The non-invasive colorectal cancer test detected 64% of samples from patients with pre-cancer polyps, and 85% of cancers in stool samples. The test looks for altered DNA from cancer cells that get shed into the stool.
That's probably not as good as a colonoscopy, but it provides a good alternative for the 40% of people over the age of 50 who should get a colonoscopy, but don't because ... well ... let's just say I hear they aren't much fun.
The test had a specificity of 88%; in short, 12% of the samples it called positive didn't actually have a sign of cancer. Again, that's not 100% exact, but since the patients with false positives would just go on to get an uncomfortable-but-generally-safe colonoscopy to confirm the test, a false positive rate in that range is more than acceptable.
How well could Cologuard sell? There's no real competition, and insurance companies UnitedHealth Group
But we're getting a little ahead of ourselves. Exact Sciences still has to run a clinical trial to confirm the results. These results came from frozen samples of well-characterized patients.
Exact Sciences opened down 14% today, but that may be mostly profit taking after a substantial run-up going into the data release. There's still some risk here -- witness Sequenom