NMR test fails to identify cancer.

NMR test fails to identify cancer

A much-touted experimental technique cannot detect signs of cancer after all, scientists report.

In 1986, a study by Eric T. Fossel and his colleagues at Beth Israel Hospital in Boston generated excitement among researchers and physicians by suggesting that nuclear magnetic resonance (NMR) spectroscopy could identify telltale cancer clues in people's blood samples (SN: 12/6/86, p.356).

But in the April 5 NEW ENGLAND JOURNAL OF MEDICINE, two separate research teams report that NMR analysis cannot reliably distinguish between blood samples from people with cancer and those from healthy individuals. Both teams took the clear portion of the blood, or plasma, placed it in an NMR spectrometer, and studied the resulting NMR signal from the fat-containing lipoproteins.

Paul Okunieff of the Harvard Medical School in Boston led one study, while Terje Engan of Trondheim Hospital in Norway led the other. Both found that the NMR signals of people with cancer looked remarkably similar to those of healthy controls. Okunieff showed that NMR yields a false negative rate of 56 percent and a false positive rate of 52 percent.

"The NMR test for cancer has not fulfilled the great expectations that accompanied its initial description," writes Robert Shulman of the Yale University School of Medicine in an editorial accompanying the two reports. The discouraging results mean researchers must continue searching for a blood test that reliably homes in on cancer at an early stage, Okunieff told SCIENCE NEWS. Physicians could use such a test to monitor people at high risk of cancer so that early treatment could attack malignant cells before they spread, he adds.