Lung Cancer Screening. A Different Paradigm

Abstract

Am J Respir Crit Care Med. 2003;168:2003. McWilliams A, Mayo J, McDonald S, et al Comments: Several recent 1 studies have shown that screening chest computed tomography (CT) has higher sensitivity than chest x-ray and it detects lung cancers at an earlier stage (Lancet 1999;354:99–105, Am J Respir Crit Care Med 2002;165:508–13 , Chest 2002;122:15–20 ). However, CT has a high false-positive rate and it cannot detect early cancers in the central airways. The current study describes a new approach combining sputum markers, thoracic CT, and autofluorescence bronchoscopy to screen for lung cancer among smokers. The study enrolled 561 volunteers between 50 and 74 years of age with more than or equal to a 30 pack-year smoking history. Sputum was induced using nebulized 3% hypertonic saline. Automated quantitative cytometry (AQC) was performed and DNA content of at least 3000 epithelial cells was measured using an investigational automated high-resolution image cytometer. Atypia was defined as the presence of ≥5 cells with a DNA index higher than 1.2. All subjects subsequently underwent a thoracic CT. Three hundred seventy-eight study subjects also underwent autofluorescence bronchoscopy using the LIFE-lung device. Biopsies were obtained from the areas of abnormal fluorescence, and the histology was graded according to World Health Organization criteria (Eur Respir J 2001;18:1059–68 ). Sputum analysis showed atypia in 423 (75%) patients. CT scans showed noncalcified lung nodules in 259 (46%) patients. Autofluorescence bronchoscopy showed dysplasia in 48% of patients with sputum atypia and in 32% of patients without sputum atypia (P = 0.05). Overall, the screening detected 14 primary lung cancers. Thoracic CT detected lung tumors in 10 participants and autofluorescence bronchoscopy detected tumors in the remaining 4 subjects. Thus, 29% of subjects with tumor had normal CT findings. Thirteen of 14 (93%) cancers were found in subjects with abnormal sputum findings. The majority of tumors were stage 0–II. The histologic cell-type distribution of the tumors was similar to that seen in the general population. The detection rate for CT alone in this study was 1.8%. Prescreening with sputum AQC and limiting CT for those with sputum atypia would have increased the CT detection rate to 2.1% and would have avoided 25% of initial CT scans. The addition of autofluorescence bronchoscopy would have further increased the detection rate to 3.1%. The results of this pilot trial are clearly encouraging. For the first time, the authors have shown that hybrid screening by combining sputum analysis, autofluorescence bronchoscopy, and thoracic CT increases the detection rate of early lung cancer. The screening method used in this study has the potential to reduce unnecessary CT scans. A clear benefit is a better ability to detect early-stage central tumors that cannot be detected by CT imaging. The limitations of this screening method should also be noted. The new screening paradigm did not reduce the incidence of thoracotomy for benign disease process. Similar to a prior report (Radiology 2002;222:773–81 ), 3 of 15 (20%) patients in the present study had surgery for lesions that subsequently proved to be benign. Also, lung cancer screening with spiral CT is not considered cost-effective in its present form (JAMA 2003;289:313–22 ). The addition of sputum analysis and autofluorescence bronchoscopy to the screening protocol raises further cost-related issues. More studies are required to confirm the findings from this study. Identification of better sputum markers could further improve our ability to target screening for patients at highest risk of lung cancers.