A SYSTEMATIC REVIEW ON RENAL CELL CARCINOMA DIAGNOSIS USING ARTIFICIAL INTELLIGENCE APPROACHES

A kidney tumor is a dangerous disease that affects mostly males more than females, with a high mortality rate. The kidneys help to maintain fluid and solute balance in the body by filtering waste materials. A kidney tumor is categorized into three types: benign, malignant and indolent. Renal Cell Carcinoma (RCC) is a form of malignant tumor caused by defective genes. RCC classification is essential for further treatment planning. Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and ultrasound are some of the imaging modalities used to diagnose kidney tumors. Ultrasound is a screening technology that differentiates between solid and cystic lesions. However, cross-sectional imaging systems such as CT and MRI are also essential for the detection of kidney tumors. Mostly, CT is used for the identification of tumors. The goal of this review paper is to evaluate Deep Learning (DL) and Machine Learning (ML) models for detecting kidney tumors. Initially, existing review articles are evaluated to analyze previous techniques. After that, it offers a thorough examination of ML and DL models along with their datasets and limitations. Furthermore, this paper analyzes segmentation and classification-based techniques for detecting kidney tumors. The analysis of limitations of DL, ML, segmentation and classification techniques is utilized to highlight the challenges. Afterwards, different types of datasets are analyzed in this review paper to predict the best dataset. The limitations in this review paper help to construct a novel model for the detection of kidney tumors.