Manual histopathological assessment of tumors is currently the main mode to establish breast cancer (BC) diagnosis. However, there is a shortage of pathology expertise in many parts of the world, and also a high inter-assessor variability between pathologists exists. This leads to prolonged response times, unnecessary patient anxiety, and unequal access to top-quality histopathology assessments for cancer patients. Furthermore, misclassifications can cause over- and under-treatment.
In this research we will develop state-of-the-art AI-based models for BC routine histopathology. Novel methodologies for stain-free and multi-stain histological analysis will also be developed. The project aims to improve the quality of BC histopathology assessments by reducing errors and inter-assessor variability, enhancing patient stratification and reducing over- and under-treatment of patients, while also contributing toward more efficient and reliable routine pathology.