|
Instructions Using volBrain is a very simple process, consisting of four steps: 1. First, you have to register as a new user, or log into the system if you are already registered. 
2. Second, upload your compressed anonymized T1w brain MRI data in NIFTI format. Zip, Rar and Gzip compression formats are accepted. Most of MRI scanners produce DICOM data so a previous conversion to NIFTI format is required. There are many software packages that can do the job but we recommend dcm2nii (part of MRICRON software) which is a free and powerful tool. 3. Once your data is uploaded, volBrain will process your request as soon as possible and generate a report containing the results. 4. Finally, you will receive an e-mail informing about your request completion. The report will be included as an attachment, and you will be able to download a package including image files generated during the process.
Pipelines
Currently five segmentation pipelines are available:
1- volBrain pipeline volBrain is a pipeline of processes aimed to automatically analyze MRI brain data. It works as a black box from the user point of view as it gets an anonymized MRI brain volume in NIFTI format and produces a pdf report with the volumes of the main IntraCranial Cavity (ICC) tissues (that is, CSF, GM and WM). It also provides volume information of some macroscopic areas such as brain hemispheres, cerebellum and brainstem. Finally, automatic subcortical structure segmentation is performed and related volumes and label maps are provided. The average processing time of the whole pipeline is around 12 minutes. However this time can vary depending on the number of jobs queued in our server. In the next figure a brief outline of the process is displayed. 
volBrain labelling protocol Intracranial cavity mask, hemisphere/cerebellum masks and Subcortical labels were manually segmented by our expert on a set of T1w MRI. This training library is then used to perform automatic segmentation of your new cases. All structures were outlined according to our expert definition (our Lateral ventricles definition did not include choroid plexuses) with the exception of hippocampus which was segmented following the EADC protocol. volBrain example report Once the process is finished you will be notified by e-mail so you will be able to download a package including some image files and two (CSV and PDF) reports gathering all the volumetry values calculated from the segmentations. As you can see in the figure below the PDF includes patient information, parenchyma, brain tissues, macrostructure and subcortical structure volumes and also asymmetry indexes. It also includes several snapshots from the different labelling results as a quality control.
2- Ceres pipeline Ceres is a pipeline to automatically analyze Cerebellum MRI brain data. As for volBrain, it gets an anonymized MRI brain volume in NIFTI format and produces a pdf report with the volumes of the main cerebellum tissues (WM and GM) and the cerebellum lobules. It also provides cerebellar cortical thickness for each lobule. The average processing time of the whole pipeline is around 10 minutes. However this time can vary depending on the number of jobs queued in our server. In the next figure a brief outline of the process is displayed. 
Ceres labelling protocol Cerebellum white matter and lobules were manually segmented by experts in a set of T1w MRI. This training library is then used to perform automatic segmentation of your new cases. All structures were outlined according to the definition described in Park et al., 2014 (see reference below). Ceres example report Once the process is finished you will be notified by e-mail so you will be able to download a package including some image files and two (CSV and PDF) reports gathering all the volumetry values calculated from the segmentations. As you can see in the figure below the PDF includes patient information, parenchyma, cerebellum tissues and cerebellum lobules volumes, cerebellum cortical thickness and also asymmetry indexes. It also includes several snapshots from the different labelling results as a quality control.
3- lesionBrain pipeline lesionBrain is a pipeline to automatically segment white matter lesions from MRI data(T1 + FLAIR). As for volBrain, it gets two anonymized MRI brain volumes in NIFTI format and produces a pdf report with the volumes of the lesions and their locations The average processing time of the whole pipeline is around 20 minutes. However this time can vary depending on the number of jobs queued in our server. In the next figure a brief outline of the process is displayed. 
lesionBrain example report Once the process is finished you will be notified by e-mail so you will be able to download a package including some image files and two (CSV and PDF) reports gathering all the volumetry values calculated from the segmentations. As you can see in the figure below the PDF includes patient information, lesion clases , volumes and their locations in MNI space. It also includes several snapshots from the different labelling results as a quality control.
4- HIPS pipeline HIPS is a pipeline for automatic hippocampus subfield segmentation from monomodal (T1) of multimodal MRI data (T1 + T2). As for volBrain, it gets two anonymized MRI brain volumes in NIFTI format and produces a pdf report with the volumes of diferent subfields using two different delimitation protocols. The average processing time of the whole pipeline is around 20 minutes. However this time can vary depending on the number of jobs queued in our server. In the next figure a brief outline of the process is displayed. 
HIPS example report Once the process is finished you will be notified by e-mail so you will be able to download a package including some image files and two (CSV and PDF) reports gathering all the volumetry values calculated from the segmentations. As you can see in the figure below the PDF includes patient information, subfield volumes and their asymetries in MNI space. It also includes several snapshots from the different labelling results as a quality control.
5- pBrain pipeline pBrain is a pipeline for automatic segmentation of Parkinson related structures(substantia nigra, red nucleus and subthalamic nucleus) from monomodal (T2) at high or standard resolution. As for volBrain, it gets one anonymized MRI brain volumes in NIFTI format and produces a pdf report with the volumes of different structures. The average processing time of the whole pipeline is around 10 minutes. However this time can vary depending on the number of jobs queued in our server. In the next figure a brief outline of the process is displayed. 
pBrain example report Once the process is finished you will be notified by e-mail so you will be able to download a package including some image files and two (CSV and PDF) reports gathering all the volumetry values calculated from the segmentations. As you can see in the figure below the PDF includes patient information, subfield volumes and their asymetries in MNI space. It also includes several snapshots from the different labelling results as a quality control.
6- vol2Brain pipeline vol2Brain pipeline provides automatic brain segmentation dividing the volume in 135 structures. It also provides tissues, macrostructures and lobes segmentations as well as cortical thickness. When age and sex are provided, vol2Brain provides expected bounds for normal development. It works with anonymized MRI standard resolution T1-weighted images in NIFTI format. The average processing time of the whole pipeline is around 20 minutes. However this time can vary depending on the number of jobs queued in our server. In the next figure a brief outline of the process is displayed. 
vol2Brain example report Once the process is finished you will be notified by e-mail so you will be able to download a package including some image files and two (CSV and PDF) reports gathering all the volumetry values calculated from the segmentations. As you can see in the figure below the PDF includes patient information, structure volumes and their asymetries in MNI space. It also includes several snapshots from the different labelling results as a quality control.
Some papers describing parts of the system DenoisingJosé V. Manjón, Pierrick Coupé, Luis Martí-Bonmatí, Montserrat Robles, Louis Collins. Adaptive Non-Local Means Denoising of MR Images with Spatially Varying Noise Levels. Journal of Magnetic Resonance Imaging, 31,192-203, 2010. Inhomogeneity correctionNJ Tustison, BB Avants, PA Cook, Y Zheng, A Egan, PA Yushkevich. 2010. N4ITK: improved N3 bias correction. Medical Imaging, IEEE Transactions on 29 (6), 1310-1320. J Ashburner, KJ Friston. 2005. Unified segmentation. Neuroimage 26 (3), 839-851. RegistrationAvants BB, Tustison N, Song G. 2009. Advanced normalization tools (ANTS). Insight Journal. Intracranial cavity ExtractionJosé V. Manjón, Simon F. Eskildsen, Pierrick Coupé, Jose E. Romero, D. Louis Collins, Montserrat Robles. Nonlocal Intracranial Cavity Extraction. IJBI. Article ID 820205. 2014. Tissue classificationJosé V. Manjón, Jussi Tohka, Montserrat Robles. Improved Estimates of Partial Volume Coefficients from Noisy Brain MRI Using Spatial Context. Neuroimage, 53(2), 480-490, 2010. Hemispheric segmentationJosé E. Romero, José V. Manjón, Jussi Tohka, Pierrick Coupé, Montserrat Robles. NABS: Non-local Automatic Brain Hemisphere Segmentation. Magnetic Resonance Imaging. DOI: 10.1016/j.mri.2015.02.005, 2015. Subcortical structure segmentationPierrick Coupé, Jose V. Manjón, Vladimir Fonov, Jens Pruessner, Montserrat Robles, D. Louis Collins. Patch-based Segmentation using Expert Priors: Application to Hippocampus and Ventricle Segmentation. NeuroImage, 54(2): 940-954, 2011. Cerebellar lobule segmentationJose E. Romero, Pierrick Coupé, Rémi Giraud, Vinh-Thong Ta, Vladimir Fonov, Min Tae M. Park, M. Mallar Chakravarty, Aristotle N. Voineskos, Jose V. Manjón. CERES: A new cerebellum lobule segmentation method. Neuroimage, 147:916-924, 2017 Giraud, R., Ta, V.-T., Papadakis, N., Manjón, J.V., Collins, D.L., Coupé, P. and Initiative, A.D.N. An Optimized PatchMatch for multi-scale and multi-feature label fusion. Neuroimage In press, 2015. Cerebellar lobule segmentation protocolPark, M.T.M., Pipitone, J., Baer, L.H., Winterburn, J.L., Shah, Y., Chavez, S., Schira, M.M., Labaugh, N.J., Lerch, J.P., Vineskos, A.N. and Chakravarty, M.M. Derivation of high-resolution MRI atlases of the human cerebellumat 3 T and segmentation using multiple automatically generated templates. NeuroImage 95: 217 - 231, 2014. Cerebellar cortical thicknessTustison, N.J., Cook, P.A., Klein, A., Song, G., Das, S.R., Duda, J.T., Kandel, B.M., van Strien, N., Stone, J.R., Gee, J.C., Avants, B.B. Large-scale evaluation of ANTs and FreeSurfer cortical thickness measurements. Neuroimage. 99: 166 - 79, 2014. Lifespan normal brain volume modelsPierrick Coupe, Gwenaelle Catheline, Enrique Lanuza, José V. Manjón. Towards a unified analysis of brain maturation and aging across the entire lifespan: A MRI analysis. Human Brain Mapping, 2017. Hippocampus subfield segmentationJosé E.Romero, Pierrick Coupé, José V. Manjón. HIPS: A new hippocampus subfield segmentation method. NeuroImage. DOI: 10.1016/j.neuroimage.2017.09.049, 2017. vol2BrainJose V. Manjon, Jose E. Romero, Roberto Vivo-Hernando, Gregorio Rubio, Fernando Aparici, Mariam de la Iglesia-Vaya, Pierrick Coupe. vol2Brain: A new online Pipeline for whole Brain MRI analysis. arXiv:2202.03920, 2022. |
