Klasifikasi Kanker Renal Cell Carcinoma melalui Cira Ct Scan Menggunakan Probabilistic Neural Network

Abstract

One of the most common types of kidney cancer is renal cell carcinoma. Renal cell carcinoma is divided into three types, Clear Cell Carcinoma, Papillary Renal Cell Carcinoma, and Chromophobe Renal Cell Carcinoma. This cancer classification can be done by radiological examination through CT scan medical imaging. However, to classify the cancer, expert doctors diagnose it manually first. Therefore we need a method that can perform early processing of radiological detection through medical images quickly and accurately. This research uses Probabilistic Neural Network (PNN). The data used were CT scan of 3 types of Renal Cell Carcinoma as much as 211 data which is then divided into two parts, 151 data as training data and 60 test data. The first stage is to divide the data into training data and test data. Then, Grayscaling, Scaling and Contrast Limited Adaptive Histogram Equalization (CLAHE) were performed at the preprocessing stage. Then, the segmentation stage uses the Thresholding method. Then, the Feature Extraction process is carried out using the Gray level Co-occurance Matrix (GLCM) method. The last stage is classification using PNN, the results of the classification are Clear Cell Carcinoma, Papillary Renal Cell Carcinoma, and Chromophobe Renal Cell Carcinoma. The test results reach an accuracy of 90%.