Classification of lung cancer patients and controls by chromatography of modified nucleosides in serum.

A wide spectrum of modified nucleosides has been quantified by high-performance liquid chromatography in serum of 49 male lung cancer patients, 35 patients with other cancers, and 48 patients hospitalized for nonneoplastic diseases. Data for 29 modified nucleoside peaks were normalized to an internal standard and analyzed by discriminant analysis and stepwise discriminant analysis. A model based on peaks selected by a stepwise discriminant procedure correctly classified 79% of the cancer and 75% of the noncancer subjects. It also demonstrated 84% sensitivity and 79% specificity when comparing lung cancer to noncancer subjects, and 80% sensitivity and 55% specificity in comparing lung cancer to other cancers. The nucleoside peaks having the greatest influence on the models varied dependent on the subgroups compared, confirming the importance of quantifying a wide array of nucleosides. These data support and expand previous studies which reported the utility of measuring modified nucleoside levels in serum and show that precise measurement of an array of 29 modified nucleosides in serum by high-performance liquid chromatography with UV scanning with subsequent data modeling may provide a clinically useful approach to patient classification in diagnosis and subsequent therapeutic monitoring.

Risk of breast cancer in relation to cigarette smoking.

To provide additional data on the smoking-breast cancer association, a case-control study of 456 cases of breast cancer and 1693 matched controls was conducted among participants in a cancer screening program. The adjusted risk of breast cancer for current smokers was 1.38 (95% confidence interval, 1.01 to 1.90). Analysis of smoking habits restricted to premenopausal women revealed a risk estimate of 2.33 (confidence interval, 1.10 to 4.96) among current smokers and increasing linear trends in risk for number of cigarettes smoked per day and for number of years of smoking. Although smokers had an earlier natural menopause than nonsmokers, there was no evidence of a protective effect of early menopause after adjustment for other factors. These findings suggest that smoking may increase the incidence of breast cancer, especially in premenopausal women.