Artificial Intelligence-Assisted Breast Cancer Detection Through Tumor Density Estimation from Histopathological Images

Computer-assisted breast cancer detection is a useful and widely used technique that aids pathologists in clinical diagnoses. Presently, identifying breast cancer is primarily reliant on a single imaging trait. This study contributes to the creation of an artificial intelligence model that can forecast information on the density of tumor cells. Multiple AI models have been experimented with for estimating the tumor density, and the best model was selected by comparing the results of all the models and iterations. Our analyses reveal that the Resnet 101 model, which has all layers trainable and by using the learning rate 1e-3, outperforms baselines on normalized histopathological validation picture data than the model obtained by performing hyperparameter tuning. Commonly, doctors or pathologists with extensive experience are frequently advised to analyze a sample and determine the tumor density. The proposed AI model will support the process and will have a significant impact on the time taken for manual analysis, resulting in faster decision-making, and influencing the patient’s waiting time.