Classification of Ventricular Septal Defect Disease Using Deep Learning

Öz

Ventricular Septal Defect (VSD) disease is the most prevalent type of congenital heart disease. VSD is a hole between the left and right ventricles of the heart structure. VSD disease accounts for approximately one-fifth of all congenital heart disease types. Therefore, accurate disease diagnosis is paramount in determining the most appropriate treatment methods. This study aims to classify VSD disease using the deep learning algorithms VGG16, ResNET50, and Inceptionv3 on Computed Tomography (CT) images and compare the pre-trained algorithms used. One of the reasons why imaging methods such as echocardiog raphy are generally used to detect congenital heart diseases is that there are almost no CT datasets related to this disease. The dataset used in this study is the ImageCHD dataset, which comprises 3D CT scans encompassing 16 distinct types of congenital heart defects. Hyperparameter optimization was performed using the grid search method to enhance the model performance, identifying the VGG16 model as the most effective. The model demonstrated a very high classification accuracy of 99.99% in the training dataset and 99.94% in the test dataset. Gradient-weighted Class Activation Mapping was employed to enhance model explainability, providing visualizations of the regions most critical for the classification, thereby enabling medical professionals to validate AI-driven predictions. An optimized model that successfully classifies VSD using 3D CT image data has been introduced to the literature for the first time. Therefore, this study assumes greater significance in the existing literature and sets a benchmark for future studies.

Anahtar Kelimeler

• Computed tomography
• deep learning
• medical image classification
• transfer learning
• Ventricular Septal Defect disease

Kaynakça

1. Abbas, S., Ojo, S., Al Hejaili, A., Sampedro, G. A., Almadhor, A., Zaidi, M. M., & Kryvinska, N. (2024). Artificial intelligence framework for heart disease classification from audio signals. Scientific Reports, 14 (i), 3123. doi:10.1038/s41598-024-53778-7 google scholar
2. Arslan, N. N., & Ozdemir, D. (2024). Analysis of CNN models in classifying Alzheimer’s stages: comparison and explainability examination of the proposed separable convolution-based neural network and transfer learning models. Signal, Image and Video Processing, 18 (SI), 447-461. DOI: 10.1007/s11760-024-03166-5 google scholar
3. Aziz, S., Khan, M. U., Alhaisoni, M., Akram, T., & Altaf, M. (2020). Phonocardiogram signal processing for the automatic diagnosis of congenital heart disorders through the fusion of temporal and cepstral features. Sensors, 20(13), 3790. doi:10.3390/s20133790 google scholar
4. Bernier, P. L., Stefanescu, A., Samoukovic, G., & Tchervenkov, C. I. (2010). The challenge of congenital heart disease worldwide: epidemi-ologic and demographic facts. Seminars in Thoracic and Cardiovascular Surgery: Pediatric Cardiac Surgery Annual, 13(1), 26-34. https://doi.org/10.1053/j.pcsu.2010.02.005 google scholar
5. Deng, J., Dong, W., Socher, R., Li, L.-J., Kai Li, & Li Fei-Fei. (2009). ImageNet: A large-scale hierarchical image database. İn 2009 IEEE Conference on Computer Vision and Pattern Recognition (pp. 248-255). IEEE. doi: 10.1109/CVPR.2009.5206848 google scholar
6. Dillman, J. R., & Hernandez, R. J. (2009). Role of CT in the evaluation of congenital cardiovascular disease in children. American Journal of Roentgenology, 192(5), 1219-1231. doi: 10.2214/AJR.09.2382 google scholar
7. Dörterler, S., Dumlu, H., Özdemir, D., & Temurtaş, H. (2024). Hybridization of meta-heuristic algorithms with k-means for clustering analysis: Case of medical datasets. Gazi Journal of Engineering Sciences, 10(1), 1-11. https://doi.org/10.30855/gmbd.0705N01 google scholar
8. Geva, T., Martins, J. D., & Wald, R. M. (2014). Atrial septal defects. The Lancet, 383(9932), 1921-1932. https://doi.org/l0.1016/S0140-6736Cl3) 62145-5 google scholar

9. Hallıoğlu, O., Karpuz, D., Giray, D., Demetgül, H., & Öztaş, A. (2018). Frequency of the congenital heart diseases according to the risk groups: Fetal echocardiographic screening. The Journal of Gynecology-Obstetrics and Neonatology, 15(1), 1-4. google scholar
10. Hassani, K., Jafarian, K., & Doyle, D. J. (2017). Heart sounds feature used for the classification of ventricular septal defect size in children. İn IFMBE Proceedings (Vol. 61, pp. 28-31). Springer Verlag. doi:10.1007/978-981-10-4220-1_6 google scholar
11. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 770-778). IEEE. doi: 10.1109/CVPR.2016.90 google scholar
12. Kara, İ. (2023). Query by image examination: Classification of digital image-based forensics using deep learning methods. Açta Infologica, 7(2), 348-359. doi:10.26650/acin.1282567 google scholar
13. Koundinya, T. S., Sudhanva, S., Shashank, C. S., Chandrasekar, S., & Munavalli, J. R. (2023). Congenital heart disease detection using spectral analysis and CNN. In 2023, 4th International Conference for Emerging Technology, INCET 2023. Institute of Electrical and Electronics Engineers Inc. https://doi.org/10.1109/INCET57972.2023.10170429 google scholar
14. Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017). ImageNet classification with deep convolutional neural networks. Communications of the ACM, 60(6), 84-90. https://doi.org/10.1145/3065386 google scholar
15. Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324. doi: 10.1109/5.726791 google scholar
16. Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., ... Fei-Fei, L. (2015). ImageNet large scale visual recognition challenge. International Journal of Computer Vision, 115(3), 211-252. doi:10.1007/s11263-015-0816-y google scholar
17. Sapitri, A. I., Nurmaini, S., Sukemi, Rachmatullah, M. N., & Darmawahyuni, A. (2021). Segmentation atrioventricular septal defect by using convolutional neural networks based on U-NET architecture. IAES International Journal of Artificial Intelligence, 10(3), 553-562. doi: 10.11591/ijai.v10.i3.pp553-562 google scholar
18. Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017). Grad-CAM: visual explanations from deep networks via gradient-based localization. İn 2017 IEEE International Conference on Computer Vision (ICCV) (pp. 618-626). IEEE. doi: 10.1109/ ICCV.2017.74 google scholar
19. Şevli, O. (2023). A deep learning-based classification study for diagnosing corneal ulcers on ocular staining images. Açta Infologica, 7(2), 281-292. https://doi.org/l0.26650/acin.1173465 google scholar
20. Simonyan, K., & Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition. 3rd International Conference on Learning Representations (ICLR 2015), 1-14. Retrieved from http://arxiv.org/abs/1409.1556 google scholar
21. Sun, S., Jiang, Z., Wang, H., Fang, Y., & Tao, T. (2012). Heart sound feature parameter distribution and support vector machine-based classification boundary determination method for ventricular septal defect auscultation. Journal of Computational Science and Technology, 6(3), 198-206. doi:10.1299/jcst.6.198 google scholar
22. Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016). Rethinking the Inception architecture for computer vision. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 2818-2826). IEEE. https://doi.org/10.1109/CVPR.2016.308 google scholar
23. Tan, C., Sun, F., Kong, T., Zhang, W., Yang, C., & Liu, C. (2018). A survey on deep transfer learning. The 27th International Conference on Artificial Neural Networks (ICANN 2018). Retrieved from http://arxiv.org/abs/1808.01974 google scholar
24. Tay Lik Wui, E., Yip, J. W. L., & Li, W. (2011). Echocardiography. In Diagnosis and Management of Adult Congenital Heart Disease (pp. 2843). Elsevier. https://doi.org/l0.1016/B978-0-7020-3426-8.00005-8 google scholar
25. Tüzün, B. N., & Özdemir, D. (2023). Classification of brain tumors using deep learning models. Journal of Scientific Reports-A, 054, 296306. doi: 10.59313/jsr-a.1293119 google scholar
26. Wang, J.-K., Chang, Y.-F., Tsai, K.-H., Wang, W.-C., Tsai, C.-Y., Cheng, C.-H., & Tsao, Y. (2020). Automatic recognition of ventricular septal defect murmurs using convolutional recurrent neural networks with temporal attentive pooling. Scientific Reports, 10(1), 21797. doi:10.1038/s41598-020-77994-z google scholar
27. Xu, X., Wang, T., Zhuang, J., Yuan, H., Huang, M., Cen, J., ... Shi, Y. (2021). ImageCHD: A 3D computed tomography image dataset for classification of congenital heart disease. Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data, 4, 77-87. Retrieved from http://arxiv.org/abs/2101.10799 google scholar
28. Y amashita, R., Nishio, M., Do, R. K. G., & Togashi, K. (2018). Convolutional neural networks: an overview and application in radiology. Insights into Imaging, 9(4), 611-629. doi:10.1007/s13244-018-0639-9 google scholar
29. Y ang, Y., Wu, B., Wu, H., Xu, W., Lyu, G., Liu, P., & He, S. (2023). Classification of normal and abnormal fetal heart ultrasound images and identification of ventricular septal defects based on deep learning. Journal of Perinatal Medicine, 51(8), 1052-1058. doi:10.1515/ jpm-2023-0041 google scholar
30. Y ıldız, J., Çetin, İ. İ., Aktaş, D., Arı, M. E., Kocabaş, A., Ekici, F., & Şaylı, T. R. (2015). Is an echocardiographic evaluation necessary for all children with a cardiac murmur? Turkish Journal of Pediatric Disease. doi:10.12956/tjpd.2015.158 google scholar