Akademik Veri Yönetim Sistemi
Diğer, ss.1, 2024
Nasophayngeal Carcinoma (NPC) is one of the most commonly encountered cancer types in nasopharynx area. Utilization of radiation therapy guided by radiation oncologists is the main strategy when combatting against NPC. The planning of this procedure must be done precisely and without failure as it could lead to recurrence [1,2]. The planning of this procedure is mostly handle manually, takes time, and shows variance from expert to expert. In order to gain allocate more time to patient care, and make this process as accurate and less variable as possible, the usage of machine learning and computation techniques have been proposed. In this work, a machine learning algorithm is created from radiomics features with the intent of classifying and segmenting NPC tumor from computer tomography (CT) images. The CT images of 73 patients from Acibadem Maslak hospital were preprocessesd and prepared before the tumor classification and segmentation, then the performance of the model was evaluated by comparing the results with the ones that were identified by the radiation oncologists. After obtaining the results, on average, 95% accuracy score was obtained for the classification of NPC and 0.75 Jaccard score was obtained for the segmentation parts of the algorithm. Based on these results, it was concluded that the algorithm was capable of classifying NPC tumor from the CT images, but it lacked performance when it came to its segmentation.