Background: Thoracic computed tomography (CT) for lung cancer screening is sensitive for the detection of early peripheral lung cancer but is not sensitive for detecting central preinvasive and microinvasive cancer. Our hypothesis is that the use of a two-step strategy, using a sputum biomarker, may increase the detection rate of lung cancer by identifying individuals at highest risk.
Methods: We completed a pilot study of 561 volunteer current or former smokers 50 years of age or older, with a smoking history of more than or equal to 30 pack years. All subjects received induced sputum examination and low-dose thoracic CT scan and were offered autofluorescence bronchoscopy.
Results: CT detected 2408 pulmonary nodules, 80% of which were less than or equal to 4 mm in diameter. During 2-year follow-up, 95% of these nodules were stable or resolved, with only 4% showing growth at any time. A total of 28 cancers were detected in 22 subjects: 21 by CT scan and seven by autofluorescence bronchoscopy. Overall, 0.9% nodules were malignant, but growth on more than or equal to two CT scans increased the malignancy rate to 75%. The mean diameter of malignant nodules on detection was 12.8 mm (range, 3 to 36.4 mm). However, 18% of malignant nodules were less than or equal to 4 mm in diameter when first seen.
Conclusions: Multi-detector row CT scanners found multiple small nodules in most subjects screened, but most were stable over the 2-year follow-up. Persistent interval growth increases the probability of malignancy from less than 1% to 75%. One quarter of detected cancers were CT occult and only seen with autofluorescence bronchoscopy. Prescreening using a sputum biomarker improved the detection rate of lung cancer from 3 to 5%.