Abdominal pain: coronal reformations from isotropic voxels with 16-section CT--reader lesion detection and...

To retrospectively assess if reader detection of intraabdominal pathologic findings on coronal reformations from isotropic voxels at 16-section computed tomography (CT) was similar to reader detection on transverse scans.

The institutional review board approved this HIPAA-compliant study, and a waiver of informed consent was obtained. Twenty-nine consecutive patients (12 men, 17 women; mean age, 48 years; age range, 21-93 years) with abdominal pain underwent 16-section CT with coronal reformations. Eight independent readers reviewed randomized scans (transverse and coronal) and identified pathologic findings in multiple organ systems. Timing for each interpretation was recorded. One month later, readers reviewed the scan reformatted in the other imaging plane. Agreement between transverse and coronal scans was measured by using Cohen kappa coefficients.

Agreement was moderate to near perfect between transverse and coronal interpretations for intraabdominal anatomic and pathologic findings (kappa=0.59-1.00). For transverse interpretations, more thoracic pathologic findings were noted than for coronal interpretations; for coronal interpretations, more lymph nodes were noted than for transverse interpretations. Mean transverse interpretation time was 4.9 minutes+/-1.1 (standard deviation) (range, 2.9-6.5 minutes); mean coronal interpretation time was 5.1 minutes+/-0.8 (range, 3.3-6.7 minutes). For each reader, there was no statistically significant difference in interpretation time between transverse and coronal scans (P=.06).

With regard to the presence of intraabdominal pathologic findings, coronal reformations from isotropic voxels are similar to transverse scans in terms of interpretation time and reader agreement.