Diagnostic Performance and Dose Comparison of Filtered Back Projection and Adaptive Iterative Dose Reduction Three-dimensional CT Enterography in Children and Young Adults.

Abstract

To investigate diagnostic performance and radiation dose with the use of computed tomographic (CT) enterography in children and young adults and to compare reconstruction with filtered back projection (FBP) to that with adaptive iterative dose reduction (AIDR) with three-dimensional (3D) processing. This retrospective investigation was HIPAA compliant and approved by the institutional review board. Informed consent was waived. CT enterographic examinations performed between October 2008 and June 2009 with FBP and between August 2012 and April 2014 with AIDR 3D in patients who had received histologic evaluation within 45 days of imaging were included. Two reviewers retrospectively and independently evaluated the studies for findings of active inflammation, and diagnostic performance and interreader reliability were assessed. The reference standard was histologic findings. Objective and subjective image quality also was assessed. The size-specific dose estimate was compared between the two groups. Two-sample t tests or analysis of variance tests were performed to assess for differences in diagnostic accuracy, image quality, and radiation dose between the FBP and AIDR 3D examinations. Fifty patients were included in the FBP group (mean age, 14.1 years; range, 8-21 years) and 68 patients were in the AIDR 3D group (mean age, 13.2 years; range, 2-29 years). Sensitivity and specificity for detection of active inflammation were 96% (26 of 27) and 96% (22 of 23), respectively, for the FBP group and 90% (45 of 50) and 89% (16 of 18), respectively, for the AIDR 3D group. Dichotomous interreader reliability (κ) for the entire group was 0.86. The mean size-specific dose estimate for all weights was significantly lower for the AIDR 3D group (6.1 mGy ± 2.1) than that for the FBP group (16.7 mGy ± 5.2; P < .0001). No significant difference was found in objective image noise for soft-tissue structures (P = .2-.8). CT enterography is highly accurate for detection of active inflammation in pediatric patients and has excellent interreader reliability. Reduced-dose CT enterography with AIDR 3D allowed substantial dose reduction compared with that used with FBP CT enterographic examinations, while maintaining a high diagnostic performance.