The American Association for Cancer Research has published a policy statement urging the FDA to regulate high risk LDTs. The fact that FDA intends to do so certainly isn’t news to anyone who is following the issue, but this is the first such statement that I know of from a medical professional association.
Throughout this process of watching FDA come to terms with LDTs, one of the things that I continue to find wanting is a clear definition of the harm that has come from poorly validated LDTs.
We all of course know that such harm is possible, and FDA has made general statements about negative outcomes, such as its stating that the agency “has been made aware of a number of examples where clinical decisions made on the basis of faulty tests resulted in harm to patients” (see page 32 of the report referenced in this post). But never has FDA specifically identified the harmful instances of which it is aware, and I have not read statements from any other entities regarding harm from poor quality testing.
The Ovasure Case: Proof of Harm?
In talking about about the risks to patients, the AACR policy statement cites a 2011 Nature article that highlights the problems with the short-lived OvaSure test (and which is probably most useful as a strong warning against physicians practicing statistics without a license!).
And while it does describe how insufficient validation can lead to a poor performing test, what is remarkable about this particular case is how quickly the test got shut down due to hue and cry from researchers and patients, ultimately leading to FDA stopping LabCorp from offering it after only four months.
This actually seems like an example where existing processes of peer review and public scrutiny succeeded in stopping public harm from happening (if there were in fact cases of women receiving false negatives from OvaSure and delaying treatment as a result they are not mentioned in the article).
Are there any examples of poor-performing, high-risk tests that managed (or continue to manage) to stay on the market for appreciable periods of time? Please let me know if you are aware of one as I am genuinely curious.
Along with poorly-validated tests, the AACR’s policy statement also cites mis-interpretation of germline DNA analyses as a possible risk from LDT testing. While an important issue to be sure, I would expect such issues to arise more often from mis-interpretation of personal genomics services than from the current and next generations of oncologists being insufficiently grounded in the genetic bases of cancer. If oncologists are not being trained to discriminate the disease-conferring differences of germline versus somatic mutations then we are truly in dire straits.
Make no mistake, I have no interest in seeing shoddy tests come to market, I am just proceeding from the assumption that the vast majority of CLIA labs and specialty LDT providers are run by people who have the public’s interest very much at heart, and who are very well-trained in what they do. And further, that they are being overseen by their peers and regulatory bodies in very effective ways.
When weighing the need for regulation against the desire to support innovation we would ideally have quantitative metrics for both sides of the issue, in order to accurately gauge the level of regulation that is truly required to support public health and safety. From my perspective we seem to be lacking such metrics on both sides of the scale and that is unfortunate.