Brown University public health professors Constantine Gatsonis, Ilana Gareen, and Roee Gutman will lead the statistical analysis of the $100-million, four-year IDEAS study, which will test how scanning for brain plaques may affect care and outcomes for patients with dementia or cognitive impairment.

PROVIDENCE, R.I. [Brown University] — Constantine Gatsonis, professor and chair of biostatistics at Brown, will lead the statistical analysis for a new national study on whether having the results of a diagnostic scan for amyloid plaques in the brain can affect the care and medical outcomes of people with dementia or mild cognitive impairment. Amyloid plaques are associated with Alzheimer’s disease, and positron emission tomography (PET) can detect them, possibly clarifying a diagnosis for patients with those conditions.

To help find the answers, the $100-million, four-year “Imaging Dementia — Evidence for Amyloid Scanning” (IDEAS) study will assemble a registry of more than 18,000 people who have had PET scans and will select a control group who have not. Researchers will then gather data for two aims:

  1. assessing whether having a scan result affects how care is managed — for instance whether the scans result in changes in the therapies or counseling offered patients; and
  2. determining whether patients who have scans are less likely to have adverse outcomes that require visits to the hospital or the emergency room in the following 12 months.

“The purpose of the IDEAS study is to examine how brain imaging, specifically an amyloid PET scan, helps guide doctors in diagnosing and treating Alzheimer’s and other dementias in cases where the cause of cognitive impairment is difficult to diagnose,” said Dr. Gil D. Rabinovici, M.D., IDEAS study chair and associate professor of neurology at the University of California–San Francisco. “We believe the study will show that, in diagnostically uncertain cases, knowledge of amyloid status will lead to significant changes in patient management — such as earlier counseling and prescription of more appropriate drugs — that will translate into improved long-term outcomes.”

Brown researchers have direct experience with the kind of data in the study. The University’s Center for Statistical Sciences, which Gatsonis directs, hosts the biostatistics center for the studies of the National Oncology PET Registry.

The new study will generate a huge amount of data, and tasks such as identifying appropriate matches to make meaningful comparisons will be difficult, Gatsonis said. He and Brown School of Public Health colleagues Roee Gutmann and Ilana Gareen, who collaborated to help design the study, will lead the complex statistical analyses needed to answer the questions.

“The study poses major methodological challenges,” he said. “[They] require innovation in both statistics and computation.”

The study will be funded by the Centers for Medicare & Medicaid Services and various private sources.