OBJECTIVES

Early lung cancer (LC) diagnosis is key to improve prognosis. We explored here the diagnostic performance of a trained dog to discriminate exhaled gas samples obtained from patients with and patients without LC and healthy controls.

METHODS

After appropriate training, we exposed the dog (a 3-year-old cross-breed between a Labrador Retriever and a Pitbull) to 390 samples of exhaled gas collected from 113 individuals (85 patients with LC and 28 controls, which included 11 patients without LC and 17 healthy individuals) for a total of 785 times.

RESULTS

The trained dog recognized LC in exhaled gas with a sensitivity of 0.95, a specificity of 0.98, a positive predictive value of 0.95 and a negative predictive value of 0.98. The area under the curve of the receiver-operating characteristics curve was 0.971.

CONCLUSIONS

This study shows that a well-trained dog can detect the presence of LC in exhaled gas samples with an extremely high accuracy.