Water-suppressed proton nuclear magnetic resonance spectroscopy of plasma had been proposed as a technique for detecting malignant tumors although its general diagnostic value is widely contested. To assess its diagnostic value in screening for breast cancer, we collected and analyzed 108 plasma samples from healthy women and women with breast disorders, mainly adenocarcinomas. No significant differences were found between controls and patients when average methylene-methyl linewidths were compared. Significant differences, however, were observed when methylene linewidths were compared. Unfortunately, the marked overlapping of both groups greatly reduced the possible diagnostic value of the technique. Among the various biochemical parameters analyzed for each plasma sample--triglyceride, total cholesterol and HDL cholesterol concentration, altered levels of carcinoembryonic antigen, phosphohexose isomerase, 5'-nucleotidase and phosphatase alkaline in patient samples, and estrogen and progesterone receptors of tumors--only triglyceride concentrations presented a clear inverse linear correlation with methylene linewidths.