TU-H-207A-04: How Patient Positioning in CT Affects More Than the AEC: Image Noise Uniformity and CT Number Changes as a Function of Positioning




Studies dealing with patient positioning in CT have mainly dealt with dose and image noise changes. For diagnosis, only dose decreases can negatively affect the diagnostic utility of the exam via an increase in image noise. However, a little studied consequence of poor positioning is CT number and noise uniformity changes.


An anthropomorphic phantom was imaged in 2 cm vertical steps between the highest and lowest couch positions. CT number and noise were measured using regions of interest (ROI) in the upper-thorax, mid-thorax, and abdomen. The degree of noise non-uniformity and CT number change was assessed by comparing values obtained in the anterior vs. posterior regions. Additionally, to add clinical support for the study, we analyzed 5,746 abdominal patients at our institution and obtained their lateral and vertical mis-positioning from iso-center. The patients ranged in size from teens to large adults. Box and whisker plots were used to identify the range of patient poisonings seen in our clinic.


Absolute CT number changes between anterior and posterior ROIs over 10 HU were measured for ROIs at 4 cm off-centering. Noise uniformity varied by greater than 2 times for all regions at an offset of −10 cm. The 25–75 percentile ranges for patient positioning was −.91 – 1.4, −.87 – .88, −1.3 – .93, −1.8 – .60 cm for pediatrics, small adults, medium adults, and large adults respectively. However, 1.4% of our patients had mis-positioning errors outside of 4 cm.


The CT number changes shown here are large enough to warrant further investigation on how positioning might affect clinical decision making that relies on CT number for pathology characterization, especially due to the fact our clinical positioning data demonstrates we will have HU differences on the order of 10–20 HU.

Research support GE Healthcare. TPS supplies CT protocols to GE under a licensing agreement.