It is really never a good sign when a miraculous new breakthrough in veterinary medicine is announced on the internet. That isn’t where true scientific breakthroughs show up. Legitimate science is a slow, detailed process where even brilliant, revolutionary ideas have to go through a long, rigorous process of critique and exploration before they are accepted. Public relations press announcements are about making money and generating hype, not about exploring good science.
It is even less inspiring of confidence when the supposed breakthrough is offered by a person or institution with a track record of promoting pseudoscience.
In the case of OncoK9 Liquid Biopsy, such red flags abound. Press announcements with hyperbolic language like “pioneering,” “revolution,” and “breakthrough innovation” are not only meaningless but outright misleading when the only evidence offered to support them is a company study that supposedly “will be submitted for publication in a leading peer-reviewed journal” someday. Legitimate science does not start with press releases and then move to possiblepublication in the scientific literature.
It is also not inspiring of confidence that the product is being initially offered for sale by Petco, a company which has a sketchy history when it comes to pet health topics. Selling raw diets, grain-free foods and bogus nutritional supplements while running a vapid and unscientific public relations campaign lauding its ban on “artificial” ingredients in the foods sold at PetCo stores does not establish the chain as a reliable source of information on pet health.
So what is OncoK9? According to the company that developed the product, it is “a multi-cancer early detection (MCED) test for the detection and characterization of cancer-associated genomic alterations in DNA isolated from canine whole blood samples, using next-generation sequencing (NGS) technology.” Sounds fancy, eh? The idea is plausible. Cancer cells may, depending on the type of cancer and many other factors, circulate in the blood. If such cells are present, it is potentially possible to detect them by identifying genetic markers associated with such abnormal cells. If there are then other tests available to identify and locate the cancer, and if there are therapies that might be more effective in treating the disease when caught at an earlier stage, such a test could improve health and lifespan in dogs.
Notice all the “ifs” and caveats here? The idea is sound, but the devil is in the details, and this company has not even come close to proving that all the steps in this process will work in a way that actually benefits dogs. The test was developed by mixing DNA from cancer cells and healthy canine cells and showing that the test could detect the abnormal DNA. Apparently, the test was then applied to samples from dogs know to have and thought not to have cancer (this is the study that has not yet been submitted for publication but from which the company is reporting results in its marketing materials). That’s a good start, but a long way from proving that regular clinical use of this test will benefit canine patients.
The company recommends using the test as a screening tool, “as an annual screening test for dogs that are 8 years and older and/or belong to breeds that are highly predisposed to cancer.” Screening is a particularly tricky kind of testing because you are looking at dogs without any signs or symptoms of disease. Most apparently healthy dogs are actually healthy, so the risk of false positive results, finding disease that isn’t actually there, is higher than when testing sick patients.
The risk of overdiagnosis is also high. This is where a test finds a disease that is truly there but isn’t likely to ever cause any symptoms. Many kinds of screening tests in humans, based on the same idea that early detection of asymptomatic disease is better, have been scaled back because it turns out they do more harm than good. Finding a type of cancer that isn’t ever going to progress or make someone sick isn’t in their best interests if subsequent testing and treatment cause harm without any benefit in terms of health or increased lifespan.
Assuming the methods and controls for error and bias were appropriate and the data is accurate (all of which we have to blindly assume in the absence of any scientific publication), the company study suggests an overall sensitivity of 48% for the test. This means that 48%, less than half, of dogs with cancer will correctly test positive. This is not, on the face of it, a very sensitive test. However, if we are looking for cancer in dogs without any clinical symptoms, that might actually be a good thing. We would rather miss a case in a dog with cancer that is not causing any problems than mistakenly identify a dog as having cancer who doesn’t.
The specificity, that is the proportion of dogs without cancer who correctly test negative, is reported to be 97%. This is the more important value in this case since we don’t want to falsely diagnose cancer and send a lot of dogs into unnecessary, costly, and potentially dangerous further testing or treatment they don’t need.
However, given that the company actively recommends the test as a screening tool, it is likely most of the dogs tested will be healthy. A rule for screening tests is that the less likely the patients you test are to actually have the disease you are looking for, the more wrong answers your test will give you. One way of looking at this is with the Positive Predictive Value (PPV) of a test.
The company advertising materials calculate the PPV for OncoK9 based on guessing that 10%-20% of dogs tested will have cancer. This seems high (which would make the test look better than guessing a lower rate of hidden cancers), and there aren’t good research data to tell us what the actually prevalence of asymptomatic cancer is in dogs of a specific age or breed. The best data we have are variable and not consistent between studies, so I will give the company the benefit of the doubt on these guesses for the time being.
Using these numbers, then, the PPV for OncoK9 would be between 64% and 80%. At the low end, that means that 64% of the dogs with a positive test actually have cancer and 36% do not. It seems a bit concerning that we are going to tell over a third of the dogs we test that they might have cancer when they don’t. Presumably, this would be followed by x-rays, ultrasound, other tests, and potentially months to years of anxiety about this supposed hidden cancer, none of which would improve the health or longevity of these dogs. Even at the more optimistic figure of 80% PPV, 20% or 1/5 dogs testing positive would actually be healthy. This raises the potential for significant costs and harms from unnecessary testing and treatment for a lot of dogs.
And remember all of the other caveats about screening tests for cancer? What if the OncoK9 test is correct but we can’t find the cancer anywhere until it progresses? What if it is a type of cancer for which there is no effective treatment? What if it is a type of cancer that is non-progressive and is never going to make the dog sick or lead to death? What if treating the cancer earlier, and for a longer time, doesn’t actually work better than treating it when it causes symptoms that we can detect without this test? All of these are likely scenarios, and in none of these cases would the cost of the test or the cost, anxiety, and risks of further testing and treatment offer any benefits for the dogs. This test has significant potential to cause harm not only when it gets the wrong answer but even when it is correct!
Of course, having a test that reliably detects cancer at a stage where the dog is still healthy and can be effectively treated in a way that preserves health and prolongs life would be a great thing. I would welcome such a test once it was proven to work as intended. Unfortunately, there is no way to know if OncoK9 will lead to net benefit or harm to dogs based on the evidence currently available. Despite the undoubtedly good intentions of the people developing and promoting this test, it is disappointing, even irresponsible to market it without having first done the necessary work to show it will actually make life better for our canine companions.
I will keep an eye out for the promised publications. I will even hold out a little bit of hope that eventually studies will be done to determine if the test as actually used in practice reduces disease and saves lives or not. Unfortunately, my fear is that once the test is on the market, the research will stop and continued use will be supported by heartwarming anecdotes about dogs “saved” by an early cancer diagnosis. We will likely never know how much misdiagnosis and overdiagnosis comes from this test since there will be no financial incentive to do the necessary research to figure this out. For now, I would urge dog owners to think very carefully about spending $400 for a test that involves big promises, little evidence, and a lot of uncertainty.