Building an open-data software package to enable data-driven innovation in breast CT

Breast cancer is one of the most diagnosed cancer types worldwide. Early detection achieved through dedicated screening programs (based on mammography or tomosynthesis) reduces the mortality rate. However, there is still a large number of missed cancers (22%) and false positives (46%) [1], that result in unnecessary follow-up procedures. In two-dimensional breast imaging, especially the tissue overlap can make images difficult to interpret for women with dense breast. 

This problem can be addressed by scanning the breast with a tomographic imaging device (CT scanner). Such dedicated breast CT (BCT) scanners were already introduced to clinical practice and allow to measure full volumetric information with a high resolution. Grating-interferometry-based CT scanners, as they are investigated in our group, not only exploit the attenuation of the X-rays in the breast, but also the beam refraction (phase contrast) within one single scan. Combining the information of both physical processes into one image, has shown the potential to outperform conventional (attenuation-based) BCT scanners [2]. 

In order to exploit the full potential of (GI-)BCT for clinical application and to accelerate its development, data taking of clinical samples with our setups and sharing the results with the research community is an essential step. Building up and maintaining a respective data base, which hosts reliable and well documented measurements of clinical samples, will allow to answer research questions on image quality and clinical benefit of GI-BCT with underlying statistical significance. 

For this reason, we started to develop the repository qRIP, the quantitative refraction imaging package, which will give access to existing and newly measured (GI-)BCT data from our and other research groups. The software package will allow easy handling of volumetric data, raw data (projections) and breast tissue properties. Furthermore, it will provide tools for image analysis, reconstruction and simulation. This will allow to accelerate the development of GI-BCT towards clinical applicability.