Artificial Intelligence for Prediction and Early Diagnosis of Cancer

• Grant number: 23/15728-1
• Support Opportunities: Scholarships in Brazil - Doctorate
• Start date: December 01, 2023
• End date: November 30, 2027
• Field of knowledge: Physical Sciences and Mathematics - Mathematics - Applied Mathematics
• Agreement: MCTI/MC
• Principal Investigator: José Soares de Andrade Júnior
• Grantee: Higor da Silva Monteiro
• Host Institution: Centro de Ciências. Universidade Federal do Ceará (UFC). Ministério da Educação (Brasil). Fortaleza , SP, Brazil
• Company: Universidade Federal do Ceará (UFC). Reitoria
• Associated research grant: 20/09706-7 - CEREIA - Reference Center on Artificial Intelligence, AP.PCPE

Abstract

Cancer is one of the main threats to public health in the new millennium. Its increasing incidence, along with cardiovascular diseases, currently results in the highest global disease burden. The high complexity and costs of the procedures used to identify and treat neoplasms make the burden even greater, especially in developing countries. In this context, early detection of cases and the prediction of new cases are fundamental goals for the development of effective treatment and management strategies. Recently, the use of artificial intelligence technologies has emerged as one of the main tools for achieving these goals. In particular, the use of deep neural networks in natural language and computer vision has produced promising results in the identification and characterization of neoplasms through supervised learning of medical records and imaging of thousands of patients. In this project, we will use anonymized healthcare data associated with the monitoring of millions of beneficiaries of the Hapvida NotreDame Intermédica health plan network. These data are characterized by the richness of patient information, covering demographic, geographical, and clinical aspects. With this, we will use deep learning techniques to build optimized computer vision and natural language models for cancer detection, with a focus on the most common types of cancer in the Brazilian population. Furthermore, we will use machine learning models to assess potential risk determinants in order to optimize the screening and early detection processes of new cancer cases. Thus, we hope that the results will have a significant impact on the generation of scientific knowledge and the direction of new measures for the treatment and management of neoplasms.