Genetic testing for prostate cancer “in the real world”

Over the past few months, the US Food and Drug Administration (FDA) has issued a slew of letters to companies marketing tests designed to provide consumers with information about their genetic profile and their risks for certain types of disease. The letters basically ask the companies to justify why they are marketing these tests in the absence of data demonstrating their clinical utility. You can click here for just one example of such a letter.

The FDA is concerned that many of these tests are being marketed to consumers with the implication that they are effective tools for the diagnosis of disease or for the prognosis of risk for a disease, but that these claims are not actually substantiated by reliable data. For a detailed account of the FDA’s viewpoint, you can read testimony to a Congressional subcommittee by the Director of the Center for Devices and Radiological Health, Dr. Jeff Shuren.

In addition, the Government Accountability Office (GAO) issued a report yesterday suggesting that misleading data developed by consumer-directed genetic testing “are further complicated by deceptive marketing and other questionable practices.” You can read the highlights here.

What’s this got to do with prostate cancer?

There are increasing amounts of data suggesting associations between the presence of specific alterations in the human genome and the presence of cancerous prostate tissue in specific individuals. Indeed, there are well over 20 single nuclear polymorphisms (SNPs) that have now been associated with risk for prostate cancer, and it does seem as though risk for prostate cancer increases in proportion with the number of such SNPs identifiable in any one individual. But what are the clinical implications of such data?

On the web site of deCODE Genetics — one of the foremost and most reputable companies seeking to develop gene-based diagnostic and prognostic information that will help to underlie personalized medicine — there is a detailed discussion of the deCODE ProstateCancer™ test. This test is based on the absence or presence of  just eight known genetic variants or SNPs. According to this detailed discussion:

Based on the presumption that these markers are independent, and the individual risks therefore multiply, the various genotype combinations have associated relative risks in the range of 0.33 (non carriers for any of the risk markers) to 17.6 (homozygous for all of the eight risk variants) compared to the general population risk.

But the presumption that these markers are all independent is a big one, and the test can still tell you and your doctor little about the aggressiveness of any cancer that might be found — although two of the eight SNP variants are associated with greater risk for more aggressive forms of prostate cancer. Another problem with the deCODE Genetics test is that it was developed primarily on the basis of data from Icelandic and other European (and therefore predominantly white) male populations. It’s relevance to African Americans and other non-white (or even non-Scandinavian) populations may therefore be limited.

In an article in The Times (an English newspaper) published in September 2009, the CEO of deCODE Genetics and members of an affiliated research team talked about a genetic test for prostate cancer that is based on as many as 20 different genetic markers. Since the currently available test is based on only eight markers, we assume that the data from the 20 available markers is yet to be integrated into a commercially viable test for risk of prostate cancer.

An article on Genomeweb yesterday discussed the promotion of “deCODE’s prostate cancer risk test” to clients of ARUP Laboratories — a national reference laboratory based in Salt Lake City. According to this article,

Clinically, the test … can identify approximately 15 percent of men in the general population who are at twice the average risk of developing [prostate cancer], and 5 percent of those who are triple the average risk.

That is not exactly a clinically important increase in risk for prostate cancer overall. If it was an accurate assessment of risk for clinically significant prostate cancer, the situation might be different, but the test can’t tell you that with accuracy.

ARUP Laboratories’ clients would presumably be physicians and hospitals and other types of provider organization. We have to assume that ARUP and deCODE Genetics believe that promotion of the deCODE Prostate Cancer test to physiciansand other providers  is still acceptable under current FDA regulations. In addition, the Genomeweb article discusses another genetic test for prostate cancer risk that is being developed by a company called Mitomics. The Mitomics test is based on an analysis of mitochondrial DNA found in man as opposed to the human DNA. Whether it is really any more clinically useful than the deCODE Genetics test isn’t known at this time.

Over the next 5 years we are likely to see the development of a whole series of genetic tests for risk of prostate cancer. Here are some of the factors that, in the opinion of The “New” Prostate Cancer InfoLink, would help to make sure that these tests are actually of real clinical value:

  • A test can clearly identify men at significant risk for clinically significant prostate cancer with a very high degree of accuracy.
  • A test can identify men who are at least 10 times more likely than the average man to be diagnosed with prostate cancer before age 60.
  • A test can clearly differentiate (or at least help to clearly differentiate) between men at risk for clinically indolent and clinically aggressive disease.
  • A test can clearly identify high risk for clinically significant prostate cancer in men for whom no other test can offer an accurate risk profile.
  • A test can clearly identify high risk for prostate cancer that will not respond to treatment for localized disease.

Obviously no one test would be likely to be able to do all of the above, but a test that could do any one of the above with a high degree of selectivity and specificity would be of considerable value. In the meantime, if you want to get one of the current genetic profiling tests for your risk of prostate cancer, remember …

  • Your health plan is highly unlikely to cover this cost.
  • You will probably need your doctor to order it for you.
  • The data that can currently be provided has limited clinical significance (because we don’t know if it is clinically meaningful).
  • It is likely that entering data into the Prostate Cancer Risk Calculator will give you at least an equally accurate estimate of the appropriateness of a prostate biopsy — and an estimate based on clinical data from > 5,500 men.

4 Responses

  1. The FDA has approved the PSA test in conjunction with a DRE for the diagnosis of prostate cancer.

    In the MedPage Today article, an ROC analysis for PSA was reported: “in analysis using ROC curves, the area under the curve (AUC) is a key guideline; an AUC of 1 means the test is perfect — catching every case and not giving any false positives — while an AUC of 0.5 means the test is no better than random chance. For cancer overall, Dr. Thompson and colleagues found an AUC of 0.678 — better than random but not much. ”

    I do not see statistics reported by the the FDA or the GAO demonstrating AUC results for the prostate cancer markers reportedly used by DeCODE or others that prove that DeCODE’s results are inferior to the PSA test approved by FDA. The impression from reading the above proceedings is that DeCODE’s AUC is 0.5 and does not add to AUC if used in conjunction with the PSA. These governmental bodies are concerned that doctors will surgically remove prostate glands based on DeCODE’s risk predictions instead of actually diagnosis. In any case, many well-informed people are concerned about over-treatment on the order of 50 treated to 1 actually saved. With over 200,000 new men each year joining the prostate cancer men’s club, there appears to be strong need even for minor improvements. Putting DeCODE out of business probably will not spur investors to risk money in this area. Your hope for genetic diagnostics being developed in the future most likely has taken a significant hit much like the situation in Germany.

  2. Dear Tasker:

    There is nothing to stop deCODE applying for approval of their test(s). And as far as I am aware, they are fully able to provide their test based on a sputum sample provided by a physician. What concerns the FDA is the direct marketing of these tests to consumers.

  3. Your information on deCODE is in error. I am a subscriber to deCODEme. The following SNPs are being used to compute my risk for prostate cancer: Rs 7127900, 10896449, 1859962, 8102476, 5759167, 10207654, 2660757, 10934853, 7679673, 1512768, 16902104, 620861, 1050483, 14447295, 10086908, 2710646, 10486567,2735839, 6465657, 10943994, 5945572, 6983267, 9364554, 4403796, 401681. These 25 SNPs — including the number of cases and controls in each study — are being reported. Also your explanation of the risk calculation used is incorrect, too. By the way, some time ago when I asked one of my doctors to allow me to get this test done, he refused. When I was able to subscribe, I found it valuable and well worth the cost.

  4. Tasker: “My” information on the deCODE Genetics prostate cancer test reports exactly what is on their web site. Apparently they haven’t updated the web site to reflect what they are actually doing. If you are subscriber, perhaps you should suggest to them that updating their web site would be a good idea.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: