When a patient steps into the doctor’s office and complains of memory problems, it would be nice if a few brain scans could reliably predict the person’s risk of future dementia. Don’t expect a single technique to win crystal ball status, but a recent batch of imaging studies suggests that the idea of combining methods to foretell impending Alzheimer disease is becoming less utopian. In last week’s Neurology, scientists present data to strengthen the case that subjective memory complaints may be visualized as specific brain pathologies. In a separate study of patients with mild cognitive impairment (MCI), a new statistical approach suggests that rates of whole brain atrophy predict progression to AD. Periventricular white matter lesions could also provide a means for distinguishing progressors from non-progressors, according to another recent paper. Meanwhile, a multi-modal imaging study using data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) indicates that positron emission tomography (PET) and magnetic resonance imaging (MRI) offer complementary and differentially sensitive ways of assessing memory.

Previous research has attributed early memory deficits in AD to hippocampal atrophy (Dickerson et al., 2001) and white matter lesions (de Groot et al., 2002). Additional studies have linked those pathologies to subjective cognitive complaints reported prior to any formal diagnosis but did not adjust for other abnormalities or for depressive symptoms, a possible confounder. In the new work—a prospective cohort study—Frank-Erik de Leeuw and colleagues at Radboud University Nijmegen in the Netherlands took such variables into account. Reporting in the 7 October issue of Neurology, first author A. G. van Norden and colleagues analyzed 503 seniors who were enrolled in an ongoing cohort study at the university. All had incidental white matter lesions but came up “normal” in a standardized battery of neuropsychological tests. A third of the participants had depressive symptoms, and 91 percent had subjective cognitive failure (SCF) as assessed by a questionnaire that asks, for instance, whether they have trouble remembering names of friends or finding their way around the neighborhood.

Adjusting for age, sex, education, depression, and intracranial volume, the researchers found that participants with subjective cognitive failure had smaller hippocampi than those without—regardless of white matter lesion severity. “The inference is that the subjects themselves are detecting that they have a problem, and the imaging is verifying it,” said Cliff Jack of Mayo Clinic in Rochester, Minnesota, whose earlier study of older MCI patients linked hippocampal atrophy with future conversion to AD (Jack et al., 1999).

The challenge of identifying who within the heterogeneous MCI population will eventually succumb to AD has become increasingly urgent with the realization that intervention needs to begin well in advance of overt dementia. Though there is no shortage of studies showing that regional brain atrophy patterns predict progression from MCI to AD, comparatively few have looked at global brain atrophy. Those that have (see, e.g., Sluimer et al., 2008 and Ridha et al., 2006) took serial MRI scans of subjects from a wide spectrum of disease states, showing that dementia severity correlates with extent of brain shrinkage. Within the MCI population, measures of hippocampal volume and either whole brain or ventricle atrophy rates were found to be somewhat useful for predicting conversion to AD (Jack et al., 2005). However, significant overlap between converters and non-converters made it hard to use such measures to determine individual outcomes.

Publishing 29 September in the Neurobiology of Aging, a group led by Hilkka Soininen at Kuopio University Hospital in Finland used a different statistical approach (iterative principal component analysis, or IPCA) to compute rates of whole brain atrophy in 102 MCI subjects. With this automated algorithm (Chen et al., 2004)—which can apparently detect tiny changes in brain volume without confounding effects from different scan protocols and intervals in longitudinal data—the authors report a similar lack of separation between those with stable and progressive MCI. In other words, they found no difference in annualized mean atrophy rates between MCI groups who did or did not convert to AD within the study’s 3.5-year interval.

“However, the big surprise came with the finding of a small but consistent difference when the studied population was stratified according to atrophy rate instead,” first author Gabriela Spulber wrote in an e-mail to ARF. Among subjects with no ongoing atrophy (0 percent/year), only 5 percent converted to AD within 3.5 years. However, among those with the highest atrophy rates (4 percent/year or greater), about 66 percent developed AD within the same timeframe. Given that 10-15 percent of MCI patients typically convert to AD each year (Petersen et al., 1999), it is likely that many in the so-called “stable MCI” group would eventually convert to AD if followed long enough, Spulber explained. The strength of the IPCA approach, in her view, is that it enabled her team to identify a single criterion with reliable predictive value over the few years of the study. Offering a broader interpretation of the new data, Jack told ARF it confirms, using a different method to analyze an independent cohort of subjects, “the underlying biological essence, which is that those whose brains are shrinking faster have a greater likelihood of clinical decline.”

A study led by Charles Decarli at the University of California, Davis, suggests another radiologic feature that might help distinguish MCI progressors and non-progressors—periventricular white matter lesions. In particular, first author Elisabeth van Straaten of VU Medical Center in Amsterdam, Netherlands, found that white matter hyperintensities in the periventricular zone, but not in deep subcortical areas, hastened progression to dementia. Their study is presented in a paper published 25 September in the Journal of Neurology online.

At first glance, this seems to contradict the data of de Leeuw and colleagues, which showed a relationship between subjective cognitive complaints and hippocampal atrophy irrespective of white matter lesions. However, the studies differed in several ways. Unlike the Decarli study, the analysis by de Leeuw’s group did not distinguish between white matter lesions in the periventricular and subcortical regions. The de Leeuw study looked at a community-based group of seniors who were cognitively normal by objective measures. The Decarli study, on the other hand, analyzed data on 152 amnestic MCI (aMCI) patients enrolled in a clinical trial studying the effect of donepezil or vitamin E on progression from aMCI to AD.

“People will look at two studies that may come up with different conclusions, and they'll attribute it to imaging methodological differences,” said Jack, who is a co-author on the Decarli study. “But when you look closely at the subject populations, they're totally different. That is far more likely to cause the different findings than any specifics about imaging.”

The prospect of assessing various imaging methods in a single population was key to the launch of ADNI, a five-year, multi-site study that will establish a brain imaging and biomarker database in hopes of standardizing image acquisition and determining which approaches are best for measuring drug efficacy in clinical trials (see ARF related news story). In a new study relating fluorodeoxyglucose (FDG)-PET and morphometric MR imaging to memory scores in healthy elderly and patients with MCI and AD from the ADNI database, “both methods were sensitive, and neither stood out as clearly superior,” lead author Kristine Walhovd of the University of Oslo, Norway, wrote in an e-mail to Alzforum. Walhovd communicated the findings, which appeared in the October 5 Neurobiology of Aging, on behalf of ADNI.

In normal people, FDG-PET was more sensitive than MRI in predicting memory performance, whereas MRI seemed to be the better predictor in those who were cognitively impaired. “PET has always been touted as the more sensitive measure,” Jack said, but considering the overall data, “that’s not really how it worked out. They had complementary information. When you combined MRI and FDG-PET, you got the best predictor.”

Considering the findings of these and other imaging studies, past and present, those who see the glass as half empty may wonder whether the reams of data are actually refining our understanding of pathology and progression, or confusing the picture. “The more studies that come out, the more complicated things become,” Jack said. “That's inevitably the way it is.”

However, in his more than two decades of work in this area, he sees within the thicket of data several lines of consensus. People with brain pathology are more likely to be impaired and, if followed over time, more likely to decline than those without pathology, Jack said. Furthermore—though it may be hard to predict, for instance, whether someone is worse off with a shrinking hippocampus or a buildup of white matter lesions—people with multiple indices of pathologies are more likely to be impaired and to progress more rapidly than those without. “It all hangs together in a thematically consistent way, but I appreciate the difficulties people would have in interpreting all these hundreds of papers that have come out,” Jack said. “That's just the nature of science.”—Esther Landhuis


Make a Comment

To make a comment you must login or register.

Comments on this content

No Available Comments


News Citations

  1. Sorrento: ADNI Imagines the Future of AD Imaging

Paper Citations

  1. . MRI-derived entorhinal and hippocampal atrophy in incipient and very mild Alzheimer's disease. Neurobiol Aging. 2001 Sep-Oct;22(5):747-54. PubMed.
  2. . Periventricular cerebral white matter lesions predict rate of cognitive decline. Ann Neurol. 2002 Sep;52(3):335-41. PubMed.
  3. . Prediction of AD with MRI-based hippocampal volume in mild cognitive impairment. Neurology. 1999 Apr 22;52(7):1397-403. PubMed.
  4. . Whole-brain atrophy rate and cognitive decline: longitudinal MR study of memory clinic patients. Radiology. 2008 Aug;248(2):590-8. PubMed.
  5. . Tracking atrophy progression in familial Alzheimer's disease: a serial MRI study. Lancet Neurol. 2006 Oct;5(10):828-34. PubMed.
  6. . Brain atrophy rates predict subsequent clinical conversion in normal elderly and amnestic MCI. Neurology. 2005 Oct 25;65(8):1227-31. PubMed.
  7. . An automated algorithm for the computation of brain volume change from sequential MRIs using an iterative principal component analysis and its evaluation for the assessment of whole-brain atrophy rates in patients with probable Alzheimer's disease. Neuroimage. 2004 May;22(1):134-43. PubMed.
  8. . Mild cognitive impairment: clinical characterization and outcome. Arch Neurol. 1999 Mar;56(3):303-8. PubMed.

Further Reading

Primary Papers

  1. . Subjective cognitive failures and hippocampal volume in elderly with white matter lesions. Neurology. 2008 Oct 7;71(15):1152-9. PubMed.
  2. . Whole brain atrophy rate predicts progression from MCI to Alzheimer's disease. Neurobiol Aging. 2010 Sep;31(9):1601-5. PubMed.
  3. . Periventricular white matter hyperintensities increase the likelihood of progression from amnestic mild cognitive impairment to dementia. J Neurol. 2008 Sep;255(9):1302-8. PubMed.
  4. . Multi-modal imaging predicts memory performance in normal aging and cognitive decline. Neurobiol Aging. 2010 Jul;31(7):1107-21. PubMed.