Can independent coronal multiplanar reformatted images obtained using state-of-the-art MDCT scanners be used for primary interpretation of MDCT of the abdomen and pelvis? A feasibility study

Abstract 

Purpose

To evaluate if coronal reformatted images can be used for primary interpretation of MDCT of the abdomen and pelvis using 64-slice MDCT.

Materials and methods

IRB approval was obtained. We reviewed MDCT studies of the abdomen and pelvis of 220 consecutive patients performed with 64 row MDCT with constant scanning parameters. Based on a 0.625mm raw data set, transverse images were reconstructed at 5mm and coronal images at 3mm using standard reconstruction algorithms. Reader familiarity was achieved by simultaneous evaluation of transverse and coronal reformats in an initial group of 20 separate cases for findings in consensus. Two subsequent phases of image analysis were then performed in two groups of 100 patients each. In the first phase two radiologists evaluated the added utility of simultaneous review of MDCT of transverse and coronal reformatted images over transverse images alone in 100 consecutive patients referred for MDCT of the abdomen and pelvis. In the second phase, the same radiologists evaluated whether coronal multiplanar reformats could be used for primary interpretation of MDCT of the abdomen and pelvis in a separate but similar cohort of 100 consecutive abdominopelvic MDCT studies. The number of lesion(s), their location, size of smallest lesion, presence of artifacts and likely diagnosis were noted at each image interpretation. Image quality and confidence for interpretation was evaluated using five-point and three-point scale, respectively. The time required for primary interpretation of coronal reformats and transverse images were recorded. Statistical analysis was performed using Wilcoxon signed rank test.

Results

Both readers detected additional findings (n=37, 35), respectively, on simultaneous review of transverse and coronal reformats as compared with transverse images alone (p<0.001). Excellent interobserver agreement was noted (r=0.94–0.96). Both readers detected additional findings (n=62, 53), respectively, on independent review of coronal reformats as compared with transverse images alone (p<0.001). Readers’ confidence was also found to be higher on coronal evaluations as compared to axial images (p<0.01). There was good interobserver agreement between the two readers.

Conclusion

Independent coronal multiplanar reformatted images obtained using state-of-the-art MDCT scanners show promise as the preferred orientation and can be useful for primary interpretation of MDCT of the abdomen and pelvis.

Keywords: 64-Slice MDCT, Coronal reformats, Transverse images, Primary interpretation