Circular RNA Beats pTau217 at Predicting Risk
What The Molecule Does
The biomarker at the center of this week's Nature Medicine paper is not a protein. It is RNA, folded into a loop. Most RNA runs in a straight line and gets chewed up by enzymes in the blood within minutes. Circular RNA forms when the strand splices back on itself instead, closing the loop, and that shape is what lets it dodge those enzymes and survive. It is a molecule built to last, which is exactly what you want in something you are trying to measure in a blood sample. A Washington University team led by Carlos Cruchaga screened blood from roughly 2,400 people across three separate study groups and found 34 circular RNAs whose levels rise and fall in step with Alzheimer's buildup in the brain, meaning they track the disease as it progresses. Tested for how well they diagnose existing Alzheimer's, this circular RNA panel scored 877 out of 1,000. The field's current standard blood test, which measures a tau protein fragment called pTau217, scored 977. So on diagnosis, the old test still wins. But diagnosis was not the number Cruchaga's team led with. Prediction was, and that is the harder job.
The Harder Question
Diagnosing someone who is already sick is one kind of test. Predicting who is going to get sick, before any symptoms show, is a different and much harder one. It is also the problem that has quietly broken anti-amyloid Alzheimer's drug trials for a decade. Here is why. Those trials enroll volunteers who test positive for the disease's biomarkers, but that group is a mixed bag: some of them will decline within a few years, and others might never develop symptoms at all. When a trial measures a drug's effect by averaging across that whole group, the people who were never going to get sick anyway drag the numbers down, burying any real benefit the drug delivered to the people who actually needed it. This is where the new circular RNA test pulls ahead. At predicting future decline, rather than just detecting current disease, it scored 676, about 29 percent better than pTau217. And the two biomarkers' signatures started diverging two to four years before symptoms appeared. That lead time matters: it gives researchers a window to identify, in advance, which volunteers are actually on track to decline, and enroll them in a trial before that decline sets in. In other words, this is not just a diagnostic. It is a tool for building a cleaner trial. Circular Genomics licensed the patent and launched a company, CircPATH, the same week the Nature Medicine paper came out.
The study's own authors say larger prospective studies are needed before this test is used in the clinic. Meanwhile, the annual Alzheimer's Association International Conference opened in London today, where Biogen and Vaccinex are presenting trial data this week on that same problem, picking the right volunteers before they decline. CircPATH, for its part, is already selling access to a test whose validation studies have not been run yet.