Ghosting of pulmonary nodules with respiratory motion: comparison of helical and conventional CT using an in vitro pediatric model.

Abstract

The study was designed to compare helical CT with varying pitch and reconstruction intervals and conventional CT for revealing pulmonary nodules in a model that simulates respiratory motion in children. CT scans were obtained in an experimental model with one nodule (3 or 10 mm) in each scan. One-second scans were obtained at rates of 10, 20, and 30 respirations per minute using conventional CT with 4-mm collimation and table incrementation and helical CT with 4-mm collimation and either 4-mm/sec (pitch, 1:1) or 8-mm/sec (pitch, 2:1) table speed. Reconstructions were at 1-, 2-, and 4-mm intervals for scans obtained using 4-mm/sec table speed and at 1- and 4-mm intervals for scans obtained using 8-mm/sec table speed. Images were independently reviewed by three radiologists who estimated the number of nodules on each image. Ghosting (depiction of more than one nodule in a study) was seen in 79%, 80%, and 75% of helical CT scans obtained with a 1:1 pitch using 1-, 2-, and 4-mm reconstruction intervals, respectively. By comparison, ghosting was seen in only 54% and 58% of helical CT scans with a 2:1 pitch using 1-mm reconstruction intervals and 4-mm reconstruction intervals, respectively, and in 56% of conventional CT scans (p < .0001). A single nodule was detected on all other scans, and at least one nodule was seen on all scans. Ghosting of nodules is common in this model. Ghosting was seen less often on conventional scans and helical scans with 2:1 pitch than it was on helical scans with 1:1 pitch. Nonetheless, ghosting was seen on more than 50% of all scans with each technique.