Present challenges and surveillance of artificial intelligence in neuro-oncology

Abstract

Brain tumors are evaluated, characterized, and monitored using diagnostic imaging. However, because these tumors are so diverse, there are still several difficulties in each group. This might involve differences in the biology of the cancer that are linked to varying intensities of cellular invasion, proliferation, and necrosis, all of which have distinct imaging appearances. Due to these changes, tumor evaluation, including segmentation, surveillance, and molecular characterizations, has become more complex. Even though various rule-based techniques have been put into practice to relate to tumor appearance and size, these techniques naturally reduce the vast quantity of tumor imaging data to a small number of variables. Due to their efficacy in resolving image-based problems, approaches in artificial intelligence, machine learning, and deep learning have found increased use in computer vision tasks, such as tumor imaging. This section aims to provide an overview of some of these developments in tumor imaging.