Radiation dose and image quality in CT - Evaluation of how slice thickness, tube current modulation and reconstruction algorithms affect radiation dose and image noise
MetadataVis full innførsel
- Institutt for fysikk 
Background: Helse Vest has currently standardized the computed tomography (CT) Thorax/Abdomen/Pelvis protocols in the region. One of the standardized parameters was a reduction of the reconstructed slice thickness from 5 mm to 3 mm. Brooks formula, which describes the relationship between radiation dose, image noise and slice thickness in CT, has been used to estimate the effects of the slice thickness reduction on radiation dose and image noise. Exposure parameters like the tube voltage (kV) and tube current-time product (mAs) are currently not standardized. Aim: The aim of this thesis was to investigate the effects of reducing the reconstructed image slice thickness for CT Thorax/Abdomen/Pelvis protocols in Helse Vest. This was done by investigating how slice thickness, automatic tube current modulation (ATCM)and reconstruction algorithms affects radiation dose and image noise on four different CT scanners, and compare the results with the estimates made by Brooks formula. Materials and methods: An ATCM phantom was scanned with Thorax/Abdomen/Pelvisprotocols on four CT scanners from GE, Philips, Siemens and Toshiba. The protocolswere based on recommendations from the American Association of Physicists in Medicine(AAPM). Scans were performed both with a reconstructed slice thickness of 5 mm, andwhile keeping either the radiation dose or image noise constant when reducing the slice thickness from 5 mm to 3 mm. The radiation dose and image noise was kept at the desired level by adjusting the ATCM index for each scanner according to Brooks formula. All images were reconstructed with both filtered back projection (FBP) and iterative reconstruction(IR). The obtained image series were analyzed with respect to dose by looking at the volume CT dose index (CTDIvol) and mAs, and with respect to image noise by looking at the standard deviation (SD) noise in the z-direction, noise power spectrum (NPS) curves and inter-image SD noise maps. Results: The changes in radiation dose and image noise caused by the slice thicknessreduction appeared as predicted by Brooks formula, and the changes were similarwhen using both FBP and IR. When comparing a single protocol reconstructed with FBPand IR, the use of IR reduced the radiation dose on GE, and reduced the image noise on Philips, Siemens and Toshiba. When using IR on Toshiba, the amount of noise reduction performed by the IR algorithm was adapted according to the amount of noise present in the image. Toshiba applied the chosen ATCM index to a pre-defined slice thickness of 5 mm, regardless of the reconstructed slice thickness chosen by the operator. Conclusion: When image slice thickness was reduced in CT protocols, Brooks formulaprovided reliable predictions for the behavior of radiation dose and image noise, both when using FBP and IR as reconstruction methods. Due to differences found for different vendors, Helse Vest should, if they wish to further standardize the exposure parameters for their CT Thorax/Abdomen/Pelvis protocols, proceed carefully and pay particular attention to how the IR algorithms behave on the different scanners, and how each scanner applies the input exposure parameters.