The NY times recently published an article with a catchy headline: ‘F.D.A. Targets Inaccurate Medical Tests, Citing Dangers and Costs’. For those outside of the industry, it may appear that the FDA painted a picture of them riding in to save the day again, akin to Clint Eastwood in a wild west movie. The message from the NY times was clear, but I’m not sure how accurate it actually may be and in fact, seems to be taken quite out of context. Not too dissimilar to my next use of this rainbow unicorn picture in my blog to get my point across:
F.D.A. Targets Inaccurate Medical Tests, Citing Dangers and Costs
The quotes were grand: “Patients have been demonstrably harmed or may have been harmed by tests that did not meet F.D.A. requirements” but I can’t help but think there is some post hoc ergo propter hoc logic here. That is, because ‘y’ is after ‘x’, ‘x’ must be the cause of ‘y’? In this particular example, the 20 case studies must be because of the lack of FDA oversight. But how is the FDA going to help in the ‘precision medicine’ space?
Precision medicine is an ‘unsustainable paradigm’
The industry often speaks grandly about ‘precision medicine’, roughly accepted as being: the right drug, for the right patient, at the right dose, at the right time. If you read between the lines, what this is actually saying is that there needs to be a Dx device linked to a specific therapeutic. Furthermore, if the test is required for the safe and effective use of the drug, then Dx device is termed a companion diagnostic (CDx). Currently, this testing paradigm for precision medicine links a specific drug to the Dx and can be summarized as “one-drug/one-gene Dx.” However, this paradigm of one-drug/one-gene cannot be sustainable owing to a number of attributes: Firstly, cancer is an exceedingly complex molecular and epigenomic disorder, resulting from perhaps hundreds of different molecular defects, including somatic mutations, gene expression changes and genome rearrangements. Not exactly a 1:1 ratio. Then you may as well throw the prognostic, the predictive biomarkers and the diagnostics biomarkers into the cooking pot. Furthermore, the advances of NGS, being able to sequence tens, hundreds and thousands of genes in one go also contribute to the good problem (?) of too much data and knowing too much.
Lastly, let’s also not forget, large pharma/CDx companies will probably not bother with those rare diseases where the potential revenue and subsequent profit is small because the analyte is rare. How could you ever have ‘precision medicine’ for the rare diseases?
The ability of FDA oversight
A quick scan of the ‘approved FDA tests list’ shows a fairly limited number of variants. Let’s compare this to COSMIC and you can clearly see that, of the 1,000 genes loosely linked to cancer and the 560-odd that have been studied, FDA approved tests only account for a maximum of 4% of these.
It doesn’t take a scientist to calculate the vast gap between what is needed, and what we currently have. To date, no (as far as I am aware?) NGS tests have been cleared or approved through FDA’s stringent pre-marketing review process, which verifies the performance claims of the test. The FDA approved roster is relatively small and there’s a reason for that; it takes an immense amount of time, money and effort to get a Dx test FDA-approved and science is moving faster than paperwork these days.
What does this mean?
The FDA presented 20 case studies around inaccurate LDTs, but I could quite easily pull out 50 case studies in return of ‘FDA approved tests’ going wrong due to: operator error, reporting inconsistencies, nomenclature…. I’d like the NY Times to report that. Am I saying FDA approved tests are bad? Absolutely not, and they do confer a certain guarantee of the efficacy/performance if set-up and used correctly. You can read the FDA’s point of view on this on Peter Laurie’s blog here.
I’m not arguing that there doesn’t need to be improvement in genetic testing, there’s always room for improvement. The FDA has its crosshairs on Laboratory Developed Tests (LDTs) and it doesn’t appear to be giving up until it gets the oversight that it wants. But will laboratories stop using LDTs in the next few years? Highly doubtful. It’ll be interesting to see how this develops.
Main image credit: bk1bennett