Dr. Andrzej Liebert
Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU)
Virtual contrast agents for dynamic breast MRI- Unlocking the transformative potential of digitization for early breast cancer detection with breast MRI.
Approximately 1 in 8 women will face a breast cancer diagnosis during her lifetime. This makes breast cancer the most common cancer among women. Voluntary mammography screening allows women between the ages of 50-69 to have a breast cancer screening exam using X-ray mammography. X-ray mammography represents a long-established diagnostic procedure for detecting suspicious changes in the breast. Nevertheless, X-ray mammography can be limited in its informative value, especially in women with increased breast density. For this reason, too, modern diagnostic procedures, such as magnetic resonance imaging (MRI), are being investigated as a supplement or even alternative to X-ray mammography. MRI of the breast (known as breast MRI) provides imaging without the use of ionizing radiation or breast compression and has been described as having a higher cancer detection rate than X-ray mammography. However, breast MRI routinely requires intravenous administration of gadolinium-containing contrast agents for dynamic examination of tissue perfusion, which is mostly impaired in abnormal lesions. However, the contrast media can cause side effects that are rare but relevant for a screening program. Furthermore, they cause additional relevant financial and periprocedural personnel expenses related to the application. Environmental aspects of gadolinium use on contamination of water bodies and the manufacturing process are also increasingly under investigation. All these aspects complicate the implementation of breast MRI in population-based breast cancer screening programs.
This project will therefore develop and translationally research a novel digital technology (called “virtual dynamic contrast-enhanced MRI [vDCE]”). The vDCE technology should be able to derive a contrast-analog and dynamically evaluable image contrast from a selective combination of non-contrast sequences by applying artificial intelligence (AI) algorithms. Such an imaging technique could unlock the transformative application potential of AI for breast MRI without having to interfere with established reporting routines and assessment standards, thereby significantly strengthening the use of breast MRI in early breast cancer detection. The focus of this project is to explore technical advancements of vDCE technology by applying the latest AI architectures. This should pave the way for the following steps on the way to an evaluation in the context of translational studies, in order to further develop the application potential of the innovative technology in the future.