Keywords: Machine Learning/Artificial Intelligence, Screening, XAI, explainability, deep learning
In this study, we propose a simple method to improve the explainability of artificial intelligence, specifically convolutional neural networks (CNNs), for the automatic detection of early nasopharyngeal carcinoma (NPC) on magnetic resonance imaging (MRI). We show a long-short-term-memory (LSTM) unit can be introduced into a CNN to read 3-dimensional medical image series. A risk curve can be extracted from the LSTM to visualize the “thought process” of the network when it reads through the input MRI slice-by-slice. This modification improves the explainability of the network without reducing performance for the early NPC detections of the original CNN.[1] A. D. King, J. K. S. Woo, Q. Y. Ai, F. K. F. Mo, T. Y. So, W. K. J. Lam, I. O. L. Tse, A. C. Vlantis, K. W. N. Yip, E. P. Hui, B. B. Y. Ma, R. W. K. Chiu, A. T. C. Chan, Y. M. D. Lo, and K. C. A. Chan, “Early Detection of Cancer: Evaluation of MR Imaging Grading Systems in Patients with Suspected Nasopharyngeal Carcinoma,” AJNR Am J Neuroradiol, vol. 41, no. 3, pp. 515-521, Mar, 2020.
[2] Z. Liu, H. Li, K. J. Yu, S. H. Xie, A. D. King, Q. H. Ai, W. J. Chen, X. X. Chen, Z. J. Lu, L. Q. Tang, L. Wang, C. M. Xie, W. Ling, Y. Q. Lu, Q. H. Huang, A. E. Coghill, C. Fakhry, R. M. Pfeiffer, Y. X. Zeng, S. M. Cao, and A. Hildesheim, “Comparison of new magnetic resonance imaging grading system with conventional endoscopy for the early detection of nasopharyngeal carcinoma,” Cancer, vol. 127, no. 18, pp. 3403-3412, Sep 15, 2021.
[3] K. C. A. Chan, J. K. S. Woo, A. King, B. C. Y. Zee, W. K. J. Lam, S. L. Chan, S. W. I. Chu, C. Mak, I. O. L. Tse, S. Y. M. Leung, G. Chan, E. P. Hui, B. B. Y. Ma, R. W. K. Chiu, S. F. Leung, A. C. van Hasselt, A. T. C. Chan, and Y. M. D. Lo, “Analysis of Plasma Epstein-Barr Virus DNA to Screen for Nasopharyngeal Cancer,” N Engl J Med, vol. 377, no. 6, pp. 513-522, Aug 10, 2017.
[4] D. C. T. Chan, W. K. J. Lam, E. P. Hui, B. B. Y. Ma, C. M. L. Chan, V. C. T. Lee, S. H. Cheng, W. Gai, P. Jiang, K. C. W. Wong, F. Mo, B. Zee, A. D. King, Q. T. Le, A. T. C. Chan, K. C. A. Chan, and Y. M. D. Lo, “Improved risk stratification of nasopharyngeal cancer by targeted sequencing of Epstein-Barr virus DNA in post-treatment plasma,” Ann Oncol, vol. 33, no. 8, pp. 794-803, Aug, 2022.
[5] L. M. Wong, Q. Y. H. Ai, F. K. F. Mo, D. M. C. Poon, and A. D. King, “Convolutional neural network in nasopharyngeal carcinoma: how good is automatic delineation for primary tumor on a non-contrast-enhanced fat-suppressed T2-weighted MRI?,” Jpn J Radiol, vol. 39, no. 6, pp. 571-579, Jun, 2021.
[6] L. M. Wong, Q. Y. H. Ai, D. M. C. Poon, M. Tong, B. B. Y. Ma, E. P. Hui, L. Shi, and A. D. King, “A convolutional neural network combined with positional and textural attention for the fully automatic delineation of primary nasopharyngeal carcinoma on non-contrast-enhanced MRI,” Quant Imaging Med Surg, vol. 11, no. 9, pp. 3932-3944, Sep, 2021.
[7] L. M. Wong, A. D. King, Q. Y. H. Ai, W. K. J. Lam, D. M. C. Poon, B. B. Y. Ma, K. C. A. Chan, and F. K. F. Mo, “Convolutional neural network for discriminating nasopharyngeal carcinoma and benign hyperplasia on MRI,” Eur Radiol, vol. 31, no. 6, pp. 3856-3863, Jun, 2021.
[8] L. Ke, Y. Deng, W. Xia, M. Qiang, X. Chen, K. Liu, B. Jing, C. He, C. Xie, X. Guo, X. Lv, and C. Li, “Development of a self-constrained 3D DenseNet model in automatic detection and segmentation of nasopharyngeal carcinoma using magnetic resonance images,” Oral Oncol, vol. 110, pp. 104862, Nov, 2020.
[9] Y. S. Deng, C. F. Li, X. Lv, W. X. Xia, L. J. Shen, B. Z. Jing, B. Li, X. Guo, Y. Sun, C. M. Xie, and L. R. Ke, “The contrast-enhanced MRI can be substituted by unenhanced MRI in identifying and automatically segmenting primary nasopharyngeal carcinoma with the aid of deep learning models: An exploratory study in large-scale population of endemic area,” Computer Methods and Programs in Biomedicine, vol. 217, pp. 106702, Apr, 2022.
[10] F. Wang, M. Jiang, C. Qian, S. Yang, C. Li, H. Zhang, X. Wang, and X. Tang, "Residual attention network for image classification," Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 3156-3164.
[11] J. T. Springenberg, A. Dosovitskiy, T. Brox, and M. Riedmiller. "Striving for simplicity: The all convolutional net," https://arxiv.org/abs/1412.6806.
[12] R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, "Grad-cam: Visual explanations from deep networks via gradient-based localization." pp. 618-626.