The European Medicines Agency (EMA) has issued a draft opinion indicating that magnetic resonance imaging (MRI) of hippocampal volume is a valid approach for selecting people in the early, pre-dementia, stages of Alzheimer’s disease (AD) for clinical trials. The full document is available at the EMA website. The agency, headquartered in London, U.K., is now seeking comment from the public before making a final decision on adopting this opinion; the public comment period will end on 1 November 2011. “EMA’s opinion reflects the realization that loss of hippocampal volume is a marker for Alzheimer’s disease. As a marker, it is far from perfect, but there is clear evidence for a positive predictive value in the correct research situation,” said Nick Fox of University College London, U.K. “We are seeing regulators acknowledging the literature.” This draft opinion is consistent with another one the EMA issued last February regarding the use of two cerebrospinal fluid (CSF) biomarkers for selecting AD trial participants.

These two EMA opinions specifically focus on the use of markers to better select patients who are in the early stages of disease, sometimes referred to as prodromal AD. “These qualifications are not about making a diagnosis of Alzheimer's. We want to see if we can improve the selection of patients, and in that way improve the efficacy of clinical trials,” said Maria Isaac, a scientific administrator at the EMA in London. About half of patients who have clinical symptoms of mild cognitive impairment (MCI) go on to develop AD. That is because MCI, as defined using cognitive tests alone, can result from normal aging, from the disease process that will eventually lead to AD, or from some other disorder. Based on its review of the literature, the EMA concluded that if researchers select MCI patients who have clinical symptoms as well as low amyloid-β42 (Aβ42) and high total tau in their CSF, the proportion of patients who will develop AD increases. Now, the EMA is leaning toward the same conclusion for patients who have clinical symptoms of MCI plus a low hippocampal volume on an MRI scan taken at baseline, i.e., before any treatment is given. “We think that combining clinical diagnosis with biomarkers will help us better choose patients who will benefit from treatment,” said Isaac.

The EMA is responsible for regulatory approval of medicines in the European Union (EU). Its qualification opinions provide guidance on new methodologies or drug development tools used in trials to, for example, select patients or evaluate drug efficacy; such methodologies are often part of an application for a drug approval. Knowing where regulatory agencies stand on certain methods or techniques helps researchers design clinical trials that are more likely to get the regulatory go-ahead.

Researchers are already using a combination of biomarkers and/or imaging and clinical symptoms, collectively referred to as the “Dubois criteria,” to define patients with MCI who are more likely to progress to Alzheimer’s dementia (see ARF related news story on Dubois et al., 2007; Dubois et al., 2010). The Dubois criteria have also been incorporated in the revised clinical guidelines drawn up by the National Institute on Aging and the Alzheimer’s Association, which includes a preclinical stage of AD (see ARF related news story and ARF Webinar). Regulatory agencies had not yet weighed in on the value of these criteria for patient selection. The EMA’s opinions on CSF biomarkers and volumetric MRI are the first step in this direction. “Recognition of this approach [i.e., using volumetric MRI for patient selection] by a major regulatory agency should encourage more companies to move forward with these kinds of studies,” said Kevin McKenna, AstraZeneca, Wilmington, Delaware. “In the context of a research application for selection of patients in clinical trials, I think it will impact the field in a positive way.

The Food and Drug Administration (FDA), EMA’s counterpart in the U.S., has not yet issued similar guidelines. “We have told [clinical trial] sponsors that they can use the so-called Dubois criteria, or other biomarkers (in conjunction with some clinical signs) to identify and enroll early patients into clinical trials, but that we really won't know how to interpret these trials until the field more or less reaches a consensus about whether or not these markers truly identify patients with Alzheimer's disease,” wrote Russell (Rusty) Katz of the FDA in an e-mail to ARF. The Coalition Against Major Diseases (CAMD) of the Critical Path Institute in Tucson, Arizona, which asked the EMA to issue an opinion regarding volumetric MRI, is working with the FDA to obtain formal guidance both on CSF and imaging biomarkers. “We are at much earlier stages for that application process with the FDA [than with the EMA],” said Diane Stephenson, CAMD associate director. For a summary of CAMD’s projects on biomarker qualification, see ARF related news story.

These discussions come at a time of growing frustration in the AD field with the number of apparently promising treatments that have failed in large clinical trials. Some of the drugs tested have targeted Aβ, the protein that accumulates in the brains of AD patients. As a result, some researchers view these drug failures as an indication that the amyloid cascade hypothesis might be flawed. But another explanation, say others, is that the drugs have so far been tested in the wrong groups of patients. Trials have been conducted in people who already have mild to moderate AD, and may be too far along the disease process to benefit from Aβ-lowering drugs. It is also possible that patients recruited to participate in trials represent a diverse group with different pathologies, therefore masking any possible positive effects a drug might have on a more stringently selected and homogeneous group of patients. Increasingly, studies in other fields of medicine are focusing on tailoring therapies to specific subgroups of patients to achieve greater efficacy. “We have seen all the failures,” said Stephenson. “We need to go back, learn from them, and use this information to identify the right way to do these trials.”

“There is an impressive amount of evidence that loss of hippocampal volume helps predict who will progress from MCI to dementia,” said Giovanni Frisoni of the San Giovanni di Dio Fatebenefratelli Hospital in Brescia, Italy. “Essentially, people at the MCI stage [as defined by cognitive tests] who also have low hippocampal volume have about an 80 percent likelihood of progression to Alzheimer’s dementia in the following five years.” However, Frisoni and others point out that using this imaging marker as a way to enrich an MCI patient population is different from using it to gauge whether a drug is working or not, which involves taking repeated measurements at baseline and other time points. “Outcome biomarkers are very confusing. We have seen from clinical trials that sometimes a change in a biomarker does not match clinical outcome. The data on patient enrichment is much more clear cut,” said Stephenson.

That is not to say that using biomarkers to select patients is a perfect system. The EMA’s draft qualification opinion on volumetric MRI lists several caveats. For one thing, there is no standard “normal” hippocampal volume measure, or widely accepted definition of what constitutes pathological loss of hippocampal volume. An international effort led by Frisoni and Clifford Jack of the Mayo Clinic in Rochester, Minnesota, is trying to standardize measures of hippocampal volume obtained by MRI, and obtain datasets of normal hippocampal volumes that can be used for comparisons (see ARF related news story). Such efforts will be critical if low hippocampal volume is ever to become a diagnostic marker, said Frisoni. But for the purpose of selecting participants in clinical trials, the EMA concluded that individual laboratories could set their own cutoff values and obtain measurements that are internally consistent and reliable. Another caveat is that hippocampal volume measures can lead to false positives and negatives. “There are individuals with MCI who will have an apparently normal hippocampus and are in early stages of Alzheimer’s disease. Younger patients or patients with an atypical Alzheimer’s presentation will be more likely to have normal hippocampal volumes. Equally, there will be people with a slightly smaller hippocampus but they don’t have Alzheimer’s,” said Fox.

The EMA opinion does not cover people in the preclinical phase of AD. Paul Aisen, University of California, San Diego, and director of the Alzheimer’s Disease Cooperative Study (ADCS) agrees that the EMA’s decision is a positive development, but is leaning against using volumetric MRI as a selection criteria in ADCS’ planned Anti-Amyloid Treatment in Asymptomatic AD (A4) trial of asymptomatic patients, i.e., people at the pre-MCI stage (see ARF related Webinar and ARF related news story). “People with preclinical AD are scattered over a wide range of MRI measurements,” he said. Instead, Aisen plans to use CSF Aβ42 levels and imaging for amyloid in the brain, using positron emission tomography of Aβ-binding tracers such as florbetapir as selection criteria. “There are many reasonable criteria for identifying patients with early disease, as pointed out in the recent Dubois paper,” he said (Dubois et al., 2010).

Different markers may be more suitable to select participants for some types of trials than others. Recent research suggests that biomarkers follow different disease trajectories; changes in CSF Aβ42 probably occur early in the disease process, before any changes in cognition occur, whereas changes in tau and hippocampal volume occur later, when the disease is already well underway, but before dementia sets in (see ARF related news story). “Volumetric MRI seems to identify that intermediate stage where loss of hippocampal volume is more likely if a patient has AD, but clinical criteria don’t yet have the sensitivity that they would have when patients are more affected,” said Fox. Other criteria such as ApoE4 status, amyloid imaging, and CSF Aβ42 levels identify patients who are more likely to get AD, but the disease may be 10 to 15 years away, he explained. Disease-modifying drugs that target Aβ may possibly be more effective in patients with disturbances in Aβ metabolism, such as low CSF Aβ42 or high amyloid tracer binding in the brain. People without such changes but who have low hippocampal volume may respond better to other types of drugs.

While the MRI biomarker application was put forward by CAMD, the application for a qualifying opinion on CSF biomarkers was part of a proposal by the company Bristol-Myers Squibb, Princeton, New Jersey, as part of its application for a pre-dementia trial of an anti-amyloid compound. The results of that trial should be out in 2014.—Laura Bonetta.


Make a Comment

To make a comment you must login or register.

Comments on this content

  1. I applaud the EMA for taking this, and the previous, initiative, since they clearly may boost industry to use these markers in future clinical trials. Further, both EMA guidelines clearly underline the rationale of the concept reflected in the Dubois set of criteria. I would concur with Paul Aisen that MR hippocampal atrophy is too much a downstream marker to be used in the very early stage for selecting (future) AD cases. Amyloid markers are better suited for this, and within the amyloid-positive individuals, MRI may indicate the patients who are already further on the path to clinical AD.

Comments on Primary Papers for this Article

No Available Comments on Primary Papers for this Article


News Citations

  1. AD Diagnosis: Time for Biomarkers to Weigh In?
  2. DC: Biomarkers, Parkinson’s—CAMD Needs All Hands on Deck
  3. Paris: Standardization a Hurdle for Spinal Fluid, Imaging Markers
  4. Australia Report: The Way Forward—Pre-symptomatic Trials

Webinar Citations

  1. Two New Sets of Diagnostic Criteria—Which Is Right for Your Clinic?
  2. Treating Before Symptoms—ADCS Invites Ideas for Clinical Trials in Very Early AD

Paper Citations

  1. . Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria. Lancet Neurol. 2007 Aug;6(8):734-46. PubMed.
  2. . Revising the definition of Alzheimer's disease: a new lexicon. Lancet Neurol. 2010 Nov;9(11):1118-27. PubMed.

Other Citations

  1. ARF related news story

External Citations

  1. EMA website
  2. another one
  3. The Coalition Against Major Diseases
  4. international effort
  5. pre-dementia trial

Further Reading


  1. . Steps to standardization and validation of hippocampal volumetry as a biomarker in clinical trials and diagnostic criterion for Alzheimer's disease. Alzheimers Dement. 2011 Jul;7(4):474-485.e4. PubMed.
  2. . Report of the task force on designing clinical trials in early (predementia) AD. Neurology. 2011 Jan 18;76(3):280-6. PubMed.


  1. Australia Report: Lessons From a Clinical Trial—Too Little Too Late
  2. Australia Report: The Way Forward—Pre-symptomatic Trials
  3. AD Diagnosis: Time for Biomarkers to Weigh In?
  4. Revised Diagnostic Criteria for Alzheimer’s Are Published
  5. DC: Biomarkers, Parkinson’s—CAMD Needs All Hands on Deck
  6. Paris: Standardization a Hurdle for Spinal Fluid, Imaging Markers