For Prostate Cancer, Just Pee? Not So Fast

Before we get all excited about a potential urine test for prostate cancer, which is being reported tomorrow in the journal Nature, it’s worth remembering how littered the medical landscape is with promising early-detection tests that bombed. The biggest problem: if you want to do large-scale, population-wide screening, your test better be close to 100% specific (that is, not detect cancers that aren’t there). If you have even 1% false positives in a test that 25 million people take every year (25 million is how many men undergo PSA testing for prostate cancer), that’s 250,000 people you’re sending to have biopsies that will in most cases find nothing . . . and worrying those patients sick for no reason.

It’s way too early to tell if the test suggested by the new study would clear this hurdle. In the study, scientists led by Christopher Beecher and Arul Chinnaiyan of the University of Michigan describe how a molecule called sarcosine, which can be identified in urine, might act as an indicator of the progression and spread of prostate cancer (it doesn’t seem as promising as a tool to detect the simple presence of cancer as opposed to its invasiveness—that is, sarcosine seems to be better for prognosis, not diagnosis). The scientists compared molecules in the urine of 59 prostate cancer patients to those in the urine of 51 healthy men. They found that levels of sarcosine, a derivative of the amino acid glycine, were higher in the urine of patients with aggressive prostate cancer. And in healthy prostate cells growing in lab dishes, sarcosine made the cells extremely mobile (enough to push into a blob of gelatine), a trait characteristic of cancerous cells. That's strong evidence that the link between invasive prostate cancer and the presence of sarcosine is not just a coincidence.

The test obviously needs to be validated to determine its rate of false positives (yelling “prostate cancer” when none is present) and false negatives (failing to find prostate cancer when it is there). Still, the result is notable for being the first time a marker for prostate cancer has been detected in urine. As things now stand, prostate cancer can be detected by the PSA test or by biopsy, but the latter is invasive and sometimes painful, while the former is such a mess physicians hardly know what to make of readings any more. PSA tests do not even reduce mortality. Clearly, we need something better.

That's unlikely to be genetic tests, even though they hog the spotlight when it comes to early detection. Genetic tests give probabilistic answers. That is, except for single-gene disorders such as Huntington’s, which affect only 1 percent of the population, the answer to “does having this ‘disease gene’ mean I will get breast cancer/colon cancer/heart disease/schizophrenia . . . .” is: maybe. Carrying the mutated form of BRCA1, for instance, confers a risk of eventually developing breast cancer of 50 percent to 80 percent. A better solution would be a test that led to very early detection of disease, rather than one that shows your odds of getting something.

Enter biomarkers like sarcosine. These are, usually, proteins in the blood or urine whose presence means not that you have some probability of developing the disease but that you actually have it, now—but at a much earlier stage than x-rays or other tests can divine. I have written before about a company called Biophysical that offers a battery of tests to detect 250 such biomarkers in blood. The approach has yet to catch on, though some boutique providers offer the Biophysical test. Whether prostate cancer cells or any other kind of malignant cells betray their presence in urine or blood in a way that can--or should--find widespread clinical application remains to be seen.