Breast cancer is the leading cause of cancer-related death among women worldwide, largely due to the unpredictable ways in which the body responds to cancer treatment.
A collaborative study at the University of Windsor is seeking to identify breast cancer survival rates for individuals undertaking hormone therapy, radiotherapy, or surgery.
Doctoral student Ashraf Abou Tabl from the Department of Mechanical, Automotive, and Materials Engineering, and post-doctoral researcher Abedalrhman Alkhateeb from the School of Computer Science, have analyzed the genetic activity of breast cancer patients to better understand bodily responses to cancer treatment with the help of their supervisors and research team.
Abou Tabl and Dr. Alkhateeb gathered data from an online publicly accessible dataset with a total of 347 patients, both living and deceased.
Computational methods and machine learning approaches developed in the PRIB-lab at the School of Computer Science was utilized to extract and pre-process the data, and then to make predictions based on the genomes, proteins, and clinical data of the patients to determine survivability of future patients based on a given therapy.
“Every time the database is updated with further patient information the machine learns and becomes more efficient in terms of predicting patient survivability,” said Alkhateeb.
Patients that have been identified to have lived for five years or longer after their treatment are considered to have survived.
“This would allow for better understanding of which patients should be given which treatments based on their bodily response to increase the rate of cancer survivability,” Alkhateeb said.
Also assisting with the research were doctoral student Huy Quang Pham, engineering professor Waguih ElMaraghy, and the principal investigators, computer science professors Alioune Ngom and Luis Rueda.
The study has been published in their collaborative paper, A Novel Approach for Identifying Relevant Genes for Breast Cancer Survivability on Specific Therapies. Read the full publication in Evolutionary Bioinformatics.
Source: UWindsor Daily News