نهج معتمد على الذكاء الاصطناعي للكشف الاوتوماتيكي عن سرطان العظام للعلاج الإشعاعي باستخدام فحوصات كاميرا جاما التشخيصية

Abstract

Bone scan (scintigraphy) is an efficient diagnostic tool for whole-body screening for bone metastases. The whole-body bone scan image analysis is primarily dependent on manual reading by nuclear medicine doctors. On the other hand, manual analysis is time-consuming, unpleasant, and requires a great deal of experience. This work suggested a machine-learning method that employs phases to identify bone metastases in order to address such problems The first part of this work is feature extraction, which is based on integrating the Mobile Vision Transformer (MobileViT) model into the framework to extract highly complex representations from raw medical scans. ViT and a lightweight CNN with few parameters are the two main components used in this process. Feature selection (FS), the second stage of this effort, depends on the Growth Optimizer (GO) being improved by the Arithmetic Optimization Algorithm (AOA). The applicability of bone scintigraphy for real-world applications is evaluated using 2,600 bone scan images (1,300 normal and 1,300 abnormal). The results and statistical analysis .revealed that the proposed algorithm as an FS technique outperforms the other FS algorithms in this study