Improved gated recurrent unit-based osteosarcoma prediction on histology images: a meta-heuristic-oriented optimization concept.

Abstract

The major prevalent primary bone cancer is osteosarcoma. Preoperative chemotherapy is accompanied by resection as part of the normal course of treatment. The diagnosis and treatment of patients are based on the chemotherapy reaction. Contrarily, chemotherapy without operation results in persistent cancer and an osteosarcoma regrowth. Thus, osteosarcoma patients should receive comprehensive therapy, which includes tumor-free surgery and global chemotherapy, to improve their survival. Hence, early diagnosis and individualized care of osteosarcoma are essential since they may lead to more effective therapies and higher survival rates. Here, the main goal of the recommended research is to use a unique deep learning approach to predict the osteosarcoma on histology images. Initially, the data is collected from the navigation confluence mobile osteosarcoma data of UT Southwestern/UT Dallas dataset. Next, the pre-processing of the collected images is accomplished by the Weiner filter technique. Further, the segmentation for the pre-processed images is done by the 2D Otsu's method. From the segmented images, the features are extracted by the linear discriminant analysis (LDA) approach. These extracted features undergo the final prediction phase that is accomplished by the novel improved recurrent gated recurrent unit (IGRU), in which the parameter tuning of GRU is accomplished by the osprey optimization algorithm (OOA) with the consideration of error minimization as the major objective function. On contrast with various conventional methods, the simulation findings demonstrate the effectiveness of the developed model in terms of numerous analysis.