Final Report Summary - CONNECTMS (Brain connectomics: modeling disconnection syndrome in Multiple Sclerosis)
Brain architecture or connectivity is complex and defines the basis of cognitive processing. However, the complexity of brain connectivity and the technical limitations prevents comprehensive understanding of its organization. For this reason, the use of disease models producing disconnection syndromes, such as Multiple Sclerosis (MS), is a useful approach for improving our understanding on brain organization. MS is a common neurological disease that attacks the young population in the northern countries, women preferentially afflicted. It remains as a very challenging disease to diagnose and treat, due to its variability in the clinical presentation, and involvement in both white and gray matter of the brain. The diagnosis of MS is at present mainly guided by conventional structural MRI of the brain and spinal cord, seeking for evidence of both dissemination in time and dissemination in space of the MS lesions. However, this approach (conventional MRI) has several limitations, which do not reflect the actual disease state, and this contributes to the “clinical-MRI paradox”, namely the low correlation between MRI findings and clinical course.
In ConnectMS project we have developed novel methodology for understanding the brain organization using state-of-the-art, non-conventional MRI modalities. In other words in about 30 minute time interval we can acquire several types of MRI scans. One type called T1 can help us segment the white matter and the gray matter and different structures inside the gray matter (e.g. thalamus, motor area etc.). Also using this image we can segment the lesions in the brain that come from MS. Using the scan called diffusion MRI we can reconstruct the fibers in the brain. And using a brain scan called functional MRI we can capture the functional alterations in the brain. We can combine all these scans and create complex networks of the brain that afterwards we can analyze with already established methods in network theory. This gives a more complete picture of the state of the brain and development of the disease. We have applied this methodology to a patient cohort of 130 MS patients and 30 healthy controls. We have furthermore tested this methodology on different brain disorders such as Bi-polar disorder. To test the generality of the solution we worked together with Harvard Medical School and their pool of data. Moreover we started an active collaboration with the San Rafaelle Institute in Milan who sent us over 100 MS datasets to analyze with our pipeline. Finally we work on a collaborative study between TU/e in Netherlands and University of Sherbrooke in Canada where we expanded parts of the pipeline to clinical application of a brain tumor. This way we have created truly multi-center study that proves the benefits of the work done in the ConnectMS project.
The main benefits of this project is a general pipeline that can be applied to different brain disorders to not only understand more about the disorder but to guide the neurologist in the decision making for the diagnosis and prognosis of the brain diseases by giving additional information about the disease progression and offering possibilities to new therapies for the patients.
Bellow we summarize some of the main results:
• 30 healthy controls undergo several state-of-the-art non-conventional MRI protocols. Here we compared the scanning time, quality of data and test-re-test reproducibility in order to decide for the optimal protocol for our cohort of MS patients. Two publications arose from this work and several conference abstracts including one oral communication on the biggest international conference for MRI ISMRM (In Milan 2014) (see figure 1 and figure 2)
• 30 healthy controls from the data pool of Harvard Medical School (HMS) were analyzed with network techniques developed together between the PI of the ConnectMS project (Vesna Prchkovska) and the collaborator in Harvard Jorge Sepulcre. This was done during the 3-month visit of Vesna Prchkovska at HMS and there is a publication that is currently in preparation (see figure 3)
• 120 MS patients from our cohort were assessed with multiple volumetric measures along the time to improve the understanding of the brain atrophy in this disease. The collaboration with San Rafaelle Institute in Milan (SRI) provided 100 MS patients more. We used this data as a validation cohort for the generated models (figure 4)
• There is a publication currently in preparation. 130 MS subjects were analyzed with the developed pipeline. The results are currently being gathered together and a scientific publication is being prepared. This will be the main scientific publication arising from the work done with ConnectMS project (see figure 5 and 6)
• We tested some parts of the developed framework on a cases with MS and brain tumor in collaboration with TU/e and Univ. of Sherbrooke and found that the data can be enriched and more informative using our framework (see figure 7 and 8)
Other main contributions can be summarized here:
• 9 Bi-polar patients and matched controls have been analyzed with the developed framework and served as a proof of concept that the framework is general enough to be applied not only to healthy subjects and patients with neurodegenerative diseases but also for patients with psychiatric diseases. Several publications arose (some of them in preparation) and 1 masters thesis and co-supervision of 2 PhD students (in collaboration with University Ss Cyril and Methodius, Macedonia)
• Supervision of internship student
• Personal training of the fellow and presence on international conferences:
o 3 month visit of Harvard Medical School, Boston USA
o 1 week visit of Ecole Polytechnique Federale de Lausanne, Switzerland
o INCF Neuroinformatic course imaging the brain at different scales: how to integrate multi - scale structural information? - Antwerp, Belgium
o NITP UCLA Advanced Neuroimaging Summer program, Los Angeles, USA
o MICCAI conference 2013, Nagoya, Japan
o ISMRM conference 2014, Milan Italy
o Workshop on Connectomics in Multiple Sclerosis, Paris, France
o Invited talk on EU programmes for higher education and their role and impact on the Western Balkans, Belgrade, Serbia
===
Captions for attached figures for the publishable summary:
Figure 1. Summarized results of the differences of the MRI protocols
Figure 2. Summarized results of the quality of the fibers that different protocols can recover
Figure 3. Important parts of the brain so called ‘hubs’ found for switching information processing when signal is observed multiple times
Figure 4. Individual atrophy process for all MS patients can be seen separated by loos of complete brain, white and gray matter
Figure 5. Visualization of brain dataset of an MS patient. The lesions represented with red cubes are showing the brain damage along the tracks which can be explored interactively
Figure 6. Analysis of the Optic Radiation fibers of an MS patient where we can see how the lesions damage the tracts
Figure 7. MS patient processed with 3 different approaches. Our approach in c) shows how tracts can overcome the MS lesions and allow track-based analysis of the data
Figure 8. Our methodology (E-ODFs) compared to standard DTI for patient with tumor before and after surgery. Our approach explains the clinical findings in this case
In ConnectMS project we have developed novel methodology for understanding the brain organization using state-of-the-art, non-conventional MRI modalities. In other words in about 30 minute time interval we can acquire several types of MRI scans. One type called T1 can help us segment the white matter and the gray matter and different structures inside the gray matter (e.g. thalamus, motor area etc.). Also using this image we can segment the lesions in the brain that come from MS. Using the scan called diffusion MRI we can reconstruct the fibers in the brain. And using a brain scan called functional MRI we can capture the functional alterations in the brain. We can combine all these scans and create complex networks of the brain that afterwards we can analyze with already established methods in network theory. This gives a more complete picture of the state of the brain and development of the disease. We have applied this methodology to a patient cohort of 130 MS patients and 30 healthy controls. We have furthermore tested this methodology on different brain disorders such as Bi-polar disorder. To test the generality of the solution we worked together with Harvard Medical School and their pool of data. Moreover we started an active collaboration with the San Rafaelle Institute in Milan who sent us over 100 MS datasets to analyze with our pipeline. Finally we work on a collaborative study between TU/e in Netherlands and University of Sherbrooke in Canada where we expanded parts of the pipeline to clinical application of a brain tumor. This way we have created truly multi-center study that proves the benefits of the work done in the ConnectMS project.
The main benefits of this project is a general pipeline that can be applied to different brain disorders to not only understand more about the disorder but to guide the neurologist in the decision making for the diagnosis and prognosis of the brain diseases by giving additional information about the disease progression and offering possibilities to new therapies for the patients.
Bellow we summarize some of the main results:
• 30 healthy controls undergo several state-of-the-art non-conventional MRI protocols. Here we compared the scanning time, quality of data and test-re-test reproducibility in order to decide for the optimal protocol for our cohort of MS patients. Two publications arose from this work and several conference abstracts including one oral communication on the biggest international conference for MRI ISMRM (In Milan 2014) (see figure 1 and figure 2)
• 30 healthy controls from the data pool of Harvard Medical School (HMS) were analyzed with network techniques developed together between the PI of the ConnectMS project (Vesna Prchkovska) and the collaborator in Harvard Jorge Sepulcre. This was done during the 3-month visit of Vesna Prchkovska at HMS and there is a publication that is currently in preparation (see figure 3)
• 120 MS patients from our cohort were assessed with multiple volumetric measures along the time to improve the understanding of the brain atrophy in this disease. The collaboration with San Rafaelle Institute in Milan (SRI) provided 100 MS patients more. We used this data as a validation cohort for the generated models (figure 4)
• There is a publication currently in preparation. 130 MS subjects were analyzed with the developed pipeline. The results are currently being gathered together and a scientific publication is being prepared. This will be the main scientific publication arising from the work done with ConnectMS project (see figure 5 and 6)
• We tested some parts of the developed framework on a cases with MS and brain tumor in collaboration with TU/e and Univ. of Sherbrooke and found that the data can be enriched and more informative using our framework (see figure 7 and 8)
Other main contributions can be summarized here:
• 9 Bi-polar patients and matched controls have been analyzed with the developed framework and served as a proof of concept that the framework is general enough to be applied not only to healthy subjects and patients with neurodegenerative diseases but also for patients with psychiatric diseases. Several publications arose (some of them in preparation) and 1 masters thesis and co-supervision of 2 PhD students (in collaboration with University Ss Cyril and Methodius, Macedonia)
• Supervision of internship student
• Personal training of the fellow and presence on international conferences:
o 3 month visit of Harvard Medical School, Boston USA
o 1 week visit of Ecole Polytechnique Federale de Lausanne, Switzerland
o INCF Neuroinformatic course imaging the brain at different scales: how to integrate multi - scale structural information? - Antwerp, Belgium
o NITP UCLA Advanced Neuroimaging Summer program, Los Angeles, USA
o MICCAI conference 2013, Nagoya, Japan
o ISMRM conference 2014, Milan Italy
o Workshop on Connectomics in Multiple Sclerosis, Paris, France
o Invited talk on EU programmes for higher education and their role and impact on the Western Balkans, Belgrade, Serbia
===
Captions for attached figures for the publishable summary:
Figure 1. Summarized results of the differences of the MRI protocols
Figure 2. Summarized results of the quality of the fibers that different protocols can recover
Figure 3. Important parts of the brain so called ‘hubs’ found for switching information processing when signal is observed multiple times
Figure 4. Individual atrophy process for all MS patients can be seen separated by loos of complete brain, white and gray matter
Figure 5. Visualization of brain dataset of an MS patient. The lesions represented with red cubes are showing the brain damage along the tracks which can be explored interactively
Figure 6. Analysis of the Optic Radiation fibers of an MS patient where we can see how the lesions damage the tracts
Figure 7. MS patient processed with 3 different approaches. Our approach in c) shows how tracts can overcome the MS lesions and allow track-based analysis of the data
Figure 8. Our methodology (E-ODFs) compared to standard DTI for patient with tumor before and after surgery. Our approach explains the clinical findings in this case
