Use of machine learning techniques for serial assessment of systemic inflammatory markers in breast cancer patients undergoing neoadjuvant chemotherapy

• Grant number: 24/08967-2
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: July 01, 2025
• End date: May 31, 2028
• Field of knowledge: Health Sciences - Medicine
• Principal Investigator: Afonso Celso Pinto Nazário
• Grantee: Sara Socorro Faria
• Host Institution: Escola Paulista de Medicina (EPM). Universidade Federal de São Paulo (UNIFESP). Campus São Paulo. São Paulo , SP, Brazil

Abstract

Breast cancer is a heterogeneous disease and can be divided into subtypes, according to the immunophenotypic and gene expression profile. Patients undergoing neoadjuvant chemotherapy present different degrees of tumor response, ranging from no response to complete response. In addition to already established clinicopathological characteristics, such as staging, systemic inflammation can also have an impact on outcomes. In blood counts, systemic inflammatory indices, identified by the absolute count of neutrophils, lymphocytes and monocytes, represent important tools related to chronic inflammation. Additionally, computer vision methods have been applied and developed in the epidemiological and biological fields, and mainly in the development of new tumor biomarkers. Objectives: Develop and evaluate predictive machine learning models, based on the absolute blood cell count of women with breast cancer undergoing neoadjuvant chemotherapy; Correlate the intensity of the tumor lymphocytic infiltrate with the neutrophil/lymphocyte ratio and the pathological response rate in women with breast cancer undergoing neoadjuvant chemotherapy; characterize the intensity of the inflammatory infiltrate by immunohistochemistry and the expression of CD8 and CD3 on the inflammatory cells that infiltrate the tumor; to correlate the intensity of the tumor lymphocytic infiltrate with systemic inflammatory indices and the pathological response rate in women with breast cancer undergoing neoadjuvant chemotherapy. Methods: This is a prospective and multicentric study, in which blood count data will be used before, during and prior to surgery in women with breast cancer undergoing neoadjuvant chemotherapy. Additionally, tumor infiltrating lymphocytes (TIL) will be analyzed before and after neoadjuvant chemotherapy. Machine learning and/or deep learning models will be implemented, evaluated and improved with the aim of predicting disease risk. (AU)