The impact of noise on the results of automated slice sensitivity profile measurements in computed tomography

Purpose: This study was aimed to investigate the impact of noise on full-width at half maximum (FWHM) of slice sensitivity profile (SSP) of computed tomography (CT) images which was automatically calculated using Matlab-based software on the AAPM CT performance phantom. Methods: The slice thickness part of AAPM CT performance phantom was scanned by the 128 slice GE Healthcare CT scanner with four different tube currents of 50 mA, 100 mA, 150 mA and 200 mA. The noise calculation was carried out on the homogeneous region of the phantom which contains water. The noise was calculated as the standard deviation (SD) within the region of interest (ROI). The automated measurement of FWHM of SSP was carried out on the middle stair object of the image. Results: The noises were 7.67, 5.95, 4.98, and 4.00 HU for tube currents of 50, 100, 150, and 200 mA, respectively. The FWHMs of SSPs were 5.5 ± 0.1, 5.3 ± 0.3, 5.4 ± 0,0, and 5.2 ± 0.1 mm for noises of 7.67, 5.95, 4.98, and 4.00 HU, respectively. The result of automated FWHM of image with noise of 7.67 HU was 5.5 mm and the difference from the nominal slice thickness was 10%. The FWHM of image with noise of 4.00 HU was 5.2 mm and the difference from set slice thickness was 4%. The automated FWHM results tended to be similar to the nominal slice thickness for lower noise. Conclusions: The automated FWHM results tends to be similar to the nominal slice thickness for lower noise. Although the noise of 7.67 HU still produces an accurate FWHM of SSP, the results suggested to use low noise for measuring FWHM and SSP.