Arushi Singh

Predicting breast cancer subtype outcomes and the driver mutations that lead to different subtypes

Started June 15, 2023

Abstract or project description

Breast cancer is the most common type of cancer among people and demands for accurate predictive models to help improve patient outcomes. As machine learning models have become more advanced than ever, this can be used to our advantage and help save lives. In this research, we answer two questions and utilize machine learning models to predict the outcomes. Firstly, we investigate applying linear and logistic regression models to predict breast cancer patient's survival rate and status, respectively, with infiltrating duct and lobular carcinoma. These models use a dataset that contains data about patients' cancer stages based on the TNM system and this dataset is used to train/predict patients’ survival rate and status. Secondly, we investigate using a K-means clustering model to predict patients’ breast cancer subtypes based on DNA mutations. The model uses a proteome dataset to help train/predict which one of the three cancer subtypes a patient has and which DNA mutation most likely causes that. Our research is aimed to explore the potential of machine learning to help breast cancer patients and develop models to aid doctors in recommending treatments and making well-informed decisions for patients.