The diagnostics space is choked with countless cancer tests – from biopsies and imaging analyses to a simple blood test. Bay Area startup Miroculus is working on a device that can conduct multiple cancer tests from a single blood sample.
Wired wrote an in-depth profile of the company, which is worth a read. It begins:
Early detection, we’re often told, is the surest way to beat cancer. It’s the reason why, year after year, men and women of a certain age dutifully visit their doctors and undergo uncomfortable tests to screen for things like prostate and breast cancer.
But what about the other hundred or so types of cancer out there—the brain cancers, the ovarian cancers, the leukemias and lymphomas? And what of the millions of young people who never get tested at all, even though they’ve been found to have worse outcomes than adults?
The device, called Miriam, is a low-cost and open-sourced project that debuted at the TEDGlobal conference in Rio De Janiero last week, Wired says. The company’s goal is to make it simple enough for untrained health workers around the world use it to test for a variety of different cancers. Wired nicely describes how Miriam works:
The Miroculus technology is based on microRNA, a class of small molecules that can act as a type of biological warning sign, appearing and disappearing based on what is happening in our bodies at that moment. As a result, they’ve become effective indicators of diseases—including cancer—ever since they were first discovered in 1993. They can reveal not just whether a person may have cancer, but what specific type of cancer that person might have.
For years, however, researchers believed microRNA could only be found inside of cells, making these biomarkers less accessible. But in 2008, a group of researchers discovered microRNA circulating in blood, spawning a wave of interest from other scientists, who viewed microRNA as the key to early cancer detection.
The whole article’s really worth a read. Also worth reading is the article’s top comment, which casts a critical eye on the efficacy of such a project:
No! This is a perfect example of why everyone should be required to learn statistics.
There’s a very good reason we do not do mass screening for cancers and other serious diseases with low incidence, no risk of transmission, and invasive treatment regimens. Medical tests are not crystal balls. Any test you perform has a risk of false positives and a risk of false negatives, and even if you have a test with good *specificity* and *sensitivity*, applying any test to any large population with a low incidence of disease guarantees that most of the positive findings will be wrong. Go google specificity and sensitivity if you remain unconvinced.
Telling hundreds of thousands of healthy people that they have cancer and sending them through intensive workup and/or treatment when they actually do not is emotionally and financially harmful to them and the broader community.
This is why we do *not* do population-wide CT screening for lung cancer, why we no longer stress that young women without other risk factors have mammograms (but can choose to if they are risk averse and are okay with potentially receiving unnecessary treatment), why we are moving away from aggressive PSA screening, etc.
What’s the alternative? Targeted screening. If you reduce the size of people you screen to those most likely to be affected, then you dramatically improve your results. So people with symptoms (obviously), people with certain exposures (asbestos), genetic predispositions to certain types of cancers (BRCA), etc. In other words, the thing that works best is what we already do. This product is solving a problem that doesn’t exist.
