Bas Jasperse MD,PhD, Presenter: Nothing to Disclose

Henri A. Vrooman PhD, Abstract Co-Author: Nothing to Disclose

Marcel Koek MSc, Abstract Co-Author: Nothing to Disclose

Meike Willemijn Vernooij MD, Abstract Co-Author: Nothing to Disclose

Marion Smits MD, PhD, Abstract Co-Author: Nothing to Disclose

Wiro Niessen PhD, Abstract Co-Author: Nothing to Disclose

Aad Van Der Lugt MD, PhD, Abstract Co-Author: Nothing to Disclose

Renske De Boer PhD, Abstract Co-Author: Nothing to Disclose

Patterns of brain atrophy and vascular changes, as visualized on brain-MRI, play an important role in determining the underlying cause of dementia syndromes. The visual interpretation of these pathological changes can be very challenging due to intercurrent age-related brain changes. To facilitate the distinction between abnormal and ‘normal age-related’ brain changes, we developed an automated system that provides individual-specific lobar brain volumes and white matter lesion volume taking into account age- and sex-specific reference data from a ‘healthy’ aging population.

Automated brain tissue segmentation and atlas-based registration is performed on T1, T2 and FLAIR MR-images, generating lobar brain volumes and white matter lesion volume. For each resulting volume, reference curves are used to generate age- and sex-specific percentile values. The results are then presented in an on-line viewer to interpret the quality of tissue segmentation and the quantitative results. Reference curves are generated from 5000 healthy aging individuals (age-range (years): 46-97, Male/female-ratio (N): 2246/2754).  

An example of segmentation results and percentile plots (colored lines represent percentile-lines) for a 69 year old male patient with suspected dementia is provided in the figure. Visual inspection suggests parietal atrophy, which might be attributed to global cortical atrophy. The automated analysis results show that overall brain volume is appropriate for age (75th to 95th percentile), with a strikingly low parietal lobe volume (37th percentile), confirming the initially suspected parietal lobe atrophy.

A system prototype is currently under evaluation at our radiology department. Future work aims to incorporate automated analysis of microbleeds, shape of individual brain structures and microstructural integrity using diffusion-weighted MRI sequences.  

automated quantification of pathological brain changes with regard to reference data from healthy individuals may facilitate the interpretation of brain MRI in the diagnostic work-up of dementia.

Jasperse, B, Vrooman, H, Koek, M, Vernooij, M, Smits, M, Niessen, W, Van Der Lugt, A, De Boer, R, A Computer Based Diagnostic System to Facilitate the Diagnosis of Dementia.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12027422.html