Congress:
EuroSafe Imaging 2020
Keywords:
Multicentre study, Experimental, Not applicable, Quality assurance, Education and training, Radiation safety, Physics, Observer performance, CT, Radioprotection / Radiation dose, Action 1 - Guidelines, implementation policies, and Clinical Decision Support (ESR iGuide)
Authors:
H. de las Heras Gala, A. Giussani, G. Brix
DOI:
10.26044/esi2020/ESI-05022
Description of activity and work performed
The individual steps for the calculation of the dose-efficiency-index (DEI) are shown in Fig. 1. A minimum of 200 images of the inserts and a similar amount of images without inserts has to be acquired for each set of exposure parameters. For each insert, the value of the corresponding area under the ROC curve (AUC), using the Hotelling observer with Laguerre-Gauss channels, is obtained. Repetition of this approach for 5 different devices of the same type provided a standard deviation of 2.2 % for the AUC (step 7) [1].
A weighted average of the results for the four inserts provides the weighted AUC (AUCw), which informs the image quality. The whole procedure has to be performed for at least three CTDIvol-values between 2 and 12 mGy (in case of the head phantom) both for a basic and for a high-end configuration of the considered device.
Data acquisition and subsequent calculations must be only performed once at the vendor factory. Based on the results of AUCw as a function of the CTDIvol, a reference value of CTDIvol is determined (Fig. 2). The standard head-CTDI-phantom is then scanned at that reference CTDIvol to assess the signal-to-noise ratio (SNR) for four ROIs (Fig. 3). The obtained results for CTDIvol and SNR serve as benchmark values for the on-site acceptance test.