Analysis and Comparison of the ROMA Index with Hemogram Data and Clinical Signs and Symptoms in Brazilian Women with Ovarian Cancer.

• DOI: https://doi.org/10.22533/at.ed.15949624291010

• Palavras-chave: câncer de ovário, biomarcadores, inteligência artificial

• Keywords: ovarian cancer, biomarkers, artificial intelligence

• Abstract: One of the main discussions about women's health is the efficiency and effectiveness of using simplified methodologies, such as isolated biomarkers, to support differential diagnosis. Based on this premise, we used a database to identify the specificity and sensitivity of the ROMA index in detecting ovarian cancer compared to an intelligent computational model. This model was trained with anamnesis and blood count data to differentiate subtle nuances between patients with a potential diagnosis of ovarian cancer. The premise is that this type of model can provide better groupings for risk factors, allowing for more accurate data classifications that can better describe early patterns of silent diseases, such as ovarian cancer, in gynecological clinics, or at least in the early stages, to reduce mortality.  Our goal is to identify clusters within data on signs and symptoms, combined with data on the immune and endocrine systems, that can group characteristics for better clinical outcomes in ovarian cancer. Our methodology was based on multicriteria analysis methods, achieving a sensitivity of 94.6% and a specificity of 97%, as measured by the kappa index. These results indicate the potential of our methodology to support physicians in the gynecological monitoring of women in Basic Health Units.