There was an interesting exposé in an issue of the New York Times
last month dealing with a promising technique in personalized cancer chemotherapy that turned out to be completely worthless. And this is unlikely to be an isolated incident.
Using patterns of gene expression to detect and tailor cancer treatment has turned into a huge "cash cow" and companies are springing up like weeds on a wet Spring day, offering great promise (at high price) that often is not backed by solid scientific research and in some instances may be based on downright fraudulent work.
The FDA has generally avoiding enforcing regulations on these new labs, and naive clinicians and patients have been flocking to them to get costly information that often is of no (or even erroneous) clinical value.
The case highlighted in the NY Times article began when a Duke University researcher, Dr. Anil Potti (several GREAT name-related puns come immediately to mind!), published a series of articles, the first in Nature Medicine, on complex gene signatures to determine appropriate cancer chemotherapy regimens. When others began reviewing the data from these studies they found both careless and inexplicable errors. Subsequently, the Cancer Letter reported that Dr. Potti had falsified his résumé, wrongly claiming among other things, that he was a Rhodes Scholar. Ultimately, his four papers on gene profiling were retracted and he resigned from Duke University. In short, his career was flushed down his last name (sorry, couldn't resist). A private company set up to sell his "cancer tests" was disbanded, but not before quite a few patients had received costly and worthless results and gone on to die of their diseases. As you can imagine, lawsuits are pending.
This is certainly another stain on the Duke University research record and those of you who have been around for a while will remember the famous Duke "B72.3" debacle from a few decades ago. There have been others. Not that I have anything against Duke.... well maybe I still do a little, since I found my interview for their MD-PhD program MANY years ago to be one of the most bizarre experiences in my life... but I suspect that this problem is multi-institutional, and the larger the research program the more likely there is, by mass effect, to be a "bad egg" falsifying data. In short, large institutions risk becoming victims of their own success. As the article noted, successful research labs have become so large that it is impossible for one person, even a conscientious senior PI, to fully comprehend everything that is going on. Given the pressure to produce, fraud becomes very easy under these circumstances.
Even if all of the other cancer gene profiling tests that seem to spring up daily are not overtly fraudulent, they are in most instances NOT backed by solid clinical trial results, or they provide costly information that can be obtained much more simply by other means. Clearly this in an area where the FDA NEEDS to step in to prevent false hope, improper care, and spiraling health care costs devoid of patient benefit.