Artificial Intelligence Solution to Support Diagnosis and Screening of Head Tomography Exams

• Grant number: 24/10437-1
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: February 01, 2025
• End date: January 31, 2027
• Field of knowledge: Engineering - Biomedical Engineering - Medical Engineering
• Principal Investigator: Daniel Aparecido Vital
• Grantee: Daniel Aparecido Vital
• Company: Harpia Health Solutions Ltda
• CNAE: Atividades de atendimento hospitalar ; Atividades de serviços de complementação diagnóstica e terapêutica ; Atividades de apoio à gestão de saúde
• Pesquisadores principais: Catarina Cardoso Reis
• Associated research grant(s): 24/23395-5 - Artificial Intelligence Solution to Support Diagnosis and Screening of Head Tomography Exams, AP.PIPE
• Associated scholarship(s): 25/09751-6 - Development and Validation of Models for Ischemic Stroke and Assessment ASPECTS, BP.TT; 25/00022-1 - Artificial intelligence solution to support diagnosis and screening of head tomography exams, BP.PIPE

Abstract

Cerebrovascular accident (CVA) is one of the main causes of death and disability in Brazil and around the world. According to the Ministry of Health, Brazil records around 400,000 new cases of stroke per year, resulting in approximately 100,000 deaths annually. Furthermore, one in six people will have a stroke in their lifetime, and around 70% of survivors are left with sequelae that compromise their ability to carry out daily activities. Skull Computed Tomography (CT) is the exam of choice for the quick and accurate diagnosis of neurological conditions such as stroke, and is essential for defining appropriate treatment. It is estimated that, in Brazil, more than 6 million cranial CT exams are performed annually, reflecting the high demand for fast and accurate neurological diagnoses. Rapid diagnosis is crucial for effective stroke treatment, significantly reducing sequelae and mortality. The therapeutic window for interventions such as thrombolysis is extremely short, generally up to 4.5 hours from the onset of symptoms. Because the vast majority of head CT examinations for stroke cases are performed urgently in hospitals, a correct diagnosis, including an accurate understanding of the affected area, is crucial to determining the appropriate care protocol and initiating the correct treatment as soon as possible. as soon as possible. However, studies indicate delays and inconsistencies in diagnosis due to the exhaustive working hours of professionals and the scarcity of qualified specialists in the most affected regions to deal with emergency cases. To overcome this situation and mitigate risks, computational solutions based on Artificial Intelligence have been developed and applied to highlight information in medical image exams. By providing additional, precise and objective information, these solutions are increasingly present in hospital care, making it possible to increase diagnostic speed and accuracy, in addition to supporting clinical decision-making. Therefore, the general objective of this project is to develop a computational solution based on AI for screening and supporting the diagnosis of Skull Tomography exams, focused on the detection and analysis of neurological conditions. The tool will consist of a series of methods and computational models that will be researched, developed and improved by the development team, aiming to detect and analyze hemorrhages, identify ischemic stroke and provide assessment of ASPECTS, detect cerebral aneurysms and identify skull fractures. As a result, we aim to transform the landscape of neurological diagnosis in Brazil, significantly improving patient outcomes, with the potential to save lives, reduce disabilities and optimize hospital resources, meeting a critical need in public health. (AU)