Evaluation of open source medical imaging systems and assessment of multiple sclerosis MRI imaging.

Abstract

Multiple sclerosis (MS) is an autoimmune neurological disease which causes affects the myelinated axons of the central nervous system (CNS). Its characteristic is the formation for demyelinated plaques in white matter of the CNS. The disease progresses from reversible neurological deficits to permanent neurological deterioration. MS affects commonly young individuals, between 20 to 50, mostly men and imposes considerable burden for patients, presented in childhood or late middle age. Patients’ daily routine and activities are diminished since almost half of them need assistance with their basic needs, like walking, as the disease progresses over time making thus MS a vital problem to handle. Some of the symptoms include numbness, weakness, loss of balance and their diagnosis should be made by experienced physicians. The diagnosis is done by clinical findings and can be supported by further auxiliary tests, like magnetic resonance imaging (MRI).

Diagnosis is one of the steps to treat MS. Early prediction of the disease’s course is another. To do so, medical imaging techniques like MRI are obtained from the patients in order to estimate the symptoms’ path. To be able to accurately predict the progress of the disorder, large amount of data should be acquired and analysed, knowledge should be shared across research teams and collaboration should be promoted. In doing so, scientific workflow management systems (SWMS) in life sciences are used. Those systems utilize technological advancements promoting use of distributed resources as well as data integration, such as shared access to common data sources.

Three frameworks were chosen, MeVisLab, KNIME and Loni Pipeline, to be evaluated as open source medical neuroimaging systems with workflow management features. One of them, Loni Pipeline, presented several connectivity issues which has not been solved until recently.Therefore, it was discarded from the study. The rest were examined and evaluated in terms of image registration and image segmentation, 3D-volume reconstruction, lesion visualization and lesion measuring. The two remaining systems were compared against 3D-slicer, a software tailored to medical imaging research.