Lung cancer screening results in the National Cancer Institute New York Study

Abstract

BACKGROUND Anecdotal reports of radiologically occult early stage lung carcinomas detected by sputum cytology suggested that screening by cytology might lead to earlier diagnosis, more effective surgical therapy, and lower death rates from lung carcinoma. Thus, a randomized study was undertaken to evaluate sputum cytology as a lung carcinoma screening technique supplementing the chest X-ray. METHODS Three major medical centers participated in the study, recruiting approximately 10,000 cigarette smoking men older than 45 years of age at each center: Memorial Sloan-Kettering Cancer Center (MSKCC) in New York, NY, Johns Hopkins Hospital in Baltimore, MD, and The Mayo Clinic in Rochester, MN. At MSKCC, the men were divided randomly into two groups: a dual screen group received four monthly sputum cytology examinations in addition to annual posteroanterior and lateral chest X-rays and an X-ray only group received annual chest X-rays but no sputum examinations. The men suspected of lung carcinoma because of radiologic or sputum cytology findings were referred immediately for evaluation, and those with operable lung carcinoma were recommended for surgery and treated with intent to cure. RESULTS The men who entered into the study remained in the screening program for 5–8 years, depending on their date of enrollment, and were followed for 2 years after screening. Follow-up was completed on more than 99%. There were 53 of the 10,040 men in the study who were found to have lung carcinoma on initial examination (prevalence): 23 were in the X-ray only group; of 30 found in the dual screen group, 9 (all with squamous cell carcinoma) were detected by cytology alone. During the entire study and the 2-year follow-up period, 354 of the 10,040 men developed lung carcinoma, equally divided between the dual screen and X-ray only groups. Nearly two-thirds (190 men) had lesions that were detected by screening, and over 50% (100 men) were in Stage I. Excluding oat cell carcinoma, during the screening period 175 of 250 carcinomas (70%) were detected by screening. In contrast, during the 2-year post-screening period, 61 lung carcinomas were diagnosed of which only 12 (20%) were Stage I. Chest X-ray was most effective in detecting peripheral adenocarcinomas of the lung, which were the most common cell type. Cytology was most effective in detecting early epidermoid carcinomas of major bronchi. The epidermoid carcinomas grew slowly, metastasized late, and after becoming visible by X-ray could be treated equally effectively as in the earlier occult stage. Forty percent of all the lung carcinomas were detected in Stage I, and at least two-thirds of the patients with Stage I lung carcinoma treated by complete resection did not die of their disease. Overall 5-year survival of all patients with lung carcinoma who had enrolled in the detection program was 35%, compared with 13% for the United States as a whole during this same time period. CONCLUSIONS Sputum cytology and the chest X-ray complemented each other as lung carcinoma detection techniques. The chest X-ray best detected peripheral adenocarcinomas of the lung, which are the most common type of lung carcinoma. Sputum cytology detected epidermoid carcinomas arising in major bronchi, but these are slow growing tumors that can be resected and cured after becoming visible by chest X-ray. Thus, for subjects at risk of lung carcinoma who could be followed by annual chest X-rays, sputum cytology did not improve survival, but for high risk subjects who had only a single screening examination, sputum cytology increased the number of early lung carcinomas detected. The design of the current study did not permit evaluation of chest X-ray screening versus nonscreening for prevention of death from lung carcinoma. However, the large proportion of Stage I lung carcinomas and the high survival rate of patients in this study compared with Surveillance, Epidemiology, and End Results program data strongly suggested that screening for lung carcinoma in high risk populations is a valuable public health measure. Cancer 2000;89:2356–62. © 2000 American Cancer Society.