At the Human Amyloid Imaging Conference, held on 12-13 January 2012 in Miami, Florida, the overriding story was that new data on amyloid imaging are largely confirming the key issues of its neuropathological correlation and its predictive power for cognitive decline. Remaining scientific debate on those fundamentals is moving on to finer points. And yet, underneath that fundamental accord, there were plenty of question marks. For example, on the clinical application of amyloid imaging and other biomarkers, scientists are confronting mismatches that leave them puzzled for the time being. In particular, amyloid imaging appears to contradict the clinical diagnosis of a significant fraction of patients, leaving the clinician-researcher to wonder whether the doc or the scan got it right. This question will find its answer in longitudinal observation, but in the meantime, amyloid biomarkers as currently incorporated into new diagnostic criteria leave the practicing physician with considerable uncertainty, researchers agreed.

Take, for example, the Alzheimer’s Disease Neuroimaging Initiative, (ADNI). Analyzing the data of this flagship study, several groups of scientists are discovering that some of its participants clinically diagnosed according to standardized criteria with probable AD might have something else. Either they are misdiagnosed or current biomarkers appear not to be worth their salt. At HAI, Susan Landau of the University of California, Berkeley, told the audience that, in a study addressing the separate topic of how amyloid and FDG-PET compare, she noticed to her surprise that 22 percent of clinically diagnosed AD patients from her ADNI sample were negative on their florbetapir scan. Similarly, Norman Foster of the University of Utah, Salt Lake City, reported on a poster comparing amyloid and FDG-PET in ADNI that he, too, found that 21 percent of his sample of 70 "probably AD" subjects were negative on their PIB or florbetapir scan (Foster et al., 2012). This, in Foster’s mind, raises concerns about the accuracy of the clinical diagnosis in ADNI. This is true in primary care, as well, Chris Rowe added in a subsequent discussion. Up to a third of people referred to Rowe’s center at Austin Hospital near Melbourne, Australia, with a diagnosis of AD turn out to be negative for amyloid PET, Rowe said.

Scientists want to understand where this discrepancy comes from. Does the problem lie with technical aspects of the PET scan, or with the clinical diagnosis? One way to answer this question is to follow "discordant" patients over time. Pascual Sanchez-Juan at the University of California, San Francisco, showed on a poster what happened to 15 patients at the UCSF Memory and Aging Center over the course of four years after the discordance appeared. Six of 69 people clinically diagnosed AD patients had a negative PIB scan, and nine of 65 people clinically diagnosed with a frontotemporal lateral dementia (which are non-amyloid diseases) had a positive scan. After that, these patients returned repeatedly for further assessments.

Of the PIB-negative AD cases, two stayed stable and their diagnosis changed to MCI due to psychiatric and vascular causes; in other words, the scan appears to have been right. Ditto for three more who evolved to an FTLD syndrome and had their diagnosis changed. One patient, however, continued losing memory and kept the diagnosis of AD; in this case, the scan might have been wrong. Of the nine PIB-positive FTLD cases, four progressed on a typical FTLD course, keeping their original diagnosis. Because they had brain amyloid, doctors prescribed cholinesterase inhibitors for them. Here, the amyloid deposition could have been incidental to their FTD. The five other PIB-positive FTLD patients evolved clinically toward AD, and their diagnosis was changed to AD; three added cholinesterase inhibitors (Sanchez-Juan et al., 2012).

Beyond this data presentation, the issue of how to weigh clinical diagnosis versus biomarker result when the two don’t match sparked intense discussion. Some scientists tended to place more trust in the biomarker, whereas others leaned toward the clinical finding. Attendees agreed that the issue needs to be clarified before diagnostic biomarkers can be widely implemented, which will require more standardization and defined cut points. Many expected that amyloid deposition may eventually sit at the top of a diagnostic hierarchy because it is seen to be the only biomarker that is specific to AD, but there was general consensus that it is too early to draw this conclusion. More data are needed.

As it is, biomarkers have been incorporated for the first time in the recently revised diagnostic criteria. That allows scientists to test how the new criteria perform in existing datasets. At HAI, Val Lowe of the Mayo Clinic in Rochester, Minnesota, presented one such exercise. He noted that the criteria contain a category called “Dementia Unlikely Due to AD” and applied it to ADNI data. In clinical practice, the diagnosing physician would invoke this category when a demented patient is negative for biomarkers of amyloid and downstream neuronal injury. How about it? In Lowe’s analysis, 9 percent of clinical AD cases in the ADNI sample he analyzed were negative for amyloid. Of those, some were negative for neuronal injury markers as well, and hence would be diagnosed as having dementia unlikely due to AD. However, some were negative for one set of biomarkers and positive for the other; those people are not captured at all by the new criteria as currently published, Lowe said (Lowe et al., 2012).

“It is pretty upsetting that these highly clinically selected, supposedly typical AD cases by the new criteria would either be indeterminate or have dementia not due to AD,” said Foster, “I would like to institute these criteria in my practice, but find it is quite often indeterminate.” Rowe agreed, noting, “It happens that I think patients have typical AD, and then they are negative on amyloid PET. I have to tell them I really do not know what they have.” Pathologists across the field have, of course, been noting for years that a fraction of people who died with a clinical diagnosis of probable AD turn out upon autopsy to have had either multiple mixed pathologies or no AD pathology at all. And at last year’s Alzheimer’s Association International Conference in Paris, Nigel Cairns of Washington University, St. Louis, Missouri, noted that, of the ADNI AD cases that have so far come to autopsy, 40 percent meet McKeith diagnostic criteria for dementia with Lewy bodies. In the past, this reality check simply never made its way back to the clinician during the patient’s lifetime, but now there is a technique to image at least one pathology.

Where does this leave the field? More biomarker data may well lead to further revision of the diagnostic criteria. In the process, there may be an unexpected upside. As amyloid imaging becomes more widely applied, “it will reveal a lot of non-AD dementia,” said Cliff Jack of the Mayo Clinic, Rochester, Minnesota. In this sense, an amyloid-based diagnostic tool may, in effect, bend Alzheimer’s case estimates downward a bit, and in its wake give added prominence to a collection of other dementias that have tended to languish in AD’s shadow in terms of awareness, research attention, and funding.—Gabrielle Strobel.

This is Part 8 of a nine-part series. See also Part 1, Part 2, Part 3, Part 4, Part 5, Part 6, Part 7, Part 9. Download a PDF of the entire series.

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. News Focus: 2012 Human Amyloid Imaging Conference
  2. Miami: Amyloid PET in the Clinic: What Are the Issues?
  3. Miami: Scan and Tell? Amyloid Imaging Confronts Disclosure Dilemma
  4. Miami: Can the Naked Eye Tell When a Scan Is Positive?
  5. Miami: When Does Amyloid Deposition Start in Familial Alzheimer’s?
  6. Miami: Age and Amyloid—What Has ApoE Got to Do With It?
  7. Miami: Longitudinal Amyloid PET Data Start Converging
  8. Miami: Scientists Angle for Way to Image Tangle

Paper Citations

  1. . Diagnostic Classification with Amyloid PET and FDG-PET among Clinically Diagnosed Alzheimer’s Disease Patients in the Alzheimer’s Disease Neuroimaging Initiative. Human Amyloid Imaging Abstract. 2012 Jan 1;
  2. . PET PIB Utility in Clinical Practice: Learning from the Unexpected Findings. Human Amyloid Imaging Abstract. 2012 Jan 1;
  3. . Biomarker Correlates in the ADNI AD Population Suggesting Dementia Unlikely Due to AD. Human Amyloid Imaging Abstract. 2012 Jan 1;

Other Citations

  1. Download a PDF of the entire series.

Further Reading

Papers

  1. . Amyloid imaging in the differential diagnosis of dementia: review and potential clinical applications. Alzheimers Res Ther. 2011;3(6):31. PubMed.
  2. . Differential Diagnosis in Alzheimer's Disease and Dementia with Lewy Bodies via VMAT2 and Amyloid Imaging. Neurodegener Dis. 2012 Jan 17; PubMed.