Automating spurious streamline removal to improve manual tractography in paediatric tumour patients
Poster No:
1291
Submission Type:
Abstract Submission
Authors:
Steven Greenstein1, Sila Genc1,2, Joseph Yang2,3,4, Baran Aydogan5,6
Institutions:
1Developmental Imaging, Murdoch Children's Research Institute, Melbourne, Victoria, Australia, 2Neuroscience Advanced Clinical Imaging Service,Department of Neurosurgery, Royal Children's Hospital, Melbourne, Victoria, Australia, 3Neuroscience Research, Murdoch Children's Research Institute, Melbourne, Victoria, Australia, 4Department of Paediatrics, University of Melbourne, Melbourne, Victoria, Australia, 5A.I. Virtanen Institute for Molecular Science, University of Eastern Finland, Kuopio, Finland, 6Department of Neuroscience and Biomedical Engineering, Aalto University School of Science, Espoo, Finland
First Author:
Steven Greenstein
Developmental Imaging, Murdoch Children's Research Institute
Melbourne, Victoria, Australia
Developmental Imaging, Murdoch Children's Research Institute
Melbourne, Victoria, Australia
Co-Author(s):
Sila Genc
Developmental Imaging, Murdoch Children's Research Institute|Neuroscience Advanced Clinical Imaging Service,Department of Neurosurgery, Royal Children's Hospital
Melbourne, Victoria, Australia|Melbourne, Victoria, Australia
Developmental Imaging, Murdoch Children's Research Institute|Neuroscience Advanced Clinical Imaging Service,Department of Neurosurgery, Royal Children's Hospital
Melbourne, Victoria, Australia|Melbourne, Victoria, Australia
Joseph Yang
Neuroscience Advanced Clinical Imaging Service,Department of Neurosurgery, Royal Children's Hospital|Neuroscience Research, Murdoch Children's Research Institute|Department of Paediatrics, University of Melbourne
Melbourne, Victoria, Australia|Melbourne, Victoria, Australia|Melbourne, Victoria, Australia
Neuroscience Advanced Clinical Imaging Service,Department of Neurosurgery, Royal Children's Hospital|Neuroscience Research, Murdoch Children's Research Institute|Department of Paediatrics, University of Melbourne
Melbourne, Victoria, Australia|Melbourne, Victoria, Australia|Melbourne, Victoria, Australia
Baran Aydogan
A.I. Virtanen Institute for Molecular Science, University of Eastern Finland|Department of Neuroscience and Biomedical Engineering, Aalto University School of Science
Kuopio, Finland|Espoo, Finland
A.I. Virtanen Institute for Molecular Science, University of Eastern Finland|Department of Neuroscience and Biomedical Engineering, Aalto University School of Science
Kuopio, Finland|Espoo, Finland
Introduction:
Diffusion MRI tractography is a valuable tool for pre-surgical planning in brain tumour surgery (Essayed, 2017). However, achieving high anatomical precision in surgical tractography remains challenging due to pathology presence (Wasserthal, 2018). Modern techniques produce false-positive streamlines that need manual removal requiring neuroanatomy expertise (Dumais, 2023). We evaluated Purifibre (Aydogan, 2022), a tool to automatically identify and filter false-positive streamlines, in a paediatric brain tumour cohort to evaluate its potential to facilitate manual tractography process
Methods:
Fourteen brain tumour surgery patients were studied (8 males, median age 10.8 years [interquartile range, IQR 8-13], median tumour volume 81.89 cm³ [IQR 57.87-131.77]). Expert manual tractography was performed for pre-surgical planning.
3T MR used included 3D T1-weighted data (voxel-size=1.0mm), multi-shell diffusion MRI (dMRI) data acquired with isotropic 2.3 mm³ voxels, TE/TR = 77/3500ms,11 interleaved b0s and 2 diffusion-weighting (b=1000, 3000s/mm²), 30 and 60 directions, respectively. dMRI data were pre-processed with MRtrix3, FSL, ANTs, and underwent denoising, Gibbs ringing correction, motion and susceptibility distortion correction, and bias field correction
Tractography was performed using fibre orientation distribution (FOD) estimated from multi-shell multi-tissue constrained spherical deconvolution and the iFOD2 probabilistic algorithm. The tracking ROIs were manually delineated based on neuroanatomical priors. Per tract streamline numbers varied (2500, 7500, or 10000), with FOD cutoff set at 0.1 (n=173), 0.07 (n=38), or 0.03 (n=1), with lower values used to accommodate greater disruptions to FOD caused by tumour presence. A neuroanatomy expert created the manually-edited tractography set (~5-20 minutes editing time per tract). Fourteen pairs of tracts were reconstructed in both hemispheres, with 11 bilateral tracts in 8 cases (n=176), and 3 bilateral tracts in 6 cases (n=36), totalling 212 tracts
Purifibre (Aydogan, 2022) assigns a fibre-to-bundle-coupling (FICO) scalar to each streamline (Figure 1), which represents how well a streamline is aligned with the rest of the bundle. Low values suggest a potential false-positive, and are the first to be filtered out. The percentage of streamlines filtered out is determined by the purify parameter. All pre-manually edited tracts were filtered to produce Purifibre-edited tracts, first using the default 5% purify value, then with a range of purify values from 0-40%
Dice similarity coefficients (DSC) between manually-edited and Purifibre-edited tracts were calculated using Scipy (v1.10.1). Kruskal-Wallis test was used to compare DSC on tumour- versus non-tumour side tracts, and between different FOD cutoff values
3T MR used included 3D T1-weighted data (voxel-size=1.0mm), multi-shell diffusion MRI (dMRI) data acquired with isotropic 2.3 mm³ voxels, TE/TR = 77/3500ms,11 interleaved b0s and 2 diffusion-weighting (b=1000, 3000s/mm²), 30 and 60 directions, respectively. dMRI data were pre-processed with MRtrix3, FSL, ANTs, and underwent denoising, Gibbs ringing correction, motion and susceptibility distortion correction, and bias field correction
Tractography was performed using fibre orientation distribution (FOD) estimated from multi-shell multi-tissue constrained spherical deconvolution and the iFOD2 probabilistic algorithm. The tracking ROIs were manually delineated based on neuroanatomical priors. Per tract streamline numbers varied (2500, 7500, or 10000), with FOD cutoff set at 0.1 (n=173), 0.07 (n=38), or 0.03 (n=1), with lower values used to accommodate greater disruptions to FOD caused by tumour presence. A neuroanatomy expert created the manually-edited tractography set (~5-20 minutes editing time per tract). Fourteen pairs of tracts were reconstructed in both hemispheres, with 11 bilateral tracts in 8 cases (n=176), and 3 bilateral tracts in 6 cases (n=36), totalling 212 tracts
Purifibre (Aydogan, 2022) assigns a fibre-to-bundle-coupling (FICO) scalar to each streamline (Figure 1), which represents how well a streamline is aligned with the rest of the bundle. Low values suggest a potential false-positive, and are the first to be filtered out. The percentage of streamlines filtered out is determined by the purify parameter. All pre-manually edited tracts were filtered to produce Purifibre-edited tracts, first using the default 5% purify value, then with a range of purify values from 0-40%
Dice similarity coefficients (DSC) between manually-edited and Purifibre-edited tracts were calculated using Scipy (v1.10.1). Kruskal-Wallis test was used to compare DSC on tumour- versus non-tumour side tracts, and between different FOD cutoff values
Results:
With the default 5% purify parameter, a median DSC of 0.87 (IQR 0.83-0.96) was achieved across all tracts when comparing manually-edited and Purifibre-edited tracts. The median DSC was significantly lower for tumour-side tracts (0.83, IQR 0.80-0.96), compared to non-tumour side tracts (0.9, IQR 0.87-0.97; p=0.0299). Tumour-side tracts showed greater DSC variance, and this was tract-specific. Tracts reconstructed with a 0.07 FOD cutoff had significantly lower DSC than those with a 0.1 FOD cutoff (p < 0.001; Figure 2). Adjusting the percentage of filtered streamlines, from 0-40%, revealed tract- and FOD cutoff-specific DSC trends across all cases
Conclusions:
Our findings suggest that Purifibre effectively aids manual tractography in brain tumour cases, and can be fine tuned for improved performance. Lower DSC for the tumour-side tracts was likely due to tumour-distorted anatomy and altered peri-tumoural white matter properties, a persistent challenge in surgical tractography. The purify parameter can be adjusted to enhance tumour-side tracking results, although a threshold exists beyond which true-positive streamlines were removed
Modeling and Analysis Methods:
Diffusion MRI Modeling and Analysis 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
White Matter Anatomy, Fiber Pathways and Connectivity 2
Keywords:
Open-Source Code
PEDIATRIC
Tractography
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
Other - neurosurgery, brain tumour, automation
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Dumais, F. et al (2023): FIESTA: Autoencoders for accurate fiber segmentation in tractography. Neuroimage 279
Essayed, W. I. et al (2017): White matter tractography for neurosurgical planning: A topography-based review of the current state of the art. NeuroImage Clin. 15, 659–672
Wasserthal, J. et al (2018): TractSeg - Fast and accurate white matter tract segmentation. Neuroimage 183, 239–253