Saratov JOURNAL of Medical and Scientific Research

Application of a neural network to restore the lost surface of skull bones

Summary:

Abstract. Objective: to evaluate the sensitivity, specificity and accuracy of a digital algorithm based on convo-lutional neural networks to restore of bones of cranium defects. Material and methods. Neural network training was carried out as a result of 6,000 epochs on 78,000 variants of skull models with artificially generated skull injuries. The evaluation was performed on 222 DICOM series of patients computerized tomography with bones of cranium defects. Results. The indicators of sensitivity, specificity and accuracy were 95.3%, 85.5% and 79.4% respectively. A number of experiments were carried out with step-by-step sorting of three-dimensional models in order to find the reasons for the unsatisfactory skull reconstructing results. Incorrect detection of the skull defect most often occurred in the area of the facial skeleton. After excluding the series with artifacts, the average increase in metrics was 2.6%. Conclusion. Correct determination of the bone defect at the scull model (specificity) by the algorithm had the greatest impact on the surface accuracy. The maximum accuracy of the algorithm, which allows using the obtained surfaces without additional processing in a three-dimensional modeling environment, was achieved on series without the presence of artifacts during computed tomography (83.5%), as well as with defects that do not extend to the skull base (79.5%).

Bibliography:
1. Mishinov SV, Stupak VV, Panchenko AA, Krasovskiy IB. Reconstruction of frontomalarorbital region with use of the individual titan implant developed with direct laser sintering with 3D printer. Clinical case. Russian Polenov Neurosurgical Journal. 2017; 9 (1): 80-2.
2. Okishev DN, Cherebylo SA, Konovalov AN, et al. Features of modeling a polymer implant for closing a defect after decompressive craniotomy. Voprosy Neirokhirurgii Imeni N.N. Burdenko. 2022; 86 (1): 17-27.
3. Bratsev IS, Smetanina OV, Yashin KS, et al. Cranioplasty of post-trepanation skull defects using additive 3D printing technologies. Neyrokhirurgiya = Russian Journal of Neurosurgery 2021; 23 (2): 34-43.
4. Ivanov OV. Simultaneous repair of the skull base and the frontal lobe defect using CAD-CAM technology. Extreme Medicine. 2021; 23 (4): 72-7.
5. Zhang Q, Xu Y, Zhou J, et al. Neural network-based repairing skull defects: an initial assessment of performance and feasibility. Journal of Mechanics in Medicine and Biology. 2021; 21 (5): 2140012. DOI: 10.1142/s0219519421400121.
6. Wodzinski M, Daniol M, Socha M, et al. Deep learning-based framework for automatic cranial defect reconstruction and implant modeling. Comput Methods Programs Biomed. 2022; (226): 107173. DOI: 10.1016/j.cmpb.2022.107173.
7. Masouleh MK, Sadeghian S. Deep learning-based method for reconstructing three-dimensional building cadastre models from aerial images. Journal of Applied Remote Sensing. 2019; 13(02): 1. DOI: 10.1117/1.JRS.13.024508.
8. Morais A, Egger J, Alves V. Automated computer-aided design of cranial implants using a deep volumetric convolutional denoising autoencoder. In: Rocha A, Adeli H, Reis L, Costanzo S. (eds). New knowledge in information systems and technologies. Advances in Intelligent Systems and Computing. Springer, Cham. 2019; 932 p. DOI: 10.1007/978-3-030-16187-3J5.
9. Matzkin F, Newcombe V, Glocker B, et al. Cranial implant design via virtual craniectomy with shape priors. arXiv: 2009.13704 [eess. IV]. DOI: 10.48550/arXiv.2009.13704.
10. Kodym O, Spanel M, Herout A. Skull shape reconstruction using cascaded convolutional networks. Computers in Biology and Medicine. 2020; (123): 103886. DOI: 10.1016/j.comp-biomed.2020.103886.
11. Li J, von Campe G, Pepe A, et al. Automatic skull defect restoration and cranial implant generation for cranioplasty. Medical Image Analysis. 2021; (73): 102171. DOI: 10.1016/j.media. 2021.102171.
12. Wu CT, Yang YH, Chang YZ. Three-dimensional deep learning to automatically generate cranial implant geometry. Scientific Reports. 2022; 12 (1): 2683. DOI: 10.1038/S41598-022-06606-9.

AttachmentSize
2023_01_34-40.pdf513.59 KB

No votes yet