Micronucleus analysis through AI in breast cancer patients

• Grant number: 25/01151-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2025
• End date: November 30, 2025
• Field of knowledge: Biological Sciences - Morphology - Cytology and Cell Biology
• Principal Investigator: Andréa Cristina de Moraes Malinverni
• Grantee: Ana Beatriz Paiva Primo
• Host Institution: Escola Paulista de Medicina (EPM). Universidade Federal de São Paulo (UNIFESP). Campus São Paulo. São Paulo , SP, Brazil

Abstract

The micronucleus test (MNT) has been widely used in human biomonitoring due to its sensitivity in detecting DNA damage caused by genotoxic and cytotoxic agents. In patients with breast cancer, a significant increase in the frequency of micronuclei has been observed. Therefore, this project aims to analyze micronuclei using artificial intelligence (AI) in patients with breast cancer, optimizing time and ensuring greater reliability of results. To achieve this, a prospective study will be conducted, applying the Free and Informed Consent Form (ICF) to thirty patients diagnosed with breast cancer and thirty healthy participants (undiagnosed), with both groups being age-matched. The collection of oral exfoliated cells will be performed by scraping the mucosa from the inner cheek. For each sample obtained, 2,000 cells will be analyzed using the micronucleus test to identify cell morphology and quantify both the percentage of micronuclei and metanuclear anomalies. Subsequently, the images will be used to train artificial intelligence models, including ChatGPT 4.0, Gemini 2.0, and Google AI Studio, with the goal of evaluating the effectiveness of these tools in identifying image patterns containing micronuclei. This study aims to determine which AI model is more efficient in detecting micronuclei across different molecular subtypes and to compare the results of the trained models with manual analyses, paving the way for future applications in breast cancer prognosis.