การพัฒนาเว็บแอปพลิเคชันสำหรับช่วยวินิจฉัยโรคหลอดเลือดสมอง  โดยใช้ปัญญาประดิษฐ์แบบการเรียนรู้เชิงลึกบนภาพเอกซเรย์คอมพิวเตอร์

Titipong Kaewlek; Ketmanee  Sitinwan; Kunaporn  Lueangaroon; Wasita   Sansuriyawong; Rawiwan   Pattaweerakul; Thatchai Hantrakul; Bhuwid  Chinwatanawongwan

Authors

Titipong Kaewlek Department of Radiological Technology, Faculty of Allied Health Sciences, Naresuan University https://orcid.org/0000-0001-6400-0489
Ketmanee Sitinwan Department of Radiological Technology, Faculty of Allied Health Sciences, Naresuan University https://orcid.org/0009-0008-0246-7874
Kunaporn Lueangaroon Department of Radiological Technology, Faculty of Allied Health Sciences, Naresuan University https://orcid.org/0009-0000-7472-4253
Wasita Sansuriyawong Department of Radiological Technology, Faculty of Allied Health Sciences, Naresuan University https://orcid.org/0009-0009-5735-4823
Rawiwan Pattaweerakul Department of Radiology, Faculty of Medicine, Naresuan University
Thatchai Hantrakul Department of Radiology, Faculty of Medicine, Naresuan University
Bhuwid Chinwatanawongwan Department of Radiology, Faculty of Medicine, Naresuan University https://orcid.org/0000-0002-9430-0930

Keywords:

Artificial Intelligence, Computer aided detection, Deep Learning, Stroke, Web Application

Abstract

Introduction: Stroke causes a high risk among the elderly. Accurate diagnosis using computed tomography (CT) imaging is essential, and expert interpretation is required. Artificial intelligence (AI) has increasingly contributed to improving the accuracy of radiologists’ interpretations. Objective: This study aimed to develop a web application for stroke detection in brain CT images using artificial intelligence. Methods: A total of 1,636 brain CT images (1,111 normal images and 525 stroke images) were used in this study. Approximately 70% (1,175 images) of the dataset were used for model training, 20% (329 images) for validation, and 10% (132 images) for testing the deep learning model (VGG-16). Accuracy, sensitivity, specificity, F1-score, and area under the curve (AUC) were used to evaluate the performance of the web application. Design and usability were evaluated using a 5-point Likert scale by three radiologists. Results: The results shows that the accuracy, sensitivity, specificity, F1-score, and AUC were 0.969, 0.952, 0.978, 0.952, and 0.965, respectively. The design and usability scores evaluating by radiologists were 4.13 ± 0.38 and 4.37 ± 0.33, respectively. Conclusion: The developed web application demonstrated high performance in diagnosing stroke from brain CT images and is easy to use via internet-based platforms.

Downloads

Download data is not yet available.

References

World Health Organization (WHO). World stroke day 2022 [cited 2022 November 17]. Available from: https://www.who.int/srilanka/news/detail/29-10-2022-world-stroke-day-2022

The Ministry of Public Health of Thailand, Public Health Statistics A.D. 2022. [cited 2022 May 14]. Available from: https://spd.moph.go.th/wp-content/uploads/2023/11/Hstatistic65.pdf

González RG. Current State of Acute Stroke Imaging. Stroke. 2013; 44(11):3260-3264 https://doi.org/10.1161/STROKEAHA.113.003229

Birenbaum D, Bancroft LW, Felsberg GJ. Imaging in Acute Stroke. West J Emerg Med. 2011;12(1):67-76.

Guberina N, Dietrich U, Radbruch A, et al. Detection of early infarction signs with machine learning based diagnosis by means of the Alberta Stroke Program Early CT score (ASPECTS) in the clinical routine. Neuroradiology. 2018; 60:889–901.

Kuang H, Najm M, Chakraborty D, et al. Automated ASPECTS on noncontrast CT scans in patients with acute ischemic stroke using machine learning. American Journal of Neuroradiology 2019; 40:33–38

Shafaat O, Bernstock DJ, Shafaat A, et.al. Leveraging artificial intelligence in ischemic stroke imaging. Journal of Neuroradiology. 2022; 49; 343–351.

Pandey N. Lung segmentation from Chest X-Ray dataset. Kaggle Inc [cited 2022 April 28]. Available from: https://www.kaggle.com/code/nikhilpandey

/lung-segmentation-from-chest-x-ray-dataset/

notebook

Tasnia N. brain-stroke-prediction-ct-scan-image-dataset. [cited 2022 May 1]. Available from: https://www.kaggle.com/datasets/noshintasnia/brain-stroke-prediction-ct-scan-image-dataset

Vbookshelf. Brain CT Images with Intracranial Hemorrhage Masks. [cited 2022 May 1]. Available from: https://www.kaggle.com/datasets/vbookshelf/

computed-tomography-ct-images

Kaewlek T, Sitinwan K, Lueangaroon K, Sansuriyawong W. Comparative analysis of deep learning techniques for accurate stroke detection. Journal of Associated Medical Sciences. 2024; 57(2), 49–55.

Varshney P. VGGNet-16 Architecture: A Complete Guide. [cited 2022 January 13]. Available from: https://www.kaggle.com/code/blurredmachine/vggnet-16-architecture-a-complete-guide

Vogt WP. Dictionary of statistics and methodology. 1999, 2nd ed, Sage: Thousand Oaks, California.

Joshi A, Kale S, Chandel S, Pal DK. Likert Scale: Explored and Explained. Current Journal of Applied Science and Technology, 2015; 7(4), 396–403.

Mcleod S. Likert Scale Questionnaire: Examples & Analysis. [cited 2023 August 13]. Available from: https://www.simplypsychology.org/likert-scale.html

Chen YT, Chen YL, Chen YY, Huang YT, Wong HF, Yan JL, et al. Deep Learning–Based Brain Computed Tomography Image Classification with Hyperparameter Optimization through Transfer Learning for Stroke. Diagnostics 2022; 12, 807.

Ozaltin O, Coskun O, Yeniay O, Subasi A. A Deep Learning Approach for Detecting Stroke from Brain CT Images Using OzNet. Bioengineering 2022; 9: 783.

Li S, Zheng J, Li D. Precise segmentation of non-enhanced computed tomography in patients with ischemic stroke based on multi-scale U-Net deep network model. Computer methods and programs in biomedicine. 2021; 208:106278.

Guerrero R, Qin C, Oktay O, Bowles C, Chen L, Joules R, et al. White matter hyperintensity and stroke lesion segmentation and differentiation using convolutional neural networks. Neuroimage Clin. 2018; 17:918-34.

Phong TD, Duong HN, Nguyen HT, Trong NT, Nguyen VH, Hoa TV, et al. Brain hemorrhage diagnosis by using deep learning. Proceedings of the 2017 International Conference on Machine learning and soft Computing; Ho Chi Minh City, Vietnam. Association for Computing Machinery; 2017. p. 34–9.

Gaidhani BR, Rajamenakshi RR, Sonavane S. Brain stroke detection using convolutional neural network and deep learning models. 2019, 2nd international conference on intelligent communication and computational techniques (ICCT); 2019; Sept 28-29. Jaipur, India. 2019.

Mouridsen K, Thurner P, Zaharchuk G. Artificial intelligence applications in stroke. Stroke. 2020; 51(8):2573-9.

Ding L, Liu C, Li Z, Wang Y. Incorporating artificial intelligence into stroke care and research. Stroke. 2020; 51(12): e351-e4.

Ammar M, Lamria MA, Mahmoudib S, Laidia A. Deep Learning Models for Intracranial Hemorrhage Recognition: A comparative study. Procedia Computer Science. 2022;196: 418–425.

Vamsi B, Bhattacharyya D, Midhunchakkravarthy D, Kim JY. Early Detection of Hemorrhagic Stroke Using a Lightweight Deep Learning Neural Network Model. Traitement du signal. 2021;38(6): 1727-1736.

Soun JE, Chow DS, Nagamine M, Takhtawala RS, Filippi CG, Yu W, et al. Artificial intelligence and acute stroke imaging. AJNR Am J Neuroradiol. 2021; 42(1): 2-11.