Application of Artificial Intelligence to aid in the diagnosis and prognosis of SARS-CoV-2

• Grant number: 20/09807-8
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Duration: April 01, 2021 - December 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Jairo da Silva Freitas Júnior
• Grantee: Jairo da Silva Freitas Júnior
• Host Company: Lablift Ltda
• CNAE: Desenvolvimento e licenciamento de programas de computador customizáveis ; Consultoria em tecnologia da informação
• City: Santo André
• Associated researchers: Rafael Amatte Bizão ; Tiago Botari
• Associated scholarship(s): 21/08091-1 - Application of artificial intelligence to aid in the diagnosis and prognosis of SARS-CoV-2, BP.TT; 21/03863-6 - Application of artificial intelligence to aid in the diagnosis and prognosis of SARS-CoV-2, BP.TT; 21/03661-4 - Application of artificial intelligence to aid in the diagnosis and prognosis of SARS-CoV-2, BP.PIPE

Abstract

The medical diagnostic requires the health professionals to analyze a large amount of data that include exams, patient records and other relevant information. These information are result of technological advances that produces a growing volume of data in which its analyzes is becoming increasingly complex. Furthermore, some situations demand fast responses from the health professionals, as in the recent SARS-CoV-2 pandemic. Thus, there is a lack of tools that support the medical professionals in clinical diagnosis. Among the modeling techniques currently used, machine learning is showing good results as a complementary tool, helping health professionals in achieving fast, precise and efficient diagnosis. When developed together with health professionals, these techniques may provide more precise diagnosis results and reduce costs. This research project aims to develop complementary methods for decision making in medical environments when SARS-CoV-2 infection is suspected in a patient. Mathematical and computational modeling will be developed using machine learning algorithms and other techniques of data science. State-of-the-art methodologies in computational visualization and interpretability/explainability will be used in the models to ensure transparency in the decision making process. The implementation process will be closely followed by health professionals to ensure that the solution will be adequate to the Brazilian medical environment. The data used in the machine learning models will be obtained from public databases and from future partners, to ensure that the model will be representative. Some stages of the project have already been done to show its viability. The results obtained so far show good possibilities in helping doctors in SARS-CoV-2 diagnosis. The product resulted from this project aims to create and implement mathematical models in an online platform already established (www.testedecovid.com), helping on SARS-CoV-2 diagnosis by making them fast and efficient. The technology that will be developed in the project meets expectations of the medical area and can be implemented in hospitals, laboratories and medical centers. (AU)