This non-randomised study of 1035 smokers (⩾20 pack years) evaluated screening using positron emission tomography (PET) in addition to spiral computed tomography (CT) to improve assessment of lesions. All patients had baseline CT and annual scans for 5 years. Calcified nodular lesions or non-calcified nodules (NCN) <5 mm were deemed benign (scans repeated at 2 years). Patients with NCN >5 mm had monthly high resolution CT scans (contrast enhanced if density was greater than 0 HU), and PET scans were performed on NCN >7 mm. Lesions were biopsied (VATS or open surgery) if they were NCN of >20 mm, had density >30 HU after contrast or a positive PET scan. Analysis of non-nodular lesions was not clarified.

Eleven lung cancers were detected at the baseline scan. Eleven of 29 PET scans were positive (eight lung cancers) with one false negative. In the second year, 11 new lung cancers were detected (six identified at baseline CT scan). Ten of 13 PET scans were true positive; there was one false positive, one false negative, and one true negative. Twenty one of 22 patients had complete resection over 2 years, 17 of which were stage I.

This study suggests that low dose CT scanning detects early lung cancer but it is difficult to conclude that PET scanning improves diagnostic accuracy. The study illustrates the uncertainty regarding analysis of small NCN, use of contrast or PET, and optimal follow up. Further studies will establish if detection of early lung cancer leads to lower disease specific mortality.