Share |

Identifying new biomarkers for multiple sclerosis

Illustration of the corpus callosum taken from Gray's Anatomy (1918). Image courtesy Wikimedia commons.

Multiple sclerosis (MS) is an inflammatory and degenerative brain disease that causes progressive physical and cognitive disabilities over many years. MS kicks in early and remains as the number one neurological problem affecting young adults.

Tobias Granberg, a PhD student based at the Karolinska Institutet in Sweden, used the Virtual Imaging Platform (VIP) to analyze the results of a long-term study of MS effects in the corpus callosum – an area of the brain highly sensitive to the disease (see image). VIP is a science gateway designed to provide access to grid computing and storage resources for medical imaging simulation.

Biomarkers, such as the relative volume of certain parts of the brain, are often used to monitor the progression of the disease or to provide earlier diagnostics.The results show that the volume of the corpus callosum is feasible as a quantitative biomarker for cognitive and physical disability in MS research and in clinical practice.

Although the root causes are still a mystery, the effects of MS on the brain are well understood: “The main feature of MS is an inflammatory response against myelin, a fatty substance that is wrapped around nerve fibers and functions as an insulator,” explains Granberg.

VIP in numbers

Users registered: ~800

Active users in 2014: 324

Available applications: ~15

Jobs executed in 2014: 1.7 million

CPU time consumed in 2014: 326 years

Publications by VIP users since 2013: 10 articles in journals, 12 articles for conferences.

Most used VIP applications: GATE, Freesurfer, and FSL.

Without myelin, the signals running through the nerve fibers are slowed down or interrupted. The consequences are sensory disturbances, motor problems with progressive physical disability, and cognitive impairment — all symptoms of MS.

Doctors use magnetic resonance imaging (MRI) techniques to diagnose MS and to pick the right treatment for each patient. Granberg investigated whether we can develop better MS biomarkers and for that he focused on the corpus callosum – the bridge that links the left and right halves of the brain with more than 200 million nerve fibers, most of them insulated with myelin and very sensitive to the effects of MS.

“Our main objectives were to study how the corpus callosum changes with increasing disease duration in long-term perspectives and to study if atrophy in this area is a good imaging biomarker for physical and cognitive disability in MS,” he adds. “Our hope was to find a corpus callosum measurement that could be easier to obtain than running advanced and time-consuming image post-processing.”

A long-term study

Granberg picked up a long-term study that started in 1996 with several disability and neurophysiological tests performed in 37 MS patients. The assessment was repeated first in 2004 and then in 2013.

EGI at its heart

VIP provides access to EGI resources for medical simulation and data processing. It is a complete science gateway, where applications are executed on the biomed virtual organization using the France Grilles DIRAC instance.

The platform uses a robot certificate and exposes a high-level web interface where the grid is totally hidden. In addition to this general-purpose execution service, VIP also has specific features for medical image simulation: multi-physics anatomical and pathological models are available in a repository, and simulated data are automatically annotated using concepts from an ontology.

“The physical disability was assessed according to EDSS, the expanded disability status scale, which is the most commonly used clinical measurement in MS,” he explains. Using EDSS, patients are assessed on a ten-grade scale depending on their physical mobility, where zero means no disabilities.

For the neuropsychological tests, the patients were asked to decipher symbols using a key with the numbers one to nine. This test is designed to assess cognition and speed of information processing.

The team also quantified the atrophy observed in the corpus callosum of the patients with three different methods. Two involved simple 2D measurements from side-view images of the brain. The third method was to calculate the volume of the corpus callosum. For that, Granberg used software called Freesurfer deployed on the VIP.

Virtual Imaging Platform

VIP is a web portal for medical simulation and image data analysis. It uses computing resources available in the biomed virtual organization of European Grid Infrastructure (EGI) to offer an open service to researchers worldwide.

“I used VIP to run brain tissue segmentations in MS patients and controls. In total there were 83 examinations to segment, and thus VIP greatly helped in making the volumetric analysis quick and feasible,” says Granberg.

Granberg tried using VIP following the recommendation of colleagues who were involved in its development. “I tried it initially for a few participants and was pleased with the quick segmentations, and that is why I continued using it in larger scale.”

A platform for the 'long tail' of science

With VIP, researchers from all over the world can access important amounts of computing resources and storage with no required technical skills beyond the use of a web browser. EGI is essential to the success of VIP because it provides an open infrastructure that relieves researchers of the burden of negotiating resource allocations with computing centers. Such an open policy enables the supply of services to the 'long tail' of science.

The results, published in Multiple Sclerosis Journal, suggest that the corpus callosum is a sensitive imaging biomarker for MS. “As corpus callosum area has excellent reproducibility and only takes half a minute to outline, this makes it feasible as a quantitative biomarker for cognitive and physical disability in MS research and in clinical practice,” Granberg concludes.


This article is republished with permission from EGI. Read more case studies on their website, here.

Your rating: None Average: 4 (2 votes)


Post new comment

By submitting this form, you accept the Mollom privacy policy.