Enhancing Gastric Cancer Lymph Node Detection through DL Analysis of CT Images: A Novel Approach for Improved Diagnosis and Treatment

Abstract

Although gastric cancer is a prevalent disease worldwide, accurate diagnosis and treatment of this condition depend on the ability to detect the lymph nodes. Recently, the use of Deep learning (DL) techniques combined with CT imaging has led to the development of new tools that can improve the detection of this disease. In this study, we will focus on the use of CNNs, specifically those built on the “MobileNet” and “AlexNet” platforms, to improve the detection of gastric cancer lymph nodes. The study begins with an overview of gastric cancer and discusses the importance of detecting the lymph nodes in the disease management cycle. CT and DL are discussed as potential technologies that can improve the accuracy of this detection. The study will look into the performance of CNNs, namely those built on the “AlexNet” and “MobileNet” platforms, in detecting the nodes in CT images of patients with gastric cancer. The study utilizes a dataset consisting of images of individuals with gastric cancer who have annotated lymph nodes. Various preprocessing steps, such as segmentation and image normalization, are carried out to improve the relevance and quality of the data. The two CNN architectures, namely “MobileNet” and the “AlexNet”, are evaluated for their performance in this area. Transfer learning methods are utilized to fine-tune models for detecting the lymph nodes. The results of the experiments are analyzed to determine the models' performance. The findings show that the “MobileNet” model is more accurate than the other platforms when it comes to detecting the lymph nodes. The study highlights the advantages of using DL techniques to enhance the accuracy of detecting the nodes in patients suffering from gastric cancer. It supports the notion that such techniques could help improve the diagnosis and treatment outcomes of this disease.