In line with a push to test potential Alzheimer’s therapies at earlier stages of disease (see editorial by Sperling et al., 2011, and ARF comments), new research boosts support for the use of magnetic resonance imaging (MRI) to identify seniors at the cusp of cognitive decline. In this study, some apparently healthy individuals with atrophy in a subset of cortical regions had cerebrospinal fluid biomarkers consistent with preclinical AD, and were more likely to develop subtle memory problems over the next three years. Brad Dickerson of Massachusetts General Hospital, Charlestown, and David Wolk, University of Pennsylvania, Philadelphia, reported the findings in the December 21 Neurology online.

Much structural neuroimaging research in AD focuses on the hippocampus and other medial temporal lobe (MTL) areas that are important for memory and degenerate early in the disease. Using MRI to measure thinning in these regions, researchers can separate older adults with AD from those without the disease. However, the methods cannot predict, on an individual basis, who will come down with AD—which is what a preclinical AD biomarker needs to do (see ARF related news story on Sperling et al., 2011, and Dubois et al., 2010).

Newer MRI strategies survey more of the brain, measuring atrophy not only in the MTL but in other regions as well—for example, those responsible for language and problem solving—which also deteriorate in AD. The current paper extends a line of research that began years ago when Dickerson and colleagues took newly diagnosed AD patients and compared their whole cerebral cortex to a group of healthy people of similar age, gender, and demographics. Using structural MRI, the scientists determined that thinning in a set of nine cortical areas—which they call the “AD signature” (ADsig)—correlated with symptom severity in mild AD patients. They also found slight but quantifiable ADsig loss in cognitively normal people with brain amyloid (Dickerson et al., 2009). Taking the approach into people with mild cognitive impairment (MCI), the scientists found that the thinner the patients' ADsig regions were, the faster they progressed to dementia (Bakkour et al., 2009). And more recently, Dickerson’s team detected shrinkage of these signature brain areas on MRI scans of seniors who were still cognitively normal, and determined that one standard deviation of cortical thinning tripled the risk of subsequent dementia (Dickerson et al., 2011).

That study, which came out in Neurology in April, analyzed two cohorts—one from MGH, the other from Rush University Medical Center in Chicago. Also, as presented at a recent conference, scientists collaborating with Dickerson tested ADsig in presymptomatic carriers of autosomal-dominant AD and found a signal (see ARF related news story). In the present paper, the scientists applied what they learned from those two samples to examine a third—159 cognitively normal subjects from the Alzheimer’s Disease Neuroimaging Initiative (see ARF ADNI series).

Dickerson and colleagues homed in on normal ADNI participants who fell below the 15th percentile (i.e., one standard deviation or more below mean) for ADsig cortical thickness. “We took that as an indicator of ‘silent AD’ and wanted to see if these people had other biomarkers that look like preclinical AD, and whether, over the next three years, they were more likely to develop subtle memory decline consistent with early AD,” Dickerson said.

Both turned out to be true. Of the 19 subjects deemed “high risk for preclinical AD” based on the 15th percentile ADsig cut-off, 21 percent showed signs of mental decline after three years. They had more errors on the Clinical Dementia Rating-Sum of Boxes test than the other subjects, and a standard deviation drop in at least one of three neuropsychological tests. In comparison, none of those with high cortical thickness (i.e., at least one standard deviation above mean) in the ADsig regions, and only 7 percent of those with average cortical thicknesses (i.e., within one standard deviation of the mean) showed cognitive loss. Furthermore, among “high-risk” participants with available cerebrospinal fluid (CSF) data, 60 percent had CSF Aβ1-42 levels consistent with AD, whereas only 19 percent of those with high cortical thickness and 36 percent of those with average cortical thicknesses had AD-like CSF reads.

The study “adds fuel to an already burning fire, supporting the use of a targeted, focused group of brain regions as indicators of AD,” Bill Jagust of the University of California, Berkeley, wrote in an e-mail to ARF. “I think the main novelty here is the CSF finding—showing that ADsig not only predicts cognitive decline, but also seems to be associated with a major Aβ biomarker.”

In recent years, AD researchers have puzzled over why some seniors can rack up brain amyloid without showing any signs of mental decline. In an accompanying editorial, Susan Resnick of the National Institute on Aging, Baltimore, Maryland, and Philip Scheltens of VU University Medical Center, Amsterdam, The Netherlands, suggest that the ADsig marker could help predict which of these amyloid-positive “normals” will progress toward dementia.

This capability would make ADsig useful in AD clinical trials, where biomarkers need to identify people “whose cognition will change in a meaningful way in some reasonable timeframe,” Dickerson said. “If people have amyloid in their brains but won’t develop symptoms for another five years, they won’t be informative for a drug trial that’s only going to last five years.” Toward that end, Dickerson’s group is trying to figure out how well the ADsig measure can predict the timeframe for symptoms to develop—“within the next two years, within the next four years, instead of saying changes are present or absent,” he told ARF.

How does the ADsig cortical thinning biomarker fare against other structural MRI measures in AD? It was “a little more sensitive and quite a bit more specific” than measures of hippocampal and entorhinal cortex volumes, Dickerson said, citing his earlier findings (Bakkour et al., 2009). ADsig has not been compared with functional MRI measures.—Esther Landhuis

Comments

  1. This paper is very interesting, and another in a series by Brad Dickerson showing that his "cortical signature" set of regions of interest (ROIs) seems to be a pretty good biomarker for AD. I think in the aggregate, the work is impressive—he also had a prior report (Dickerson et al., 2011) that was in Neurology this year showing predictive power of this set of ROIs in two different cohorts—one at Massachusetts General Hospital and the other at Rush. So, in a way, this is the third. I guess this is not so much "novel" as fuel for the already burning fire—supporting the use of a targeted, focused group of brain regions as indicative of AD. I think the main novelty here is the CSF finding showing that not only does this predict decline, but it also seems to be associated with a major Aβ biomarker.

    How will this be used? That is a good question. So far, the main structural biomarkers that seem most widely applied are hippocampal volume and whole brain atrophy rates (i.e., the brain-boundary shift integral [BBSI]—particularly by Nick Fox's group). But I think this is very promising. In fact, we are looking at a similar approach in my lab to see how this does at picking up PIB-positive normals. The ultimate questions are always (1) how dependent are these on the particular cohorts (now that there are quite a few studies with this biomarker, it is beginning to look generalizable, though, of course, the sample of "decliners" is still small); (2) how do the different biomarkers compare (i.e., will this biomarker do better than hippocampal volume, BBSI, FDG-PET, and Aβ measures); and (3) what do markers of neurodegeneration in general tell us that is different from markers of Aβ?

    References:

    . Alzheimer-signature MRI biomarker predicts AD dementia in cognitively normal adults. Neurology. 2011 Apr 19;76(16):1395-402. PubMed.

    View all comments by William Jagust
  2. This is an interesting study that aimed to assess whether a composite measure of cerebral atrophy—called an "AD signature"—was useful in predicting cognitive decline in "cognitively normal" elderly individuals from ADNI. It showed that individuals with lower volumes did indeed have more (subtle) cognitive decline over the following years.

    The "signature" is a composite of nine regions including the medial temporal lobe and a number of other cortical grey matter regions. One point of interest is that even at this early stage of AD (which is likely, but not certain to be the cause of the cognitive decline), the disease either involves diffuse areas of the brain and/or the disease is heterogeneous, with some individuals having a medial temporal lobe-led onset and others a more posterior cortical onset. A second interesting aspect is that this is another study confirming that diffuse volume loss is occurring at a very early prodromal stage (and clearly pre-diagnostic).

    In terms of clinical trials, regardless of whether this particular composite is the one to go for, the study points out the value of multiple regions (or some composite) in tracking early structural change, which is of relevance for the (hopefully soon to come) "prevention trials."

    View all comments by Nick Fox

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Reeling In Biomarker Data in Young Carriers, API Rocks Staging Boat

Paper Citations

  1. . Testing the right target and right drug at the right stage. Sci Transl Med. 2011 Nov 30;3(111):111cm33. PubMed.
  2. . Toward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement. 2011 May;7(3):280-92. Epub 2011 Apr 21 PubMed.
  3. . Revising the definition of Alzheimer's disease: a new lexicon. Lancet Neurol. 2010 Nov;9(11):1118-27. PubMed.
  4. . The cortical signature of Alzheimer's disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals. Cereb Cortex. 2009 Mar;19(3):497-510. PubMed.
  5. . The cortical signature of prodromal AD: regional thinning predicts mild AD dementia. Neurology. 2009 Mar 24;72(12):1048-55. PubMed.
  6. . Alzheimer-signature MRI biomarker predicts AD dementia in cognitively normal adults. Neurology. 2011 Apr 19;76(16):1395-402. PubMed.

Other Citations

  1. ARF related news story

External Citations

  1. Alzheimer’s Disease Neuroimaging Initiative

Further Reading

Papers

  1. . The cortical signature of prodromal AD: regional thinning predicts mild AD dementia. Neurology. 2009 Mar 24;72(12):1048-55. PubMed.
  2. . Alzheimer-signature MRI biomarker predicts AD dementia in cognitively normal adults. Neurology. 2011 Apr 19;76(16):1395-402. PubMed.
  3. . The cortical signature of Alzheimer's disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals. Cereb Cortex. 2009 Mar;19(3):497-510. PubMed.

Primary Papers

  1. . MRI cortical thickness biomarker predicts AD-like CSF and cognitive decline in normal adults. Neurology. 2012 Jan 10;78(2):84-90. PubMed.
  2. . MRI-based biomarkers of preclinical AD: an Alzheimer signature. Neurology. 2012 Jan 10;78(2):80-1. PubMed.