Document Type : Original Paper
Department physic, faculty of sciences Rabat, University mohamed V
Hassan First University of Settat, High Institute of Health Sciences, Laboratory of Sciences and Health Technologies, Settat, Morocco
Departement of Physics, Laboratory of High Energy Physics, Modelling and Simulation, Faculty of Science, Mohammed V Agdal University, Rabat, Kingdom of Morocco
Mammographic density is an important risk factor for breast cancer. The assessment of breast density has been considered a very important part of the process of breast diagnosis.
Materials and methods
The application of artificial intelligence in medicine allows analyzing data extracted from the medical reports within the routine clinical evaluation, identifying patterns of treatment patients, and developing interventions for patients.
This study used patient data and predicted patient density in mammogram exams, based on five parameters measured by Volpara software. Next, a model based on the Multi-Layer Perceptron Neural Network was trained to predict patient density in a BIRADS category.
The artificial neural network diagram was identified by the corresponding area under the ROC curve. The model predicted the density of patients with a correct classification rate of 98.2%.
The object of this study was to determine the effectiveness of artificial neural networks in predicting patient breast density, based on data collected from patient data DICOM in a University hospital.