ECR 2013 Rec: Improved image quality for higher diagnostic accuracy of cranial computed tomography using iterative image reconstruction #B-0889 #SS1811a
B-0889 Improved image quality for higher diagnostic accuracy of cranial computed tomography using iterative image reconstruction
H. Haubenreisser, C. Fink, P. Apfaltrer, B. Schmidt, M. Sedlmair, S.O. Schönberg, T. Henzler | Monday, March 11, 14:00 – 15:30 / Room B
Purpose: To prospectively compare the image quality of cranial computed tomography (cCT) with thin slice widths using traditional filtered back projection (FBP) and sinogram-affirmed iterative image reconstruction (SAFIRE).
Materials and Methods: 40 consecutive studies (19 men; 71.6±16.6 years) referred for cCT were prospectively included. Each cranial CT raw data set was reconstructed with FBP and SAFIRE with decreasing slice widths (5 mm-1 mm). Objective image quality was assessed by measuring image noise in three predefined regions of the brain (white matter, thalamus, cerebellum) using identical regions of interest (ROIs). Subjective image quality was assessed by 2 experienced radiologists by ranking the reconstructed data sets with respect to overall image quality. The Mann-Whitney U-test and Cohen’s Kappa were used for statistical analysis.
Results: Image noise was statistically significantly reduced in all SAFIRE images at identical slice widths when compared to the images reconstructed with FBP (4.26±0.43 HU vs. 7.67±1.19 HU at 1 mm slice width) (p<0.001). Mean signal attenuation for each region and slice width remained constant between the two reconstruction methods (p>0.5). SNR was comparable between 1 mm SAFIRE images and 5 mm FBP images. Subjective image quality of SAFIRE images was rated consistently higher than that of the FBP images (p<0.001). Interobserver agreement was excellent between both radiologists (Cohen’s K = 0.79-0.86).
Conclusion: Iterative image reconstruction significantly reduces image noise, while increasing image quality. In cCT this may be used to decrease slice width and thus reduce partial volume effects, which may lead to increased diagnostic accuracy.