Artificial intelligence based brain tumor detection classification and survival prediction

Abstract

Now-a-days, human society is facing many complications with newlinehealth-related issues. The brain tumor has been considered as one of the most newlinedangerous and life threatening diseases. Brain tumor has also been considered newlineas a growth or group of abnormal brain cells, classified as benign newline(noncancerous) and malignant (cancerous). The patients, who are affected by newlinebrain tumor with a growth of tumor more than 50%, will not survive. newlineAccording to the recent medical statistical reports, it is observed that nearly newline20 million people all over the world are affected by dreadful diseases and newlineapproximately, 10 million people have lost their lives, due to this tumor. newlineMoreover, it is predicted that around 30 million people can be affected by newlinebrain tumor in the year of 2040. The growing cancer burden requires a newlinecomprehensive approach that includes prevention, early detection, accurate newlinediagnosis, effective treatment, and palliative care. Computed Tomography newline(CT), Positron Emission Tomography (PET), and Magnetic Resonance newlineImaging (MRI) have been used to acquire data from the brain tissues. MRI is newlineoften selected as the preferred technique by the medical practitioners. There newlineare several modalities of MRI used for brain imaging, including T1-weighted newline(T1), T1-weighted contrast-enhanced (T1ce), T2-weighted (T2), and fluid newlineattenuated inversion recovery (FLAIR). Each modality provides different newlinecontrasts and helps to visualize different brain tissues and abnormalities. newlineSeveral methods have been proposed by various researchers for detecting newlinemalignancies in brain images. According to the recent reviews, it is analyzed newlinethat these methods possess some major problems in terms of high newlinecomputational burden, increased time for tumor detection, ineffective survival newlineprediction, high false positives, and overlapping results. newline