Breast cancer detection using AI

This project was developed during my master's studies at the ITM (Instituto Tecnológico Metropolitano) from Medellin and was officially titled “Computer-Assisted Abbreviated Magnetic Resonance Imaging Protocol for Detection and Categorization of Suspicious Breast Cancer Lesions”.

The project aimed to revolutionize breast cancer diagnosis through the development of a computer-assisted abbreviated magnetic resonance imaging (MRI) protocol. This innovative approach utilizes advanced machine learning techniques, specifically Support Vector Machines (SVM) and Multiple Kernel Learning (MKL), to enhance the detection and categorization of suspicious breast lesions.

Traditional breast cancer diagnosis often involves lengthy and resource-intensive MRI protocols, leading to delays in treatment initiation and increased patient anxiety and money expenditures. The proposed abbreviated protocol streamlines the imaging process without compromising diagnostic accuracy. By leveraging SVM, a powerful supervised learning algorithm, we could effectively classify MRI images and identify regions of interest indicative of potential cancerous lesions.

Moreover, we employed MKL to integrate diverse sources of data, such as anatomical, functional, and textural features extracted from MRI scans. This multi-modal approach enhances the robustness and reliability of our diagnostic model, enabling more accurate identification and characterization of breast lesions.

Through the integration of SVM and MKL, our computer-assisted protocol offers several advantages over traditional methods. It significantly reduces the time required for image acquisition and interpretation, leading to faster diagnosis and treatment planning. Additionally, by leveraging machine learning algorithms, our system can adapt and improve over time, further enhancing its diagnostic capabilities.