A recent study (N. Eng. J. Med. 315 (1986), 1369), described a method of detecting malignant tumors by water-suppressed proton nuclear magnetic resonance (1 H NMR) study of plasma. We performed a similar study of the W 1/2, a mean of the full width at half height of the resonances of the methyl and methylene groups of the lipids of plasma lipoproteins which is inversely related to the spin-spin apparent relaxation time (T 2* W 1/2 values were measured at a fixed baseline width of 310 Hz. The study was prospective and blinded and comprised 182 subjects consisting of 40 controls, 68 patients with untreated malignancies, 45 with malignant tumors undergoing therapy and 29 benign tumor patients. No differences were seen between any groups that could serve as a basis for a useful clinical test. The major difficulty in the determination of W 1/2 was due to interference of metabolite protons (particularly lactate) within the lipoprotein resonance signal. Triglyceride level was seen to correlate inversely with W 1/2 within malignant patient groups. These discrepant results may be related to differing triglyceride-rich very low density lipoprotein (VLDL) levels in the patient populations of each study. We conclude that the water-suppressed 1H NMR of plasma lipoproteins is not a valid measurement for assessing malignancy. ©1988 Informa UK Ltd All rights reserved: reproduction in whole or part not permitted.