TY  - GEN
AV  - public
A1  - Klein, Laura Victoria
TI  - Patient Risk-Minimizing Tube Current Modulation in X-Ray Computed Tomography
CY  - Heidelberg
N2  - This dissertation proposes a patient-specific tube current modulation for computed tomography (CT) that minimizes the individual patient risk (riskTCM). Modern CT scanners use automatic exposure control (AEC) techniques including tube current modulation (TCM) to reduce the radiation dose delivered to the patient while maintaining image quality. Today's TCM implementations aim at minimizing the tube current-time (mAs) product as a surrogate for patient dose, which is why they are referred to as mAsTCM hereafter. However, the actual patient risk, e.g., in the form of risk measures such as the effective dose Deff representing the sensitivity of individual organs with respect to ionizing radiation, is not taken into account.
In order to be able to optimize the effective dose Deff or another biologically meaningful measure, organ doses must be estimated before the actual CT scan in order to compute an optimized riskTCM curve. This can be achieved using a machine learning approach and based on these information, the new patient risk-minimizing TCM curve can be obtained. The proposed riskTCM algorithm was evaluated in a simulation study for circular scans and compared against the current gold standard method mAsTCM and to a constant tube current as well as an organ-specific tube current modulation technique.
The results illustrate that all anatomical regions can benefit from riskTCM and a reduction of effective dose of up to 30% can be expected compared to mAsTCM. Furthermore, riskTCM was extended to a spiral trajectory that is commonly used in clinical routine and initial measurements with phantoms have been performed. The introduction of riskTCM into clinical practice would only require a software update since almost all CT systems are already capable of modulating the tube current.
ID  - heidok34016
Y1  - 2023///
UR  - https://archiv.ub.uni-heidelberg.de/volltextserver/34016/
ER  -