Statistical assessment of an interpretable AI framework for improved disease diagnosis through medical images

Raghunath, K. M. Karthick; Unhelkar, Bhuvan; Shankar, S. Siva; Chakrabarti, Prasun

doi:10.47974/JSMS-1351

Open Access Research Article

Statistical assessment of an interpretable AI framework for improved disease diagnosis through medical images

^*K. M. Karthick RaghunathCorresponding authorkarthick.km@jainuniversity.ac.inMuma College of Business University of South Florida 8350 N. Tamiami Trail Sarasota; Department of Computer Science and Engineering JAIN (Deemed-to-be-University)Muma College of Business University of South Florida 8350 N. Tamiami Trail SarasotaFlorida, FL 34243, U.S.A.View full profile → , Bhuvan Unhelkarbunhelkar@usf.eduMuma College of Business University of South Florida 8350 N. Tamiami Trail SarasotaFlorida, 33620, USA0000-0003-1118-3837View full profile → , S. Siva Shankardrsivashankars@gmail.comDepartment of Computer Science and Engineering KG Reddy College of Engineering and TechnologyHyderabad, Telangana, 500100, India0000-0002-7616-6088View full profile → , Prasun Chakrabartidrprasun.cse@gmail.comDepartment of Computer Science and Engineering Sir Padampat Singhania UniversityUdaipur, Rajasthan, 313601, lndia0000-0001-8062-4144View full profile →

* Corresponding author · click or hover a name for details

pp. 193–204Vol. 28Issue 1January 2025DOI: 10.47974/JSMS-1351XML

Received:: 07 Feb 2024
Published Online:: 15 Jan 2025
Article type:: Research Article
Language:: EN
Article no.:: JSMS-1351
Pages:: 193–204

Abstract

In medical imaging, radiographic diagnoses based on computational techniques and statistical analysis call for a solution that integrates the gap between computational methods and statistical tools to ensure the correctness and reliability of the diagnosis process. Such a gap led to the innovation of a new framework called the “Statistically Improved Convolutional Interpretation (SICI)” framework for transparent and reliable diagnosis of disease from medical images. SICI framework is explored based on Convolutional Neural Networks (CNNs), attention mechanisms, and Explainable artificial intelligence (XAI) and complements it with statistical models. The foundation of this approach is based on the statistical compilation of improvements in the visual inspection accuracy and readability through combining CNNs classes of pathologic features, attention mechanism modules for prioritization of most image-important regions, and XAI elements for bringing statistically significant transparency of diagnostic outcomes. By integrating the statistical analysis into the framework, we aim to rigorously perform experiment research to identify accuracy, efficiency, and interpretability as the performance metrics. Compared to the older methods, it is associated with a marked reduction in error, and increased predictive value of diagnostic tests; thus, tends to bring interpretability and accuracy to the realm of medical image analysis.

Keywords

SHAP Interpretability Statistical Diagnosis Accuracy Evaluation Lung nodule Cancer Attention mechanism

Subject Classifications

68T0792C55

References

[1] M. N. S. Choudary, V. B. Bommineni, G. Tarun, G. P. Reddy, and G. Gopakumar, “Predicting COVID-19 positive cases and analysis on the relevance of features using SHAP (Shapley Additive Explanation),” in 2021 Second International Conference on Electronics and Sustainable Communication Systems (ICESC), pp. 1892–1896, Aug. (2021).
[2] C. Jacobs, E. M. van Rikxoort, T. Twellmann, E. Th. Scholten, P. A. de Jong, J.-M. Kuhnigk, M. Oudkerk, H. J. de Koning, M. Prokop, C. Schaefer-Prokop, and B. van Ginneken, “Automatic detection of subsolid pulmonary nodules in thoracic computed tomography images,” Medical Image Analysis, vol. 18, no. 2, pp. 374–384 (2014).
[3] R. Kankrale, T. Jadhav, P. A. Kharat, T. Deshmukh, N. G. Pardeshi, S. Karmode, and S. Gore, “Tensor Flow-powered spam email filtering: An evaluation of performance and robustness,” Journal of Electrical Systems, vol. 20, no. 6s, pp. 509–515 (2024).
[4] S. Jiang, H. Li, and Z. Jin, “A visually interpretable deep learning framework for histopathological image-based skin cancer diagnosis,” IEEE Journal of Biomedical and Health Informatics, vol. 25, no. 5, pp. 1483–1494 (2021).
[5] X. Jin, Y. Xie, X. S. Wei, B. R. Zhao, Z. M. Chen, and X. Tan, “Delving deep into spatial pooling for squeeze-and-excitation networks,” Pattern Recognition, vol. 121, p. 108159 (2022).
[6] S. Munot, R. R. Khinde, M. A. Hussain, P. P. Chiite, P. K. Mathurkar, A. Thakur, and S. Gore, “Microwave antenna optimization for low latency and high throughput communication systems,” Journal of Electrical Systems, vol. 20, no. 6s, pp. 2410–2416 (2024).
[7] K. S. Mader, “The Lung Image Database Consortium Image Collection (LIDC-IDIR),” IEEE Dataport (2021). Available: https://dx.doi.org/10.21227/zce3-jp96
[8] Md. R. Karim, T. Dohmen, M. Cochez, O. Beyan, D. Rebholz-Schuhmann, and S. Decker, “DeepCOVIDExplainer: Explainable COVID-19 diagnosis from chest X-ray images,” in 2020 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) (2020).
[9] D. J. McLernon, D. Giardiello, B. Van Calster, L. Wynants, N. van Geloven, M. van Smeden, ... and topic groups 6 and 8 of the STRATOS Initiative, “Assessing performance and clinical usefulness in prediction models with survival outcomes: Practical guidance for Cox proportional hazards models,” Annals of Internal Medicine, vol. 176, no. 1, pp. 105–114 (2023).
[10] Z. Naz, M. U. G. Khan, T. Saba, A. Rehman, H. Nobanee, and S. A. Bahaj, “An explainable AI-enabled framework for interpreting pulmonary diseases from chest radiographs,” Cancers, vol. 15, no. 1, p. 314 (2023).
[11] M. Ribeiro, S. Singh, and C. Guestrin, “‘Why should I trust you?’: Explaining the predictions of any classifier,” in Proc. 2016 Conf. North American Chapter of the Association for Computational Linguistics: Demonstrations (2016). Available: https://doi.org/10.18653/v1/n16-3020
[12] L. O. Teixeira, R. M. Pereira, D. Bertolini, L. S. Oliveira, L. Nanni, G. D. C. Cavalcanti, and Y. M. G. Costa, “Impact of lung segmentation on the diagnosis and explanation of COVID-19 in chest X-ray images,” Sensors, vol. 21, no. 21, p. 7116 (2021).
[13] R. van de Schoot, S. Depaoli, R. King, B. Kramer, K. Märtens, M. G. Tadesse, and C. Yau, “Bayesian statistics and modelling,” Nature Reviews Methods Primers, vol. 1, no. 1, pp. 1–13 (2021).
[14] L. Wang, Z. Q. Lin, and A. Wong, “COVID-Net: A tailored deep convolutional neural network design for detection of COVID-19 cases from chest X-ray images,” Scientific Reports, vol. 10, no. 1 (2020).
[15] M. S. Kotb and Y. Abdel-Aty, “Bayesian prediction intervals for the exponential-type distributions under multiply Type-II censoring data,” Journal of Statistics and Management Systems, vol. 27, no. 7, pp. 1261–1275 (2024), doi: https://doi.org/10.47974/JSMS-1014.
[16] N.-C. Wei, J.-M. Yang, H.-C. Lin, S.-C. Chen, and C.-J. Lee, “Effects of sustaining innovation and knowledge integration ability on innovation performance,” Journal of Statistics and Management Systems, vol. 26, no. 8, pp. 1915–1927 (2023), doi: https://doi.org/10.47974/JSMS-1089.

Views: 144Downloads: 8Citations: 2