model used the three predictors to calculate a subject’s likelihood of both developing lung cancer and dying of lung cancer over the next five years. Researchers tested the new model on data culled from the U.S. Prostate, Lung, Colorectal and Ovarian Screening Trial.

During testing, the model was observed to predict the incidence of lung cancer with an estimated 83.9% sensitivity and lung cancer deaths with an estimated 85.5% sensitivity — reportedly outperforming currently used risk prediction formulas with its higher sensitivity.

The researchers explained: "We know that screening for those who have a high chance of developing lung cancer can save lives. With machine learning, we've been able to substantially simplify how we work out who is at high risk, presenting an approach that could be an exciting step in the direction of widespread implementation of personalized screening to detect many diseases early."

The machine learning model is detailed in the article “Assessing eligibility for lung cancer screening using parsimonious ensemble machine learning models: A development and validation study,” which appears in the journal PLoS Medicine.