Carolina Mondino, a bioengineering graduate of the Instituto Tecnológico Buenos Aires (ITBA), developed a way to classify tumors through the creation of an algorithm. It is a tool designed from Artificial Intelligence to assist doctors in the detection of breast cancer.
The research was carried out within the framework of a final degree project that sought to generate methodologies that allow to develop solutions for the main cause of death of women who contract oncological diseases: breast cancer. In Argentina alone, in 2020 there were 6,821 deaths of women due to this cause. In this regard, it should be noted that doctors do not detect between 10% and 30% of breast cancer and that, if the patient is diagnosed and treated in time, she has up to a 90% chance of surviving in the next 5 years. The project was based on studies that showed that, based on convolutional neural networks (CNN), a computer can automatically classify tumors on mammograms based on the type and severity of the lesion, that is, whether they are benign or malignant tumors. To make this happen, the mammogram is first entered into the program so that it examines the tumors that are present in the program (if any). The system then returns the prediction of the type and severity of the tumor. The accuracy value obtained in the project was 70% for the severity of the injury. In the classification of calcifications or masses, promising results were obtained since 90% accuracy was reached, a value that is comparable with the research published today. Even the accuracy obtained is higher than some of the published research. Carolina Mondino’s project had bioengineer Matías Nazareth Tajerian as a tutor, who in addition to also being a graduate of ITBA, works as a developer at the Italian Hospital of Buenos Aires. “The objective of this project is to be able to add the algorithm to a software so that health professionals can use it as a complementary tool,” said the ITBA graduate. WHO recommends that all women over the age of 40 have an annual mammogram so that they can diagnose breast cancer early. In this line, the tool aims to help in the detection of breast cancer at the earliest possible stage in order to have the diagnosis and treatment. That way, to be able to reduce the mortality rate from breast cancer in women.