Machine learning-based classification of chronic traumatic brain injury using hybrid diffusion imaging

Jennifer J Muller, Ruixuan Wang, Devon Milddleton, Mahdi Alizadeh, Ki Chang Kang, Ryan Hryczyk, George Zabrecky, Chloe Hriso, Emily Navarreto, Nancy Wintering, Anthony J Bazzan, Chengyuan Wu, Daniel A Monti, Xun Jiao, Qianhong Wu, Andrew B Newberg, Feroze B Mohamed, Jennifer J Muller, Ruixuan Wang, Devon Milddleton, Mahdi Alizadeh, Ki Chang Kang, Ryan Hryczyk, George Zabrecky, Chloe Hriso, Emily Navarreto, Nancy Wintering, Anthony J Bazzan, Chengyuan Wu, Daniel A Monti, Xun Jiao, Qianhong Wu, Andrew B Newberg, Feroze B Mohamed

Abstract

Background and purpose: Traumatic brain injury (TBI) can cause progressive neuropathology that leads to chronic impairments, creating a need for biomarkers to detect and monitor this condition to improve outcomes. This study aimed to analyze the ability of data-driven analysis of diffusion tensor imaging (DTI) and neurite orientation dispersion imaging (NODDI) to develop biomarkers to infer symptom severity and determine whether they outperform conventional T1-weighted imaging.

Materials and methods: A machine learning-based model was developed using a dataset of hybrid diffusion imaging of patients with chronic traumatic brain injury. We first extracted the useful features from the hybrid diffusion imaging (HYDI) data and then used supervised learning algorithms to classify the outcome of TBI. We developed three models based on DTI, NODDI, and T1-weighted imaging, and we compared the accuracy results across different models.

Results: Compared with the conventional T1-weighted imaging-based classification with an accuracy of 51.7-56.8%, our machine learning-based models achieved significantly better results with DTI-based models at 58.7-73.0% accuracy and NODDI with an accuracy of 64.0-72.3%.

Conclusion: The machine learning-based feature selection and classification algorithm based on hybrid diffusion features significantly outperform conventional T1-weighted imaging. The results suggest that advanced algorithms can be developed for inferring symptoms of chronic brain injury using feature selection and diffusion-weighted imaging.

Keywords: diffusion tensor imaging (DTI); hybrid diffusion imaging; machine learning; neurite orientation dispersion and density imaging (NODDI); traumatic brain injury.

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2023 Muller, Wang, Milddleton, Alizadeh, Kang, Hryczyk, Zabrecky, Hriso, Navarreto, Wintering, Bazzan, Wu, Monti, Jiao, Wu, Newberg and Mohamed.

Figures

Figure 1
Figure 1
Flow chart representing machine learning (ML) approach for cTBI classification. Step 1 consists of feature extraction including image acquisition, preprocessing, image normalization and skeletonization using TBSS, and atlas registration. The dataset was divided into training and test datasets for K-fold CV, calculating the mean accuracy of each model.
Figure 2
Figure 2
Heatmap showing mean accuracy performance of different ML algorithms for a trail making A and B. All 20 JHU atlas regions are used as features for the above figure. Mean accuracy results based on T1 inferences are highlighted in red and are expressed in percentages.
Figure 3
Figure 3
Feature ranking results for DTI (A), NODDI (B), and T1 (C) regions. Features are displayed if they were ranked as significant for both trail making A and B. Results of 6-feature KNN are displayed in light blue, compared with 20-feature KNN results in dark blue (D).

References

    1. Anguita D., Ghelardoni L., Ghio A., Oneto L., Ridella S. (2022). The ‘K' in K-fold Cross Validation, Eds G. Balint, Antala B, Carty C, Mabieme JMA, Amar IB, Kaplanova A (Uniwersytet ślaski), 441–446.
    1. Chong S. L., Liu N., Barbier S., Ong M. E. H. (2015). Predictive modeling in pediatric traumatic brain injury using machine learning data analysis, statistics and modelling. BMC Med. Res. Methodol. 15, 1–9. 10.1186/s12874-015-0015-0
    1. Daugherty J., Zhou H., Sarmiento K., Waltzman D. (2016). Differences in state traumatic brain injury-related deaths, by principal mechanism of injury, intent, and percentage of population living in rural areas-United States, 2016-2018. MMWR. 15.
    1. Davatzikos C. (2019). Machine learning in neuroimaging: progress and challenges. Neuroimage. 197, 652. 10.1016/j.neuroimage.2018.10.003
    1. Douglas D. B., Iv M., Douglas P. K., Vos S. B., Bammer R., Zeineh M., et al. . (2015). Diffusion tensor imaging of TBI: potentials and challenges HHS public access. Top. Magn. Reson. Imaging. 24, 241–251. 10.1097/RMR.0000000000000062
    1. Grabmeier J. L., Lambe L. A. (2007). Decision trees for binary classification variables grow equally with the Gini impurity measure and Pearson's chi-square testInt. J. Bus. Intell. Data Min. 2, 213–226. 10.1504/IJBIDM.2007.013938
    1. Hashim E., Caverzasi E., Papinutto N., Lewis C. E., Jing R., Charles O., et al. . (2017). Investigating microstructural abnormalities and neurocognition in sub-acute and chronic traumatic brain injury patients with normal-appearing white matter: a preliminary diffusion tensor imaging study. Front. Neurol. 8, 97. 10.3389/fneur.2017.00097
    1. Hu L., Yang S., Jin B., Wang C. (2022). Advanced neuroimaging role in traumatic brain injury: a narrative review. Front. Neurosci. 16, 453. 10.3389/fnins.2022.872609
    1. Hu L. Y., Huang M. W., Ke S. W., Tsai C. F. (2016). The distance function effect on k-nearest neighbor classification for medical datasets. Springerplus. 5, 1–9. 10.1186/s40064-016-2941-7
    1. Jenkinson M., Beckmann C. F., Behrens T. E. J., Woolrich M. W., Smith S. M. (2012). Review FSL. Neuroimage. 62, 782–790. 10.1016/j.neuroimage.2011.09.015
    1. Jost L. (2006). Entropy and diversity. Oikos. 113, 363–375. 10.1111/j.2006.0030-1299.14714.x
    1. Kamiya K., Hori M., Aoki S. (2020). NODDI in clinical research. J. Neurosci. Methods. 346, 108908. 10.1016/j.jneumeth.2020.108908
    1. Kramer O. (2016). “Scikit-Learn” in Machine Learning for Evolution Strategies, 45–53. 10.1007/978-3-319-33383-0_5
    1. Kraus M. F., Susmaras T., Caughlin B. P., Walker C. J., Sweeney J. A., Little D. M., et al. . (2007). White matter integrity and cognition in chronic traumatic brain injury: a diffusion tensor imaging study. Brain. 130, 2508–2519. 10.1093/brain/awm216
    1. Langs G., Menze B. H., Lashkari D., Golland P. (2011). Detecting stable distributed patterns of brain activation using Gini CONTRAST. Neuroimage. 56, 497–507. 10.1016/j.neuroimage.2010.07.074
    1. Lao Z., Shen D., Xue Z., Karacali B., Resnick S. M., Davatzikos C., et al. . (2004). Morphological classification of brains via high-dimensional shape transformations and machine learning methods. Neuroimage. 21, 46–57. 10.1016/j.neuroimage.2003.09.027
    1. Malec J. F., Brown A. W., Leibson C. L., Flaada J. T., Mandrekar J. N., Diehl N. N., et al. . (2007). The mayo classification system for traumatic brain injury severity. J. Neurotrauma. 24, 1417–1424.
    1. McCrory P., Meeuwisse W., Johnston K., Dvorak J., Aubry M., Molloy M., et al. . (2009). Consensus statement on concussion in sport: the 3rd international conference on concussion in sport held in Zurich, november 2008. Br. J. Sports Med. 43, i76–i84. 10.1136/bjsm.2009.058248
    1. Mckee A. C., Daneshvar D. H. (2015). The neuropathology of traumatic brain injury. Handb. Clin. Neurol. 127, 45–66.
    1. Minaee S., Wang Y., Chung S., Wang X., Fieremans E., Flanagan S., et al. . (2017). A machine learning approach for identifying patients with mild traumatic brain injury using diffusion MRI modeling. arXiv Preprint. arXiv:1708.09000.
    1. Mohamed M., Mohamed M., Khalid N. (2022). Prognosticating outcome using magnetic resonance imaging in patients with moderate to severe traumatic brain injury: a machine learning approach.
    1. Muller J., Middleton D., Alizadeh M., Zabrecky G., Wintering N. A., Bazzan A. J., et al. . (2021). Hybrid diffusion imaging reveals altered white matter tract integrity and associations with symptoms and cognitive dysfunction in chronic traumatic brain injury. Neuroimage. Clin. 30, 102681. 10.1016/j.nicl.2021.102681
    1. Musavi M. T., Ahmed W., Chan K. H., Faris K. B., Hummels D. M. (1992). On the training of radial basis function classifiers. Neural. Networks. 5, 595–603. 10.1016/S0893-6080(05)80038-3
    1. Mutasa S., Sun S., Ha R. (2020). Understanding artificial intelligence based radiology studies: what is overfitting? Clin. Imaging. 65, 96–99. 10.1016/j.clinimag.2020.04.025
    1. Myles A. J., Feudale R. N., Liu Y., Woody N. A., Brown S. D. (2004). An introduction to decision tree modeling. J. Chemom. 18, 275–285. 10.1002/cem.873
    1. National Academies of Sciences Engineering, and Medicine; Health and Medicine Division; Board on Health Care Services; Committee on the Review of the Department of Veterans Affairs Examinations for Traumatic Brain Injury. (2019). Evaluation of the Disability Determination Process for Traumatic Brain Injury in Veterans. Washington (DC): National Academies Press.
    1. Palacios E. M., Owen J. P., Yuh E. L., Vassar M. J., Ferguson A. R., Diaz-arrastia R., et al. . (2020). The evolution of white matter microstructural changes after mild traumatic brain injury: a longitudinal DTI and NODDI study. Sci. Adv. 6, aaz6892. 10.1126/sciadv.aaz6892
    1. Palczewska A., Palczewski J., Robinson R. M., Neagu D. (2014). Interpreting random forest classification models using a feature contribution method. Adv. Intell. Syst. Comput. 263, 193–218. 10.1007/978-3-319-04717-1_9
    1. Pease M., Arefan D., Barber J., Yuh E., Puccio A., Hochberger K., et al. . (2022). Outcome prediction in patients with severe traumatic brain injury using deep learning from head CT scans. Radiology. 304, 385–394.
    1. Pontil M., Verri A. (1998). Properties of support vector machines. Neural. Comput. 10, 955–974. 10.1162/089976698300017575
    1. Prasuhn J., Heldmann M., Münte T. F., Brüggemann N. (2020). A machine learning-based classification approach on Parkinson's disease diffusion tensor imaging datasets. Neurol. Res. Pract. 2, 1–5. 10.1186/s42466-020-00092-y
    1. Qu Y., Wang P., Liu B., Song C., Wang D., Yang H., et al. . (2021). AI4AD: Artificial intelligence analysis for Alzheimer's disease classification based on a multisite DTI database. Brain Disord. 1, 100005. 10.1016/j.dscb.2021.100005
    1. Rafało M. (2022). Cross validation methods: analysis based on diagnostics of thyroid cancer metastasis. ICT Express. 8, 183–188. 10.1016/j.icte.2021.05.001
    1. Rätsch G., Warmuth M. K. K., Glocer K. (2007). Boosting algorithms for maximizing the soft margin. Adv Neural Inf Process Syst. 20.
    1. Razzak M. I., Naz S., Zaib A. (2018). Deep learning for medical image processing: overview, challenges and the future BT-classification in BioApps: automation of decision making. Springer. 26, 323–350. 10.1007/978-3-319-65981-7_12
    1. Refaeilzadeh P., Tang L., Liu H. (2016). “Cross-validation” in Encyclopedia of Database Systems (New York, NY: Springer New York; ), 1–7. 10.1007/978-1-4899-7993-3_565-2
    1. Rodriguez-Galiano V. F., Ghimire B., Rogan J., Chica-Olmo M., Rigol-Sanchez J. P. (2012). An assessment of the effectiveness of a random forest classifier for land-cover classification. ISPRS J. Photogramm. Remote Sens. 67, 93–104. 10.1016/j.isprsjprs.2011.11.002
    1. Smith L. G. F., Milliron E., Ho M. L., Hu HH, Rusin J, Leonard J, et al. . (2019). Advanced neuroimaging in traumatic brain injury: an overview. Neurosurg. Focus. 47, E17. 10.3171/2019.9.FOCUS19652
    1. Suthaharan S. (2016). “Support vector machine.” in Machine Learning Models and Algorithms for Big Data Classification, 207–235. 10.1007/978-1-4899-7641-3_9
    1. Taylor D. D., Gercel-Taylor C. (2014). Exosome platform for diagnosis and monitoring of traumatic brain injury. Philos. Trans. R. Soc. B: Biol. Sci. 369, 20130503.
    1. Timmers I., Roebroeck A., Bastiani M., Jansma B., Rubio-Gozalbo E., Zhang H., et al. . (2016). Assessing microstructural substrates of white matter abnormalities: A Comparative study using DTI and NODDI. PLoS ONE. 11, 1–15. 10.1371/journal.pone.0167884
    1. Vergara V. M., Mayer A. R, Kiehl K. A. (2017). Detection of mild traumatic brain injury by machine learning classification using resting state functional network connectivity and fractional anisotropy. J. Neurotrauma.
    1. Wickwire E. M., Williams S. G., Roth T., Capaldi V. F., Jaffe M., Moline M., et al. . (2016). Sleep, sleep disorders, and mild traumatic brain injury. What we know and what we need to know: findings from a national working group. Neurotherapeutics. 13, 403–417. 10.1007/s13311-016-0429-3
    1. Wu Y. C., Alexander A. L. (2007). Hybrid diffusion imaging. Neuroimage. 36, 617. 10.1016/j.neuroimage.2007.02.050
    1. Wu Y. C., Mustafi S. M., Harezlak J., Kodiweera C., Flashman L. A., Mcallister T. W., et al. . (2018). Hybrid diffusion imaging in mild traumatic brain injury. J. Neurotrauma. 35, 2377–2390. 10.1089/neu.2017.5566
    1. Zhang H., Schneider T., Wheeler-Kingshott C. A., Alexander D. C. (2012). NODDI practical in vivo neurite orientation dispersion and density imaging of the human brain. Neuroimage. 61, 1000–1016. 10.1016/j.neuroimage.2012.03.072
    1. Zhang T., Du C., Wang J. (2022). Composite Quantization for Approximate Nearest Neighbor Search, 838–846.Available online at: (accessed August 31, 2022).

Source: PubMed

Patrocinadores y Colaboradores

Condiciones médicas

Intervenciones de drogas

3
Suscribir