Brain Pre-surgical white matter Tractography Mapping challenge (BrainPTM) 2021¶
Background¶
The accurate mapping of white matter (WM) tracts is critical to
the success of neuro-surgical planning
and navigation. For the last 15 years, this task was exclusively based
on white matter tractography
algorithms [1]. Deterministic algorithms, that tracks the
principal direction of diffusion (PDD), as well as
probabilistic approaches, that generate tracts in a random walk process,
have both been applied for this
purpose [1]. In both cases, the manual delineation of accurate seeding
regions is performed, requiring specific
neuroanatomical knowledge and significant amount of time. Moreover, the
tract generation itself may
require significant computing time or power, as probabilistic algorithms
are generally needed to obtain a
sufficiently complete reconstruction of the motor and optic
radiation tracts using clinical diffusion weighted
imaging scans [1].
While the fiber tracts representation is useful for brain research, as
it enables quantitative measurementssampled along the tracts,
neuro-surgeons rather need a volumetric segmentation of the tracts for
mappingpurpose. Automating and accelerating this process would
significantly reduce the amount of time spent onneuro-surgical planning
phase and it might improve the accuracy of tracts mapping which is
crucial for brain surgery. Recent progresses in multi-modal deep neural
networks, suggest these may benefit to whitematter tracts mapping,
either by automating seeding regions generation [2], or by direct
segmentation of tract voxels [3].
In this context, the
CILAB organizes
this challenge to encourage the development of machine learning
approaches
to white matter tracts mapping in clinical brain MRI scans. Significant
progress in this area
will improve the neuro-surgical planning procedure in terms of time,
accuracy, and robustness.
Challenge¶
In this challenge we ask the participants to perform direct white matter
tracts mapping in clinical brain
MRI scans we provide. The data that is provided consists of 75 cases
(patients referred for brain tumor
removal) that were acquired
at Sheba Medical Center
at Tel HaShomer, Israel [2]. Patient pathologies
include oligodendrogliomas , astrocytomas, glioblastomas and cavernomas,
on first occurrence or in a
post-surgical recurrence. According to the neuro-radiologist's
estimation, the tumor volumes ranged
from 4 (cavernoma) to 60 [cm^3] (glioblastoma multiforme). Also,
different levels of edema are present
around the dataset tumors, from inexistent to very significant.
Along with each case both T1 Structural and Diffusion Weighted
modalities are provided. For 60 cases
(training) semi-manual white matters tracts mapping is provided in the
form of binary segmentation maps.
For the rest 15 cases (test) no tracts annotations are provided as these
will be used for participants
algorithms evaluation.
More details in participation section.
Data usage agreement¶
Participants cannot share the data, cannot use it for any commercial
purpose. If this dataset or part of it is used in a published
paper (as well as test set evaluation results from the leaderboard)
please cite the following papers:
@article{avital2019neural,
title={Neural Segmentation of Seeding ROIs (sROIs) for Pre-Surgical Brain Tractography},
author={Avital, Itzik and Nelkenbaum, Ilya and Tsarfaty, Galia and Konen, Eli and Kiryati, Nahum and Mayer, Arnaldo},
journal={IEEE transactions on medical imaging},
volume={39},
number={5},
pages={1655--1667},
year={2019},
publisher={IEEE}
}
@inproceedings{nelkenbaum2020automatic,
title={Automatic Segmentation of White Matter Tracts Using Multiple Brain MRI Sequences},
author={Nelkenbaum, Ilya and Tsarfaty, Galia and Kiryati, Nahum and Konen, Eli and Mayer, Arnaldo},
booktitle={2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI)},
pages={368--371},
year={2020},
organization={IEEE}
}
Contacts¶
The challenge is organized by the CILAB @ Sheba medical center, Israel¶
If you have any questions, please contact:
- Ilya Nelkenbaum, Computational Imaging Lab (CILAB), Tel-Hashomer, Ramat Gan, Israel
- Noa Barzilay, Computational Imaging Lab (CILAB), Tel-Hashomer, Ramat Gan, Israel
- Arnaldo Mayer, Head of Computational Imaging Lab (CILAB), Tel-Hashomer, Ramat Gan, Israel
References¶
[1] A. S. Bick, A. Mayer, and N. Levin, “From research to
clinical practice: implementation of functional
magnetic imaging and white matter tractography in the clinical
environment,” Journal of the neurological
sciences, vol. 312, no. 1-2, pp. 158–165, 2012.
[2] I. Avital, I. Nelkenbaum, G. Tsarfaty, E. Konen, N. Kiryati, and
A. Mayer, “Neural segmentation of seeding
rois (srois) for pre-surgical brain tractography,” IEEE Transactions on
Medical Imaging, vol. 39, no. 5, pp.
1655–1667, 2019.
[3] I. Nelkenbaum, G. Tsarfaty, N. Kiryati, E. Konen, and A. Mayer,
2020, April. Automatic Segmentation of White
Matter Tracts Using Multiple Brain MRI Sequences. In 2020 IEEE 17th
International Symposium on Biomedical
Imaging (ISBI), pp. 368-371. IEEE.