Optimized Hybrid Prediction Method for Lung Metastases

Abstract

Brain metastases are the most prevalent intracranial neoplasm that causes excessive morbidity and mortality in most cancer patients. The current medical model for brain metastases is focused on the physical condition of the affected individual, the anatomy of the main tumor, and the number and proximity of brain lesions. In this paper, a new hybrid Metastases Fast Fourier Transformation with SVM (MFFT-SVM) method is proposed that can classify high dimensional magnetic resonance imaging as tumor and predicts lung cancer from given protein primary sequences. The goal is to address the associated issues stated with the treatment targeted at unique molecular pathways to the tumor, together with those involved in crossing the blood-brain barrier and migrating cells to the lungs. The proposed method identifies the place of the lung damage by the Fast Fourier Technique (FFT). FFT is the principal statistical approach for frequency analysis which has many engineering and scientific uses. Moreover, Differential Fourier Transformation (DFT) is considered for focusing the brain metastases that migrate into the lungs and create non-small lungs cancer. However, Support Vector Machine (SVM) is used to measure the accuracy of control patient's datasets of sensitivity and specificity. The simulation results verified the performance of the proposed method is improved by 92.8% sensitivity, of 93.2% specificity and 95.5% accuracy respectively.