Many men eventually develop prostate cancer. Some die from it; some die with it. We fear prostate cancer, as its treatment can have identity changing effects. When one careful examines this disease, one fears the disease and the treatment.

Many experts now recommend PSA screening for men aged 50 or above. Others balk, stating that we have no concrete evidence that we are helping patients.

An article in today’s NY Times refers to a new study in today’s NEJM. A Study Questions Blood-Test Results on Prostate Cancer . It summarizes the controversy as follows:

Significant numbers of older men whose results on a popular screening test for prostate cancer are normal may nonetheless have cancer, a new study has found.

The result, medical experts say, raises questions about what a normal test score should be and whether these men are better off let alone or treated when, through biopsies, cancer cells are discovered.

It also amplifies a controversy over the test, known as the P.S.A., and whether finding prostate cancer early and treating it by removing or destroying the prostate is, on balance, helping or harming men.

The P.S.A. test is a blood test that looks for prostate specific antigen, a protein released by prostate cells. When the prostate gland enlarges, whether because of cancer or benign conditions, P.S.A. levels in the blood tend to rise.

The test was initially used to look for recurrences of cancer after men had been treated. But in the 1990’s it came into widespread use as a screening test to find new cancers.

In that context, cancer experts informally agreed upon a convention: When a P.S.A. test finds more than four nanograms of the protein in a milliliter of blood, doctors usually recommend biopsies to see if cancer is present; needles are inserted into the prostate to withdraw cells for analysis. When a biopsy finds cancer, almost all men opt for treatment, usually surgery or radiation to destroy the prostate gland.

Most diagnostic tests have a trade-off between true positives and false positives. In order to increase sensitivity (true positive rate) once must sacrifice specificity (increase false positive rate).

Experts whom this study surprises have an incomplete understanding of test characteristics. They believed in a dichotomous or yes/no result. But clearly PSA is not that simple.

One can represent the trade-off between true positives and false positives using a technique called Receiver Operating Characteristic Curves. As the link demonstrates, some tests perform better at discriminating between disease and non-disease than do other tests. PSA has mediocre discrimination ability. Thus, we have a large overlap of true positives and false positives.

So how do we screen for prostate cancer in 2004? We can try to interpret the evidence, although that will not necessarily protect us from malpractice (Unbelievable malpractice case). We can just believe in the religion of PSA testing. We could biopsy all men (not me thank you very much).

We have no good answers for prostate cancer screening. I have no interest in receiving screening (although my physician insists). We will remain befuddled for some time.

Related posts:

1. Prostate cancer screening
2. The end of Pap smears
3. Not all screening helps patients
4. When guidelines differ – prostate cancer
5. Prostate cancer – to screen or not

Related posts brought to you by Yet Another Related Posts Plugin.