Effect of breathing on the detection of in vivo simulated pulmonary nodules by spiral CT

Objective: To determine the effect of breathing on the detection of small dense pulmonary nodules of threshold conspicuity by spiral computed tomography (SCT). Materials and Methods: Forty pulmonary nodules of high density were created by endobronchial deployment of 2 and 4 mm diameter beads in the peripheral airways of five anaesthetized dogs. SCT was performed during induced breath-hold and quiet breathing, using 5 mm collimation, pitch 2 and reconstruction of contiguous 5 mm slices. Scans were reviewed by six radiologists, Detection rates were measured as the number of nodules seen by at least one reader. The data was modelled using ordinal logistic regression for repeated measures, and the Wald Chi-square statistic used to test if there was a breath-hold vs. breathing effect on reader confidence level. Results: There was no difference in detection rates for breath-hold vs. breathing SCT (28 vs. 25, respectively; P = 0.48), This was also true when only the 2 mm nodules (n = 24) were considered (12 vs. 11, respectively; P = 0.77). Reader confidence level was significantly higher for breath-hold vs. breathing SCT (Wald Chi-square statistic with 6 degrees of freedom = 19.0; P = 0.0041). Conclusion: SCT can be performed during quiet breathing without a significant reduction the overall detection rate for small dense pulmonary nodules, though reader diagnostic confidence level is reduced.