CT Tube Current and X-Ray Spectrum Modulation to Maximize the Patient Benefit-to-Risk-Ratio with Respect to Image Quality and Patient-Specific Exposure Risk

• Applicants: Professor Dr. Marc Kachelrieß; Professor Dr. Michael Lell; Professor Dr.-Ing. Andreas Maier
• Subject Area: Medical Physics, Biomedical Technology
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 551239003

Modern diagnostic CT systems include numerous technical measures to keep the patient dose as low as possible. Particularly important is the so-called tube current modulation (TCM), in which the tube current is modulated depending on the expected transmission length in such a way that the tube current-time product (mAs product) is minimal for a given image quality or that the image quality is maximal for a given mAs product. In our previous project, we developed TCM method based on this mAsTCM, which minimizes the patient-specific radiation risk (and not only the tube current-time product). This riskTCM uses neural networks that calculate the patient's attenuation distribution and the segmentation of the radiation-relevant organs from the topograms available before the CT scan. Our riskTCM was simulatively studied and compared with mAsTCM. While maintaining image quality, riskTCM reduced the individual radiation risk by up to 40% compared to mAsTCM. These significant dose reduction values are also interesting because, on the part of the manufacturer, only a software adaptation of the CT systems is necessary to implement riskTCM. We now want to explore two things. On the one hand, we want to bring the riskTCM method closer to the clinical image quality with the help of a manufacturer-specific software called ReconCT, in order to be able to evaluate the method scientifically in clinical operation (translation). On the other hand, we want to extend the tube current modulation to also modulate the X-ray voltage and the pre-filtering of the CT system to further reduce the individual radiation risk. This riskXSM method requires new optimization algorithms, it requires the development of new reconstruction methods and it is necessary to calculate from a topogram and the surface information of a 3D camera, not only the attenuation value distribution of the patient, but also still possible contrast agent distributions (extension). To test riskXSM, diagnostic data will be collected, converted to the desired tube current, voltage and pre-filter density curves by adding noise and converting beam hardening, and then evaluated by experienced radiologists in comparison with mAsTCM and riskTCM.

• DFG Programme: Research Grants