Journal of Korea Society of Industrial Information Systems (한국산업정보학회논문지)

Korea Society of Industrial Information Systems (한국산업정보학회)
권호 표지
DOI QR코드

DOI QR Code

An Integrated and Complementary Evaluation System for Judging the Severity of Knee Osteoarthritis Using CNN

CNN 기반 슬관절 골관절염 중증도 판단을 위한 통합 보완된 등급 판정 시스템

  • 윤예찬 (고려대학교 산업경영공학과(산업인공지능))
  • Received : 2024.04.19
  • Accepted : 2024.08.15
  • Published : 2024.08.30
Download PDF
⟨ Previous Next ⟩

Abstract

Knee osteoarthritis (OA) is a very common musculoskeletal disorder worldwide. The assessment of knee osteoarthritis, which requires a rapid and accurate initial diagnosis, is determined to be different depending on the currently dispersed classification system, and each classification system has different criteria. Also, because the medical staff directly sees and reads the X-ray pictures, it depends on the subjective opinion of the medical staff, and it takes time to establish an accurate diagnosis and a clear treatment plan. Therefore, in this study, we designed the stenosis length measurement algorithm and Osteophyte detection and length measurement algorithm, which are the criteria for determining the knee osteoarthritis grade, separately using CNN, which is a deep learning technique. In addition, we would like to create a grading system that integrates and complements the existing classification system and show results that match the judgments of actual medical staff. Based on publicly available OAI (Osteoarthritis Initiative) data, a total of 9,786 knee osteoarthritis data were used in this study, eventually achieving an Accuracy of 69.8% and an F1 score of 76.65%.

슬관절 골관절염(OA, Osteoarthritis)은 전 세계적으로 매우 흔한 근골격계 질환이다. 빠르고 정확한 초기 진단이 필요한 슬관절 골관절염의 등급은 현재 분산된 분류 시스템에 따라 다르게 판정되며, 각 분류 시스템마다 기준이 상이하다. 또한 의료진이 X-ray 사진을 직접 보고 판독하기 때문에 의료진의 주관적인 의견에 따라 달라지며 시간이 많이 소요되어 정확한 진단과 명확한 치료 계획 수립에 시간이 지연되고 있다. 따라서 본 연구는 딥러닝 기술인 CNN을 사용하여 슬관절 골관절염 등급 판단 기준이 되는 협착 부분의 길이 측정 알고리즘과 골극의 탐지 및 길이 측정 알고리즘을 따로 설계하였다. 또한 기존 분류 시스템을 통합 보완한 등급 분류 시스템을 만들어 실제 의료진의 판단과 일치하는 결과를 나타내고자 한다. 공개적으로 사용 가능한 OAI (Osteoarthritis Initiative) 데이터를 기반으로 하여, 총 9,786개의 슬관절 방사선 데이터가 본 연구에 사용되었으며, 최종적으로 Accuracy(정확도) 69.8%, F1 score 76.65%를 달성하였다.

Keywords

References

  1. Ahn, H. (2017). A Study on Compression of Connections in Deep Artificial Neural Networks, Journal of Korea Society of Industrial Information Systems, 22(5), 17-24. https://doi.org/10.9723/jksiis.2017.22.5.017
  2. Antony, J., McGuinness, K., Moran, K. and O'Connor, N. E. (2017). Automatic Detection of Knee Joints and Quantification of Knee Osteoarthritis Severity Using Convolutional Neural Networks, In Machine Learning and Data Mining in Pattern Recognition: 13th International Conference, MLDM 2017, New York, NY, USA, July 15-20, 2017, Proceedings 13 (pp. 376-390). Springer International Publishing. https://doi.org/10.1007/978-3-319-62416-7_27
  3. Antony, J., McGuinness, K., O'Connor, N. E. and Moran, K. (2016). Quantifying Radiographic Knee Osteoarthritis Severity Using Deep Convolutional Neural Networks, In 2016 23rd International Conference on Pattern Recognition (ICP R) (pp. 1195-1200). IEEE. https://doi.org/10.1109/ICPR.2016.7899799
  4. Bayramoglu, N., Nieminen, M. T. and Saarakkala, S. (2020a). A Lightweight CNN and Joint Shape-Joint Space (JS2) Descriptor for Radiological Osteoarthritis Detection, In Annual Conference on Medical Image Understanding and Analysis (pp. 331-345). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-52791-4_26
  5. Bayramoglu, N., Tiulpin, A., Hirvasniemi, J., Nieminen, M. T. and Saarakkala, S. (2020b). Adaptive Segmentation of Knee Radiographs for Selecting The Optimal ROI in Texture Analysis, Osteoarthrits and Cartilage, 28(7), 941-952. https://doi.org/10.1016/j.joca.2020.03.006
  6. Geirhos, R., Rubisch, P., Michaelis, C., Bethge, M., Wichmann, F. A. and Brendel, W. (2018). ImageNet-trained CNNs are Biased Towards Texture; Increasing Shape Bias Improves Accuracy and Robustness, arXiv preprint arXiv:1811.12231. https://doi.org/10.48550/ arXiv.1811.12231
  7. Harman, M. (2012). The Role of Artificial Intelligence in Software Engineering, In 2012 First International Workshop on Realizing AI Synergies in Software Engineering (RAISE) (pp. 1-6). IEEE. https://doi.org/10.1109/RAISE.2012.6227961
  8. Heo, Y. S. (2022). AI Medical & Healthcare, ASTI Market Insight, 65, 1-9.
  9. Hong, J.-Y., Park, S. H. and Jung, Y.-J. (2020). Artificial Intelligence Based Medical Imaging: An Overview, Journal Radiological Science and Technology, 43(3), 195-208. https://doi.org/10.17946/JRST.2020.43.3.195
  10. Jeong, G. H. (2018). AI Based Medical Image Analysis Technology Trends, Institute for Information & Communication Technology Planning & Evaluation(I ITP), Weekly Technology Trends.
  11. Kim, M. and Bae, H. J. (2020). Data Augmentation Techniques for Deep Learning-Based Medical Image Analyses, Journal of the Korean Society of Radiology, 81(6). https://doi.org/10.3348/jksr.2020.0158
  12. Korean Association of Knee Joints, (2021). Guidebook for Degenerative Arthritis with the Association of Knee Joints, https://www.koreaknee.or.kr/pop/file/guidebook.pdf
  13. Krizhevsky, A., Sutskever, I. and Hinton, G. E. (2012). Imagenet Classification with Deep Convolutional Neural Networks, Advances in Neural Information Processing Systems, 25.
  14. Lee, J. H., Kim, B. M. and Shin, Y. S. (2018). Effects of Preprocessing and Feature Extraction on CNN-Based Fire Detection Performance, Journal of Korea Society of Industrial Information Systems, 23(4), 41-53. https://doi.org/10.9723/jksiis.2018.23.4. 041
  15. Lee, Y. H. and Kim, Y. S. (2020). Comparison of CNN and YOLO for Object Detection, Journal of the Semiconductor & Display Technology, 19(1), 85-92.
  16. MSD Manual. (2021). The Korean Orthopaedic Association, Osteoarthritis (OA), https://www.koa.or.kr/info/index_10_1.php
  17. Pesapane, F., Codari, M. and Sardanelli, F. (2018). Artificial Intelligence in Medical Imaging: Threat or Opportunity? Radiologists Again at The Forefront of Innovation in Medicine, European Radiology Experimental, 2, 1-10. https://doi.org/10.1186/s41747-018-0061-6
  18. Qi, K., Yang, H., Li, C., Liu, Z., Wang, M., Liu, Q. and Wang, S. (2019). X-net: Brain Stroke Lesion Segmentation Based on Depthwise Separable Convolution and Long-Range Dependencies, In Medical Image Computing and Computer Assisted Intervention-MICCAI 2019: 22nd International Conference, Shenzhen, China, October 13-17, 2019, Proceedings, Part III 22 (pp. 247-255). Springer International Publishing. https://doi.org/10.1007/978-3-030-32248-9_28
  19. Ravi, D., Wong, C., Deligianni, F., Berthelot, M., Andreu-Perez, J., Lo, B. and Yang, G. Z. (2017). Deep Learning for Health Informatics, IEEE Journal of Biomedical and Health Informatics, 21(1), 4-21. https://doi.org/10.1109/JBHI.2016.2636665
  20. Shen, D., Wu, G. and Suk, H. I. (2017). Deep Learning in Medical Image Analysis, Annual Review of Biomedical Engineering, 19(1), 221-248. https://doi.org/10.1146/annurev-bioeng-071516-044442
  21. Trinder, J. C., Wang, Y., Sowmya, A. and Palhang, M. (1997). Artificial Intelligence in 3-D Feature Extraction, In Automatic Extraction of Man-Made Objects from Aerial and Space Images (I I ) (pp. 257-266). Birkhauser Basel. https://doi.org/10.1007/978-3-0348-8906-3_25
  22. Wright, R. W. (2014). Osteoarthritis Classification Scales: Interobserver Reliability and Arthroscopic Correlation, The Journal of Bone and Joint Surgery, 96(14), 1145-1151. https://doi.org/10.2106/JBJS.M.00929
  23. Zou, Z., Chen, K., Shi, Z., Guo, Y. and Ye, J. (2023). Object Detection in 20 years: A Survey, Proceedings of the Institute of Electrical and Electronics Engineers, 111(3), 257-276. https://doi.org/10.1109/JPROC.2023.3238524