Development of Artificial Intelligence (AI) for urothelial carcinoma detection

Authors

  • Suchada Katedee Rajamangala University of Technology Phra Nakhon
  • Thiwaporn Thesawadwong2 สถาบันมะเร็งแห่งชาติ กรมการแพทย์ กระทรวงสาธารณสุข

Keywords:

Artificial Intelligence, urothelial carcinoma, detection, diagnosis

Abstract

This research aimed to develop an artificial intelligence program or AI program for urothelial carcinoma diagnosis software. By using artificial intelligence technology for classifying images of urothelial cells in urine liquid-based cytology specimens according to the criteria of The Paris System for Reporting Urinary Cytology version 2.0. Using Convolutional Neural Network (CNN) techniques by choosing to use the DenseNet121 algorithm. Because it is an efficient algorithm suitable for use as a model used in developing AI programs. It uses urothelial cell data from the National Cancer Institute to create a training set, validation set, and test set of the model. The researcher evaluated the model's performance by determining the sensitivity, specificity, accuracy, and precision and determining the reliability of urothelial cell diagnosis between the AI program and 50 pathologists (benchmark). The consistency of the diagnostic results was determined by finding Cohen's kappa coefficient and finding the median test value. As for finding accuracy according to criteria (Criterion volatility), the researcher used the diagnosis of a team of pathologists and cytologists as a standard criterion (gold standard). The results of the research found that the sensitivity, specificity, accuracy, and precision were 97.50%, 100%, 98.33%, and 100% respectively. It has a Kappa value of 0.716 - 0.970, which is consistent at a good - very good level and there was no difference in the median test value of the consistency in giving the decision. Therefore, it can be concluded that the artificial intelligence program for urothelial carcinoma detection is a reliable tool. It can be used as a medical decision-support tool in the diagnosis of urothelial carcinoma

References

สถาบันมะเร็งแห่งชาติ. แนวทางการตรวจคัดกรอง วินิจฉัยและรักษาโรคมะเร็งกระเพาะปัสสาวะ. กรุงเทพ ฯ: กรมการแพทย์ กระทรวงสาธารณสุข; 2563.

Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. Densely connected convolutional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017. p. 4700-8.

LeCun Y, Bengio Y, Hinton G. Deep learning. nature. 2015 May 28;521(7553): 436-44.

กอบเกียรติ สระอุบล. เรียนรู้ AI Deep Learning ด้วย Python. กรุงเทพ ฯ: อินเตอร์มีเดีย; 2565.

Barkan GA, Wojcik EM, Nayar R, Savic-Prince S, Quek ML, Kurtycz DF, Rosenthal DL. The Paris system for reporting urinary cytology: the quest to develop a standardized terminology. Acta Cytologica. 2016 Aug 26;60(3):185-97.

Wojcik EM, Kurtycz DF, Rosenthal DL, editors. The Paris system for reporting urinary cytology. New York, NY, USA: Springer; 2022 Jan 1.

Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, Thrun S. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017 Feb;542(7639):115-8.

Hashmi MF, Katiyar S, Keskar AG, Bokde ND, Geem ZW. Efficient pneumonia detection in chest x-ray images using deep transfer learning. Diagnostics. 2020 Jun 19;10(6):417.

Pattanasuwan C, Chongstitvatana P. Screening TB using deep transfer learning. In: 2021 25th International Computer Science and Engineering Conference (ICSEC). IEEE; 2021 Nov 18. p. 330-3.

Apipawinwongsa P, Limpiyakorn Y. Counterfeit luxury handbag materials image classification using deep learning and local binary pattern. Int J Emerg Technol Adv Eng. 2022;12(9):41-8.

Bibi N, Sikandar M, Ud Din I, Almogren A, Ali S. IoMT-based automated detection and classification of leukemia using deep learning. J Healthc Eng. 2020;2020(1):6648574.

Tsuneki M, Abe M, Kanavati F. Deep learning-based screening of urothelial carcinoma in whole slide images of liquid-based cytology urine specimens. Cancers. 2022 Dec 30;15(1):226.

Krishnadas P, Sampathila N. Automated detection of malaria implemented by deep learning in PyTorch. In: 2021 IEEE International Conference on Electronics, Computing and Communication Technologies (CONECCT). IEEE; 2021 Jul 9. p. 01-05.

Abadi M, Agarwal A, Barham P, Brevdo E, Chen Z, Citro C, et al. TensorFlow: Large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467. 2016 Mar 14.

จาตุรงค์ ตันติบัณฑิต. การเรียนรู้เชิงลึก ทฤษฏีและนวัตกรรมทางวิศวกรรมการแพทย์ (Deep Learning Theory and Innovation in Medical Engineering). กรุงเทพฯ: เพชรเกษมพริ้นติ้ง กรุ๊ป จำกัด; 2561.

Stevens E, Antiga L, Viehmann T. Deep learning with PyTorch. Manning Publications; 2020 Aug 4.

Subramanian V. Deep Learning with PyTorch: A practical approach to building neural network models using PyTorch. Packt Publishing Ltd; 2018 Feb 23.

ประสพชัย พสุนนท์. การประเมินความเชื่อมั่นระหว่างผู้ประเมินโดยใช้สถิติแคปปา. วารสารวิชาการศิลปศาสตร์ประยุกต์. 2558: 2-20.

อุมาพร จันทศร. สถิติที่ไม่ใช้พารามิเตอร์. กรุงเทพ ฯ: บริษัทฟิสิกส์เซ็นเตอร์; 2542.

Downloads

Published

2024-12-29

How to Cite

[1]
S. Katedee and T. . Thesawadwong2, “Development of Artificial Intelligence (AI) for urothelial carcinoma detection”, UTK RESEARCH JOURNAL, vol. 18, no. 2, pp. 1–13, Dec. 2024.

Issue

Section

Research Articles