OncoData: artificial intelligence and machine learning for cancer pathology

• Grant number: 21/14596-9
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: November 01, 2022
• End date: July 31, 2023
• Field of knowledge: Health Sciences - Medicine
• Principal Investigator: Viviane Teixeira Loiola de Alencar
• Grantee: Viviane Teixeira Loiola de Alencar
• Company: Oncodata Soluções Tecnológicas Ltda
• CNAE: Atividades de serviços de complementação diagnóstica e terapêutica
• City: São José dos Campos
• Associated research grant(s): 23/11600-0 - OncoData: artificial intelligence and machine learning for cancer pathology, AP.PIPE
• Associated scholarship(s): 23/04195-2 - OncoData: Artificial intelligence and machine learning for cancer pathology, BP.TT; 22/13935-7 - OncoData: Artificial intelligence and machine learning for cancer pathology, BP.TT; 22/12952-5 - OncoData: Artificial intelligence and machine learning for cancer pathology, BP.TT

Abstract

Pathology is essential for cancer diagnosis and appropriate treatment definition. It consists in a subjective analysis of a pathology slide based on predefined image patterns, and its accuracy is directly correlated to the examiner's previous experience and background. Therefore, variability among pathology reports are more common than expected and occur in up to 25% of cases, when slides are re-examined by an experienced pathologist. This is extremely relevant as it may potentially impact the correct diagnosis and treatment of patients with serious illnesses, such as cancer, leading to great social and financial burden. Furthermore, although cancer incidence is increasing, evidence shows progressive shortage of pathologists, which leads to delayed release of pathology reports and consequent delayed treatment initiation. In Brazil, the turnaround time for an initial pathology report is up to 3-4 weeks, especially in remote areas. If we take into account the growing number of biopsies performed in our country, this becomes an urgent issue. New technologies have emerged as a way to solve relevant problems in healthcare. Literature and international commercial data show promising use of computational vision for cancer pathology diagnosis. The initial research of our project is to develop an artificial intelligence software capable of correctly differentiating between the main histological subtypes of lung cancer. The purpose of the algorithm is to aid pathologists in analyzing slides faster and more accurately, especially difficult cases. Our vision to the company is to develop, in the future, an advanced technology, capable of generating personalized predictions for the treatment of cancer patients by analyzing pathology data and correlating it to clinical and genomic characteristics. Therefore, this technology presents great potential to create a positive impact in Brazil's healthcare system, both private and public. (AU)