Combining Diagnostic Tests Helps Predict Conversion to AD

Megan Brooks

December 18, 2012

Combining routine clinical tests with MRI, fluorine 18 fluorodeoxyglucose–positron emission tomography (FDG-PET), and cerebrospinal fluid (CSF) biomarkers is the most accurate way to predict conversion from mild cognitive impairment (MCI) to Alzheimer's disease (AD). However, FDG-PET appears to add the greatest prognostic information.

In this analysis of patients with MCI, "we asked the question: Do any of these 3 biomarkers add any new information to pencil and paper cognitive tests in terms of future prediction of Alzheimer's," Jeffrey R. Petrella, MD, from the Department of Radiology, Duke University Medical Center, Durham, North Carolina, explained in an interview with Medscape Medical News.

"We found that, in the best case, when you could do all of the tests, all of the tests added the most predictive information," he said.

"Having the luxury of doing all 3 tests is something that would be nice but it's not practical from an economic point of view, so we asked which of these 3 tests add the most diagnostic information to pencil-and-paper tests for predicting future development of Alzheimer's and the answer is FDG-PET," Dr. Petrella said.

The findings, from the Alzheimer's Disease Neuroimaging Initiative (ADNI), were published online December 11 in Radiology.

Unique Analytic Technique

Roughly 10% to 15% of patients with MCI progress to AD. Because new treatments are likely to be most effective early in the course of disease, the hunt is on for the most sensitive markers to facilitate detection and monitoring of early brain changes in at-risk individuals. MRI and FDG-PET are the 2 best-studied imaging markers to date.

Dr. Jeffrey R. Petrella

Dr. Petrella and colleagues used independent component analysis (ICA) to assess the extent to which whole-brain structural MRI, FDG-PET, and CSF proteins improve the ability to predict conversion to AD (over clinical data alone) in 97 patients (mean age, 75 years) with MCI from the ADNI.

ICA, they explain, can "isolate unique features from complex biomarkers, potentially revealing hidden patterns underlying 3-D imaging data sets," the researchers explain. To their knowledge, it's the first study to use ICA to assess the extent to which multiple biomarkers improve the ability to identify future decline in patients with MCI compared with cognitive testing and apolipoprotein E (ApoE).

During a mean follow-up of 31.5 months, 43 patients progressed to AD and 54 did not. The average time from initial screening to conversion was 20.7 months.

The researchers report that combining MRI, FDG-PET, and CSF data with routine clinical data (age, education, ApoE status, standard neuropsychological tests) significantly increased the accuracy of predicting conversion to AD compared with clinical data alone.

FDG-PET for All?

The misclassification rate based solely on clinical data and neuropsychological tests was relatively high at 41.3%, they say, but it decreased to 28.4% (P < .00001) when all 3 tests were used. "Each of these tests adds new information by looking at Alzheimer's from a different angle," Dr. Petrella commented in a statement.

Of the 3 tests, FDG-PET contributed more information to routine tests (P < .00001) than did CSF (P = .32) or MRI (P = .08), the researchers found.

"We are not recommending that every patient with MCI get an FDG-PET scan," Dr. Petrella told Medscape Medical News. The researchers say further validation, standardization, and cost-effectiveness studies are needed to translate the most useful biomarkers into routine clinical practice.

"We hope to do an economic analysis where we look at the actual dollar value costs of each of these tests compared to the diagnostic information that each of them yield, with pencil-and-paper tests obviously being the cheapest," Dr. Petrella said. "We'd also like to add 1 more test, the new amyloid scan, which has come out recently that look at the plaques in the brains of people with preclinical AD."

The study was supported by the National Institutes of Health. Dr. Petrella has disclosed no relevant financial relationships. A complete list of author disclosures is available with the original article.

Radiology. Published online December 11, 2012. Abstract