replicating the work in a prospective study”—that is, following up with a new batch of patients—“and in clarifying….how specific is the test to Alzheimer’s and not to other look-alike dementias.” Drug companies have not yet used blood tests for patient enrollment, “but pharma is definitely collecting samples in the event any of the tests can be validated,” she said.
The Eye
Imagine scanning someone’s eyes to predict Alzheimer’s. This is the most fertile ground for startups, and several are racing to build systems and prove their predictive power. Neurotrack Technologies, of Palo Alto, CA, uses a standard tablet or laptop camera to record the eye movements of a person watching images flash upon a screen. Some images are familiar, some are not. Neurotrack’s software analyzes how long the eyes linger on each image, a function of recognition memory, which is based in the brain’s hippocampus. Alzheimer’s damage starts there, likely years before the patient shows symptoms, and Neurotrack says its software can tease out subtle changes. (A healthy person spends most of the time looking at the novel images.) The test has one full data set so far: a 92-person, five-year study. A second five-year study at Emory University in Atlanta, where the test has its roots, is underway. And it has been incorporated into several studies, including the A4 presymptomatic study. “Hippocampal impairment is as early [a biomarker] as you’re going to get,” CEO Elli Kaplan told Xconomy. “We think our technology will help reduce some of the noise in the biomarker area.”
A second ophthalmological approach is a scan that measures amyloid deposits in the retina, where it turns out there are clues to many diseases, not just Alzheimer’s. But retinal amyloid seems to correlate to brain amyloid, and NeuroVision Imaging of Sacramento, CA, has a scan that can detect new plaques forming in tests three months apart, CEO Steven Verdooner told Xconomy. Those claims are based on NeuroVision’s inclusion in an Alzheimer’s trial run in part by Australia’s national science research organization. The company released interim data two weeks ago from the first 40 patients out of a planned 200 total. The number of patients tested in rigorous trials for both Neurotrack and NeuroVision are very small, to be sure. It’s early days. Both companies hope drug companies look past the small sample sizes and use the tests to help recruit patients for clinical trials. “Different companies have different criteria,” says Kaplan. Using bier tests before they’re fully validated isn’t something conservative pharma companies would typically do, but as Kaplan notes, time is passing. Vradenburg of USAgainstAlzheimer’s expresses that sense of urgency in broader fashion. If a drug has proved safe, which is critical when giving it to presymptomatic populations, the biomarker correlation shouldn’t have to be ironclad. “If you have at least some biomarkers, or strong hypotheses, regulators should lean forward” and let those at greatest risk take the drug, then follow them to understand whether the drugs are truly working. “The standard shouldn’t be that we are 100 percent certain.”
