Clinically-optimized White Matter Hyperintensity Segmentation with BrainKey
Poster No:
1654
Submission Type:
Abstract Submission
Authors:
Brady Williamson1, Owen Phillips2, Nathan Strong2, Kevin Aquino2
Institutions:
1University of Cincinnati, Cincinnati, OH, 2BrainKey, San Francisco, CA
First Author:
Co-Author(s):
Introduction:
Background: White matter hyperintensities (WMH) are a crucial marker of cerebral small vessel disease (CSVD) and are associated with several poor neurological outcomes, including stroke and vascular dementia (Meng et al., 2022). However, WMH are heterogeneous with a lack of systematic quantification leading to suboptimal predictive power (Chen et al., 2021). Quantification of WMH burden involves time-consuming and observer-dependent segmentation (Kuijf et al., 2019). While there are currently some promising automated approaches, many of these are developed with high quality, low artefact, and parameter matched images if there are multiple sites. There is still a crucial need for methods with broad generalizability for use in clinical settings. We aimed to overcome this limitation by developing a WMH segmentation platform that is specifically optimized for clinical use by using scans across sites, scanners, parameters, and quality.
Methods:
Methods: Preprocessing: MRI scans are converted to NIFTI and preprocessed to harmonize inputs into the KeyLayer WMH algorithm. T1-, T2- , and T2-FLAIR weighted images are used as input. T1 preprocessing includes image alignment, model-based bias correction using ITK-4 (Tustison et al., 2010), and non-local averaging denoising (Manjón et al., 2010). T2-FLAIR and T2 preprocessing includes alignment to the preprocessed T1 image. Training data: Sources of data are inhouse scans and submissions sent to BrainKey from various scanning sites, including international sites, with a wide range of ages, sex and demographics, ensuring training, validation, and hold-out test data is varied and un-biased to any combination of features. These data were labelled by 8 in house experts, giving further variation of quality to training. White Matter Hyperintensities Segmentation: An encoder-decoder style machine learning algorithm is used to simultaneously segment lesions from aligned T1, T2, and FLAIR inputs. Volumes, presented as raw values and volume percent within a tissue region, are extracted from the resulting WMH segmentation. Validation dataset: Results are validated with a hold-out test set of 30 scans labelled by external expert.
Results:
Results: We found high reliability with BrainKey's algorithm internally across a set of input images. A key finding was the ability to segment poor T2-FLAIR image quality (Fig. 1) validated by external experts across our test set. BrainKey is able a range of WMH loads from punctate lesions (Fig. 1) up to larger lesion loads (Fig.2).
Conclusions:
Discussion: BrainKey is developed to overcome this limitation by developing a WMH segmentation platform that is specifically optimized for clinical use by using scans across sites, scanners, parameters, and quality. Notably, our approach was able to segment small punctate WMH lesions that are difficult to capture across varying scan quality and lesion load.
Modeling and Analysis Methods:
Segmentation and Parcellation 1
Neuroinformatics and Data Sharing:
Informatics Other 2
Keywords:
Data analysis
Machine Learning
MRI
White Matter
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?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
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.
2. Kuijf, H. J., Biesbroek, J. M., Bresser, J. D., Heinen, R., Andermatt, S., Bento, M., … Biessels, G. J. (2019). Standardized Assessment of Automatic Segmentation of White Matter Hyperintensities and Results of the WMH Segmentation Challenge. IEEE Transactions on Medical Imaging, 38(11), 2556–2568. doi: 10.1109/tmi.2019.2905770
3. Manjón, J. V., Coupé, P., Martí-Bonmatí, L., Collins, D. L., & Robles, M. (2010). Adaptive non-local means denoising of MR images with spatially varying noise levels. Journal of Magnetic Resonance Imaging, 31(1), 192–203. doi: 10.1002/jmri.22003
4. Meng, F., Yang, Y., & Jin, G. (2022). Research Progress on MRI for White Matter Hyperintensity of Presumed Vascular Origin and Cognitive Impairment. Frontiers in Neurology, 13, 865920. doi: 10.3389/fneur.2022.865920
5. Tustison, N. J., Avants, B. B., Cook, P. A., Zheng, Y., Egan, A., Yushkevich, P. A., & Gee, J. C. (2010). N4ITK: Improved N3 Bias Correction. IEEE Transactions on Medical Imaging, 29(6), 1310–1320. doi: 10.1109/tmi.2010.2046908