MRI(Radiomics) and Machine Learning

Welcome​

​

NB: Use tabs at the top to see other sections.

​

This website will explore a research paper written, Machine Learning and Deep Learning Techniques to Predict Overall Survival of Brain Tumor Patients using MRI Images,  by Lina Chato and Shahram Latifi from the University of Nevada, and will attempt to reproduce the results of the paper. The replication will be completed by two undergraduate students from Virginia Tech. Both Ashley Shafer and Alex Lind are students within the Computer Science Department, as well as both students,  are minoring in Biomedical Engineering and Biological Sciences respectively.

Due to the aggressive and fatal nature of this cancer, a great deal of research has been completed to provide the best care and diagnosis for the patient afflicted with this cancer.  The importance of this survival rate prediction is to help doctors and patients work together to provide the best course of action, involving chemotherapy, electrical stimulus, bacterial intervention, surgical methods, or a hospice care plan that would best suit the patient, considering current medical technology and treatments are not extremely effective in keeping patients alive during treatment or after remission. Thus, machine learning and advancements in medical imaging are being looked at by medical professionals and researchers as a way to treat cancer more aggressively and effectively by creating a specified plan per individual patient. Thus, the paper from above from which this website will analyze and try to reproduce the results of, created a Convolutional Neural Network(CNN) to classify tumor types from patient data consisting of 4 MRI scans, age, gender, and survival time of a patient from the BraTS 2017 data set to train their neural network.[1]

​

Our job is to try and replicate the process described in the paper on other data sets to predict age and other factors for cancer patients based on MRI scans. We will be using a lung cancer MRI dataset from Kaggle[4].