P191 Effects of CT reconstruction algorithm on the quantitative assessment of Crohn's Disease: A comparison of standard and low dose CT enterography
G. Rosenfeld1, J. Brown2, P. Vos2, J. Leipsic2, R. Enns3, B. Bressler*3
1University of British Columbia, Medicine, Vancouver, Canada, 2University of British Columbia, Radiology, Vancouver, Canada, 3University of British Columbia, Division of Gastroenterology, Vancouver, Canada
Traditional Computed tomography (CT) imaging relies on filtered back projection (FBP) for image reconstruction. Newer iterative reconstruction (IR) techniques use computer modeling to generate CT images with lower radiation exposure while retaining sufficient quality for accurate interpretation. CT Enterography (CTE) is the preferred imaging modality for assessing small bowel Crohn's Disease (CD). We report here preliminary findings comparing the validity of 2 low dose CTE techniques for evaluating inflammatory changes in patients suspected of having CD.
In total, 163 patients referred for diagnostic CTE to evaluate CD will undergo a standard exam and a low dose CTE exam in a random sequence on the same day. The standard exam was used to obtain FBP and adaptive statistical IR (ASIR) images and the low dose scans, Model based IR (MBIR) images. Demographics, historic data and fecal calprotectin were collected from each patient. The Harvey Bradshaw Index (HBI) was used to determine clinical disease activity. De-identified, randomly ordered images were reviewed by 2 experienced radiologists, independently for signs of small bowel CD and an overall assessment of "active" or "inactive" was made. All available clinical data including fecal calprotectin was reviewed by an experienced Gastroenterologist to determine the presence or absence of disease activity which served as the gold standard for comparison with the CT findings.
As of Nov. 2014, a total of 124 patients had been enrolled and underwent CTE scanning. 46% of the subjects were male with a mean age of 43.6 (SD15.7 yrs). The mean HBI was 4.5, 48.4% were clinically assessed to have active disease by the gastroenterologist. The Generalized estimating equations (GEE) method was used to compare the validity of the reconstruction algorithms (Table 1).
“Comparison of the validity of CTE Reconstruction Techniques”
These values all suggest non-inferiority, 95% confidence intervals cross the pre-specified non-inferiority margin of 0.1. Completion of enrolment is expected to produce narrower confidence intervals demonstrating no significant difference between the 3 techniques. Standard dose radiation exposure was 6.7 ± 3.8 mSv and low dose exposure was 2.9 ± 2.4 mSv (p<0.001). More frequently, image noise was rated moderate or higher (28.4 - 84.7%) for MBIR compared to only 2.6 - 4.5% for ASIR and FBP.
Iterative reconstruction techniques for CTE have similar validity to traditional CTE techniques, at much lower radiation doses, for the identification of active small bowel inflammation in patients with Crohn's disease.