Return To Abstract Listing

Scientific Session 19 — Efficacy/Administration/Informatics - Safety/Quality

Thursday, May 4, 2017

Abstracts 2298-3291

2830. Patient Miscentering During CT: Findings From 57,621 Acquisitions and Impact on Dose

Akin-Akintayo O*,  Alexander L,  Neill R,  Krupinski E,  Moreno C. Emory University, Atlanta, United States

Address correspondence to O. Akin-Akintayo (

Objective: This study was performed to evaluate patient miscentering during CT of the chest, abdomen, and pelvis and to assess differences based on sex, patient position, scan region, equipment, and impact on dose.

Materials and Methods: Institutional review board approval was obtained for this HIPAA-compliant study. Data from adult chest, abdomen, and pelvic CT acquisitions performed between March 2015 and June 2016 were evaluated. Variables recorded included patient sex and position (e.g., supine or prone), scan region (e.g., chest, abdomen, and pelvis), equipment, and volume CT dose index. Distance between patient centroid and gantry isocenter in the x- and y-axes and patient diameter were determined using RadiMetrics (Bayer Healthcare). Data were analyzed with an ANOVA using the centroid-to-isocenter data as the dependent variable and the other parameters as independent variables.

Results: Single-phase CT acquisitions of 57,621 patients were analyzed (57.7% women; 42.3% men). Vertical miscentering did not differ significantly for men compared with women (p = 0.8) overall, although there was a significant interaction effect for centroid to isocenter and scan region plus sex. Mean centroid to isocenter in the y-axis varied significantly for supine versus prone positioning (supine mean ± SD, 1.7 ± 1.8 cm below isocenter; prone mean ± SD, 0.8 ± 1.8 cm above isocenter [p < 0.05]). Mean centroid to isocenter y-axis was greatest for chest and abdomen scans (mean ± SD, 1.9 ± 1.7 cm below isocenter) and least for scans of the abdomen only (mean ± SD, 1.4 ± 1.6 cm below isocenter). Mean centroid to isocenter varied significantly on the basis of device used (p < 0.05). No linear relationship was found between centroid to isocenter and patient diameter. A strong inverse linear relationship was found between centroid to isocenter y-axis and CTDIvol (–0.90), and no correlation was found between centroid to isocenter x-axis and CTDIvol (–0.007).

Conclusion: Patient centering can be influenced by a number of factors including patient position (supine or prone), scan region, and equipment with implications for patient dose. Technologists tend to center patients below the isocenter of the CT gantry.