Revolutionizing Esophageal Cancer Diagnosis: Vgg-16 Deep Learning Approach

Esophageal cancer stands as the sixth-leading cause of cancer-related fatalities globally. Less than 20% of malignancies survive, and most show symptoms later and at an advanced stage. In the medical domain, deep learning (DL) holds great significance since it facilitates sophisticated pattern identification and analysis of intricate medical data, leading to precise disease diagnosis and customized treatment regimens. The efficiency and accuracy of medical decision-making have greatly increased as a result of its capacity to extract valuable insights from massive databases, including patient information and medical pictures. In order to detect esophageal cancer, the suggested study used pretrained convolutional neural networks (CNN) that were specifically designed for image identification and tweaked VGG16. The evaluation makes use of a dataset made up of 100 HD-WLE pictures of the lower esophagus from the Endoscopic Vision Challenge, MICCAI 2015. Next, using the normal image data and the esophageal cancer image, the pretrained CNN was trained. The study used Grad-CAM, a technique for creating visual explanations, to successfully identify the reasoning behind the diagnosis of esophageal cancer. The accuracy, precision, sensitivity, and specificity of the system were 83.1%, 83.10, 81.7%, and 83.21%, respectively.