Lung Cancer Diagnosis from histopathological images using deep learning approaches

lajevardi, simasadat; Allahverdi, َAbdollah; Shojaeilangari, Seyedehsamane

All

Webpages

Books

Journals

Tarbiat Modares University Press

Modares Journal of Biotechnology

Volume 15, Issue 2 (2024) JMBS 2024, 15(2): 0-0 | Back to browse issues page

Mendeley

Zotero

RefWorks

lajevardi S, Allahverdi َ, Shojaeilangari S. Lung Cancer Diagnosis from histopathological images using deep learning approaches. JMBS 2024; 15 (2)
URL: http://biot.modares.ac.ir/article-22-71149-en.html

Lung Cancer Diagnosis from histopathological images using deep learning approaches

Simasadat Lajevardi¹

, َAbdollah Allahverdi ^*

², Seyedehsamane Shojaeilangari³

1- Biophysics Dept. Faculty of Biological Sciences, Tarbiat Modares University
2- Dept. of Biophysics, Faculty of Biological Science Tarbiat Modares University , a-allahverdi@modares.ac.ir
3- Dept. of electrical and information technology, IROST

Abstract: (550 Views)

Cancer is one of the leading causes of death worldwide. Breast cancer is the most common cancer among women and causes a high number of annual deaths. The most reliable method for successful cancer management is accurate and early diagnosis. On the other hand, the lack of timely diagnosis leads to the spread of cancer in the body, making it difficult to treat and control. The gold standard method for breast cancer diagnosis is biopsy. Usually, visual inspection and manual assesement are used to diagnose cancer, where the pathologist examines the histopathology slides under a microscope which is error-prone and time- consuming procedure and requires years of expertise. Therefore, computer-aided diagnosis is essential to help physicians improve the efficiency of interpreting medical images. In this study, we use deep learning models, especially convolutional neural networks (CNNs) to detect whether or not histopathological images are cancerous. The AUC, Precision and F1-score obtained using the pre-trained Incetion-V3 deep neural network are 98.36%, 95.28% and,97.25% respectively, and the same parameters for the pre-trained ResNet-18 deep neural network are equal to 97.90 %, 97.46% and 98.22%. The presented models are able to provide reliable diagnosis results for different morphologies of breast tissues.

Keywords: Breast cancer, histopathology images, deep learning, convolutional neural networks (CNNs).

Full-Text [PDF 1824 kb] (364 Downloads)

Article Type: Original Research | Subject: Bioinformatics
Received: 2023/08/19 | Accepted: 2024/04/16 | Published: 2024/06/24

Send email to the article author

Rights and permissions
	This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.