I think the correlation was a bit different. It's my understanding not everyone with amyloid build up had symptoms, but everyone with a confirmed AD diagnosis had amyloid build up.
I quickly found this, it looks like its based off a conference presentation but know there are older sources saying about the same:
> "So far, the results have been dramatic. Among 4,000 people tested so far in the Imaging Dementia-Evidence for Amyloid Scanning (IDEAS) study, researchers from the Memory and Aging Center at the University of California at San Francisco found that just 54.3 percent of MCI patients and 70.5 percent of dementia patients had the plaques.
A positive test for amyloid does not mean someone has Alzheimer’s, though its presence precedes the disease and increases the risk of progression. But a negative test definitively means a person does not have it."
Dementia is essentially a phenotype rather than a specific pathophysiology. (I know someone who died of a non-Alzheimer's form of dementia, specifically vascular dementia). I don't know what the exact definition of Alzheimer's is, but I believe it's based primarily on the neurodegeneration of the cerebral cortex rather than the definitive presence of amyloid buildup.
Whatever the details, they were diagnosing people and giving them "Alzheimers drugs" until they saw lack of amyloid-beta. To me it sounds like the "definitive evidence" is "telltale amyloid deposits":
“If someone had a putative diagnosis of Alzheimer’s disease, they might be on an Alzheimer’s drug like Aricept or Namenda,” said James Hendrix, the Alzheimer Association’s director of global science initiatives who co-presented the findings. “What if they had a PET scan and it showed that they didn’t have amyloid in their brain? Their physician would take them off that drug and look for something else.”
For decades, diagnosing Alzheimer’s has been a guessing game, based on looking at a person’s symptoms rather than testing for definitive evidence of the brain disorder. A firm diagnosis was not possible until an autopsy was performed.
Now, a spinal tap or PET scan can detect the telltale amyloid deposits, and researchers are trying to develop a simple blood test that would do so. PET imaging can quantify the amount of amyloid and also show where it is in a person’s brain."
