MRI May Offer Noninvasive Option for Alzheimer's Diagnosis

Megan Brooks

December 27, 2012

Magnetic resonance imaging (MRI) may provide a reasonably accurate noninvasive surrogate for cerebrospinal fluid (CSF) biomarkers to discriminate Alzheimer's disease (AD) from frontotemporal lobar degeneration (FTLD), new research suggests.

Using an MRI-based algorithm, investigators were able to differentiate between the 2 dementia types 75% of the time. This algorithm may reduce the number of more invasive diagnostic procedures such as lumbar puncture (LP), they note.

"Using this noninvasive MRI method, we obtain a single biologically meaningful value that can be used to derive a probabilistic estimate of the likelihood of AD or FTLD," first author Corey T. McMillan, PhD, from the Perelman School of Medicine and Frontotemporal Degeneration Center, University of Pennsylvania, Philadelphia, told Medscape Medical News.

"Patients are typically resistant to repeated lumbar puncture, and this approach may potentially provide a method for identifying patients for disease-modifying clinical trials and for measuring clinical trial endpoints," he added.

The study was published online December 26 in Neurology.

Reasonably Accurate

Despite distinct underlying neuropathologies, the clinical presentations of AD and FTLD may overlap, and although the ratio of total tau (tt) and β-amyloid (Aβ1-42) in CSF is a sensitive and specific in vivo diagnostic tool for discriminating between AD and FTLD, it requires a LP.

Prior studies have shown that patients with autopsy- or CSF-defined AD and FTLD have relatively distinct neuroanatomic profiles of gray matter (GM) neurodegeneration. The investigators derived a structural brain pattern from MRI that predicts the ratio of tt/Aβ in CSF to discriminate AD from FTLD.

The study involved 185 patients with a clinically diagnosed neurodegenerative disease that was consistent with AD or FTLD who had both an LP and a high-resolution volumetric MRI scan. Thirty-two of these patients had genetic or autopsy-confirmed AD or FTLD.

Dr. McMillan and colleagues found that MRI-predicted and actual CSF tt/Aβ values were highly correlated (P < .0001).

They also found that predicted tt/Aβ accurately defined the anatomic distribution of atrophy in AD and FTLD: Higher predicted tt/Aβ values were related to reduced GM density in posterior cortical regions, consistent with prior imaging studies of AD, and lower predicted tt/Aβ values were related to reduced GM density in anterior cortical regions, consistent with imaging studies of FTLD.

MRI-predicted tt/Aβ values were "reasonably accurate" at classifying individual patients as having AD or FTLD pathology. "With 75% overall diagnostic accuracy, MRI-based classification using predicted CSF is intended to serve as a screen to identify individuals requiring additional diagnostic studies," the researchers write.

Exciting and Novel

This study is "exciting and novel," and the results are "impressive," write Christian Habeck, PhD, from the Cognitive Neuroscience Division, Columbia University, New York City, and Jennifer L. Whitwell, PhD, from the Department of Radiology, Mayo Clinic, Rochester, Minnesota, in an accompanying editorial published in the January 2013 issue of Neurology.

The editorialists note the findings build on a body of work demonstrating that MRI could have an important role in predicting pathologic diagnosis.

The advantage of the multivariate approach taken by the University of Pennsylvania team over previous univariate studies, they point out, is that these complex MRI patterns can be summarized with a single diagnostically useful value that could potentially be applied in a clinical setting.

"One could also see how such a value could aid patient selection for future clinical trials that target these pathologies," Dr. Habeck and Dr. Whitwell write.

"However, an empirical test of the tt/Aβ prediction from brain data in an independent sample will be an important additional step before this method could be confidently applied to other cohorts," they add.

"The longitudinal stability and predictive accuracy of the MRI-predicted tt/Aβ ratio throughout the disease course will also require further clarification before it can be applied to heterogeneous clinical cohorts," the editorialists conclude.

The study was supported by the Wyncote Foundation and the National Institutes of Health. The study authors and editorial writers have disclosed no relevant financial relationships.

Neurology. 2013;80:126-127, 132-138. Article abstract, Editorial full text