Eli Lilly and Co.’s Amyvid, an experimental imaging agent to detect signs of Alzheimer’s disease in the brain, shouldn’t be approved because of unreliable study results, a consumer-advocacy group said.

Amyvid, acquired through Lilly’s $300 million purchase of Avid Radiopharmaceuticals last year, is difficult to interpret in brain scans and could lead to false diagnoses of Alzheimer’s, Public Citizen said Tuesday. The Washington, D.C.-based group wrote a letter in the Journal of the American Medical Association criticizing a 35-person study of Amyvid published in January.

While Alzheimer’s affects as many as 5.1 million Americans according to the National Institutes of Health, a definitive diagnosis can now only be made after death through an autopsy. The U.S. Food and Drug Administration delayed approval of Amyvid in March and asked Indianapolis-based Lilly to create a training program to ensure brain scans using the drug are read accurately.

“Given what is known about Alzheimer’s disease and the lack of effective treatments, any proposed test for diagnosing this disease must be proven to be highly accurate in large, rigorously conducted studies before it is approved by the FDA for marketing,” said Michael Carome, deputy director of Public Citizen’s health research group, in a statement.

Lilly said in a separate statement Tuesday that the claims are “inaccurate.” The company said it worked with the FDA to identify appropriate study endpoints and they were all met and disclosed appropriately.

Additional doctors tested after the study had similar success reading brain scans and “continued assessment is ongoing,” representatives from Avid wrote in a reply to Public Citizen’s letter in the medical journal.

No current evidence exists to prove any treatment prevents or delays Alzheimer’s, which is expected to afflict 16 million U.S. citizens and 100 million people worldwide by 2050, said researchers Monday in an Archives of Neurology report calling for more extensive studies to find ways to target the disease.