AI for Female Cancer Imaging

Project coordination

Danube Private University, Prof. Dr. Ramona Woitek

Project partners

• Uzhgorod National University, Uzhgorod, Ukraine

Researchers involved at DPU  

• Palak Handa, PhD
• Laura Villazan Garcia, BSc

Abstract

Aggressive breast and ovarian cancer respond to neoadjuvant chemotherapy (NACT) only in around 50% of cases and to date no non-invasive biomarkers of treatment response have been well established. Such biomarkers could aid in clinical decision making and help target treatment decisions to patients individually. 

The Research Center MIAAI at the DPU and Uzhhorod National University (UzhNU) will jointly develop AI-based prediction models of response to NACT based on radiological imaging (ultrasound, mammography, magnetic resonance imaging and computed tomography) and other medical data acquired before treatment. These AI models will be based on radiomics features extracted from imaging together with clinical, demographic and histopathological data. External validation of AI models is crucial for an algorithm’s generalisability and robustness. Therefore, this joint project specifically aims to validate AI predictors in collaboration with the UzhNU on external data that have not been used for the training and testing of AI models. A multicentric data base of female cancer will be created with images from academic and tertiary care centers collaborating with the DPU such as the University of Cambridge, UK, the Fondazione Policlinico Unversitario Gemelli in Rome, Italy, and UzhNU in Ukraine. Patients who underwent NACT for breast or ovarian cancer will be identified at the UzhNU and their demographic, clinical, histopathological and radiological imaging data will be collected. UzhNU and DPU will work jointly on the development and external validation of AI prediction models for aggressive breast and ovarian cancer to further enhance personalised cancer care for women.