Application of Radiomics Methods for the Formation of a Feature Space in the Problem of Recognition of Cerebral Aneurysms
Abstract
The paper investigates the effectiveness of using radiomics methods to extract features from medical images when solving the problem of recognizing cerebral aneurysms using a graph neural network. The construction of a graph model of the vascular network based on the results of angiography of cerebral vessels allows you to move from the original image (an ordered sequence of voxel values) to a more structured representation of the original data, consisting of a set of local node descriptors and connections between them. This representation reflects the most significant features of the recognized objects and allows you to significantly reduce the amount of processed data. In this regard, one of the key aspects of successfully solving the problem of recognizing cerebral aneurysms based on a graph model of the vascular network is the formation of an informative feature description for graph nodes. In this paper, the characteristic description of graph nodes corresponds to image biomarkers – quantitative indicators characterizing various pathological changes – obtained using radiomics methods. A graph neural network is used as an algorithm for recognizing aneurysms. The paper considers the shape characteristics of the selected image areas, histogram and texture features, as well as some other characteristics. To identify the most informative features, a number of computational experiments were conducted using various sets of features. Based on the statistical analysis of the experimental results, it was concluded that the largest part of the information significant for the recognition of cerebral aneurysms is contained in the shape characteristics of the areas highlighted in the image, as well as in histogram features. However, the experimental results also showed that the constructed feature description is not sufficient for accurate recognition of cerebral aneurysms, and it requires expansion due to the introduction of new more complex indicators.
References
2. Sichtermann T., Faron A., Sijben R., Teichert N., Freiherr J., Wiesmann M. Deep Learning-Based Detection of Intracranial Aneurysms in 3D TOF-MRA. American Journal of Neuroradiology. 2019;40(1):25-32. doi: https://doi.org/10.3174/ajnr.A5911
3. Shi Z. et al. A clinically applicable deep-learning model for detecting intracranial aneurysm in computed tomography angiography images. Nature Communications. 2020;11(1):6090. doi: https://doi.org/10.1038/s41467-020-19527-w
4. Timmins K.M., van der Schaaf I.C., Vos I., Ruigrok Y.M., Velthuis B.K., Kuijf H.J. Deep learning with vessel surface meshes for intracranial aneurysm detection. Proceedings SPIE. Medical Imaging 2022: Computer-Aided Diagnosis. Vol. 12033. SPIE; 2022. Article number: 120332D. doi: https://doi.org/10.1117/12.2610745
5. Park A. et al. Deep Learning-Assisted Diagnosis of Cerebral Aneurysms Using the HeadXNet Model. JAMA Network Open. 2019;2(6):e195600. doi: https://doi.org/10.1001/jamanetworkopen.2019.5600
6. Meng C., Yang D., Chen D. Cerebral aneurysm image segmentation based on multi-modal convolutional neural network. Computer Methods and Programs in Biomedicine. 2021;208:106285. doi: https://www.doi.org/10.1016/j.cmpb.2021.106285
7. Li T. et al. Segmentation Method of Cerebral Aneurysms Based on Entropy Selection Strategy. Entropy. 2022;24(8):1062. doi: https://doi.org/10.3390/e24081062
8. Shahzad R. et al. Fully automated detection and segmentation of intracranial aneurysms in subarachnoid hemorrhage on CTA using deep learning. Scientific Reports. 2020;10(1):21799. doi: https://doi.org/10.1038/s41598-020-78384-1
9. Timmins K. M. et al. Comparing methods of detecting and segmenting unruptured intracranial aneurysms on TOF-MRAS: The ADAM challenge. Neuroimage. 2021;238:118216. doi: https://doi.org/10.1016/j.neuroimage.2021.118216
10. Liu Y., Liu J., Yuan Y. Edge-Oriented Point-Cloud Transformer for 3D Intracranial Aneurysm Segmentation. In: Wang L., Dou Q., Fletcher P.T., Speidel S., Li S. (eds.) Medical Image Computing and Computer Assisted Intervention – MICCAI 2022. MICCAI 2022. Lecture Notes in Computer Science. Vol. 13435. Cham: Springer; 2022. p. 97-106. doi: https://doi.org/10.1007/978-3-031-16443-9_10
11. Drees D. et al. Scalable robust graph and feature extraction for arbitrary vessel networks in large volumetric datasets. BMC Bioinformatics. 2021;22:346. doi: https://doi.org/10.1186/s12859-021-04262-w
12. Chenoune Y. et al. Three-dimensional segmentation and symbolic representation of cerebral vessels on 3DRA images of arteriovenous malformations. Computers in Biology and Medicine. 2019;115:103489. doi: https://www.doi.org/10.1016/j.compbiomed.2019.103489
13. Eulzer P. et al. Vessel Maps: A Survey of Map‐Like Visualizations of the Cardiovascular System. Computer Graphics Forum. 2022;41(3):645-673. doi: https://doi.org/10.1111/cgf.14576
14. Gori M., Monfardini G., Scarselli F. A new model for learning in graph domains. In: Proceedings 2005 IEEE International Joint Conference on Neural Networks. Vol. 2. Montreal, QC, Canada: IEEE Computer Society; 2005. p. 729-734. doi: https://doi.org/10.1109/IJCNN.2005.1555942
15. Cao W., Yan Z., He Z., He Z. A comprehensive survey on geometric deep learning. IEEE Access. 2020;8:35929-35949. doi: https://doi.org/10.1109/ACCESS.2020.2975067
16. Van Griethuysen J.J.M. et al. Computational Radiomics System to Decode the Radiographic Phenotype. Cancer Research. 2017;77(21):e104-e107. doi: https://doi.org/10.1158/0008-5472.CAN-17-0339
17. Scapicchio C. et al. A deep look into radiomics. La radiologia medica. 2021;126(10):1296-1311. doi: https://www.doi.org/10.1007/s11547-021-01389-x
18. Girardi D. et al. A skeleton-based SVM-supported cerebral aneurysm detection algorithm. In: European Congress of Radiology-ECR 2012. Article number: C-0361. ECR; 2012. doi: https://www.doi.org/10.1594/ecr2012/C-0361
19. Alwalid O. et al. CT Angiography-Based Radiomics for Classification of Intracranial Aneurysm Rupture. Frontiers in Neurology. 2021;12:619864. doi: https://doi.org/10.3389/fneur.2021.619864
20. Lauric A., Ludwig C.G., Malek A.M. Enhanced Radiomics for Prediction of Rupture Status in Cerebral Aneurysms. World Neurosurgery. 2022;159:8-22. doi: https://doi.org/10.1016/j.wneu.2021.11.038
21. Liu Q.L. et al. Prediction of Aneurysm Stability Using a Machine Learning Model Based on PyRadiomics-Derived Morphological Features. Stroke. 2019;50(9):2314-2321. doi: https://doi.org/10.1161/STROKEAHA.119.025777
22. Zack G.W., Rogers W.E., Latt S.A. Automatic measurement of sister chromatid exchange frequency. Journal of Histochemistry & Cytochemistry. 1977;25(7):741-753. doi: https://doi.org/10.1177/25.7.70454
23. Lee T.C., Kashyap R.L., Chu C.N. Building Skeleton Models via 3-D Medial Surface Axis Thinning Algorithms. CVGIP: Graphical Models and Image Processing. 1994;56(6):462-478. doi: https://doi.org/10.1006/cgip.1994.1042
24. Kipf T.N., Welling M. Semi-Supervised Classification with Graph Convolutional Networks. In: Proceedings of the 5th International Conference on Learning Representations (ICLR 2017). Toulon, France; 2017. p. 1-14. Available at: https://openreview.net/forum?id=SJU4ayYgl (accessed 29.08.2022).
25. Van der Maaten L., Hinton G. Visualizing Data using t-SNE. Journal of Machine Learning Research. 2008;9(86):2579-2605. Available at: https://www.jmlr.org/papers/v9/vandermaaten08a.html (accessed 29.08.2022).
This work is licensed under a Creative Commons Attribution 4.0 International License.
Publication policy of the journal is based on traditional ethical principles of the Russian scientific periodicals and is built in terms of ethical norms of editors and publishers work stated in Code of Conduct and Best Practice Guidelines for Journal Editors and Code of Conduct for Journal Publishers, developed by the Committee on Publication Ethics (COPE). In the course of publishing editorial board of the journal is led by international rules for copyright protection, statutory regulations of the Russian Federation as well as international standards of publishing.
Authors publishing articles in this journal agree to the following: They retain copyright and grant the journal right of first publication of the work, which is automatically licensed under the Creative Commons Attribution License (CC BY license). Users can use, reuse and build upon the material published in this journal provided that such uses are fully attributed.
