Advances in imaging have given us structural MRI to reveal atrophy, and fluorodeoxyglucose-PET (FDG-PET) to assess glucose metabolism and hence areas of hypometabolism. Most recently, the introduction of PET radiotracers such as 11C-PIB have allowed us to visualize amyloid burden in vivo – both in patients and in people at risk but still in the pre-dementia phase. And we now also have PET tracers that are promising in tau imaging.

Structural MRI correlates well with several features of post-mortem brain pathology. Whole brain volume correlates with the presence of neuritic plaques, Braak stage, and cerebral amyloid angiopathy; and ventricular volume correlates with demyelination, for example.

We have more imaging techniques, more complicated analyses, and certainly a more complicated picture of AD pathology.

We have more imaging techniques, more complicated analyses, and certainly a more complicated picture of AD pathology

The McKhann criteria now acknowledge the potential contribution imaging markers can make to the sensitivity and specificity – and timeliness -- of Alzheimer’s Disease (AD) diagnosis. But do the different imaging techniques correlate with each other; and how will they relate to other biomarkers such as those measured in CSF? Or to the myriad other markers -- of synaptic function, perfusion, inflammation and connectivity – that are being assessed in AD?

A complicated picture emerges

Studies described by Gaël Chételat (INSERM, University of Caen, Normandy, France), in her plenary lecture, reveal a complicated picture. The data from these complementary techniques do not always point in the same direction.

Perhaps the pathology is not as sequential as we thought. Do we need a dual pathway model in which different disease processes act in parallel, and at least to some extent independently?

Many people with hippocampal atrophy on MRI have no amyloid deposition on PET; many with hypometabolism on PET have no corresponding hippocampal atrophy or amyloid; and many whose brains showed amyloid deposition have no atrophy or hypometabolism.

The limited agreement between the different imaging modalities suggests they reflect different pathological processes. The complementary information provided indicates that people may need amyloid deposition plus either atrophy or hyometabolism before they undergo clinical cognitive decline.

The original hypothesis suggested a single cascade of events in which aberrant tau and neuronal dysfunction – and ultimately neuronal loss with accompanying neurotransmitter deficits – all flowed from a single pathological event, i.e. the accumulation of amyloid beta in plaques. It now looks as if there may be more than one pathological process acting in parallel.  

Different techniques, different pathologies

This would explain why some people, such as carriers of Apo-E4, experience neurodegeneration similar to that in AD – and show hypometabolism in the same regions as in AD -- without any amyloid deposition.

And it may turn out that while amyloid PET is highly specific for the presence of AD, FDG-PET is more closely related to conversion to clinical cognitive decline.

Imaging information can cause re-assessment of diagnosis

The complementary information provided by the three imaging biomarkers should also add to our understanding of pathophysiological processes. Work is underway into standardizing the different imaging techniques, so that they can be combined and compared – allowing identification of the regions where the markers are discrepant. Different patterns of discrepancy between the three techniques may relate to different disease processes and so add to the precision of diagnosis.

We have intriguing evidence of the way that information obtained by imaging can cause re-assessment of diagnosis. In a multicentre study in which Professor Chételat and colleagues were involved, a clinical diagnosis of AD was changed in 68% of cases when the physicians concerned became aware that a patient was negative on imaging for amyloid beta.