Renal stone detection using unenhanced multidetector row computerized tomography--does section width matter?

We determined the effect of reconstructed section width on sensitivity and specificity for detecting renal calculi using multidetector row computerized tomography.

Three to 5 renal stones 2 to 4 mm in size were randomly placed into 14 human cadaveric kidneys and scanned by 16-row detector computerized tomography at 1.25 mm collimation and identical scanning parameters. After acquisition images were reconstructed with a section width of 1.25, 2.5, 3.75 and 5.0 mm, and reviewed independently by 2 blinded radiologists. Comparisons of sensitivity and specificity between different section widths were assessed with the McNemar test and Cochran's Q statistics.

Specificity was not significantly affected by section width (94.6% to 97.7%). In contrast, sensitivity increased as stone size increased and as section width decreased. Sensitivity to detect all stones was 80.7%, 80.7%, 87.7% and 92.1% for 5.0, 3.75, 2.5 and 1.25 mm section widths, respectively. Interobserver agreement for stone detection was excellent (kappa 0.858). Although the 2.0 mm stone detection rate improved with thinner section widths (79.4% vs 52.9% for 1.25 vs 5.0 mm, p = 0.004), stones greater than 2.0 mm were similarly detected at different slice selections (p = 0.056 to 0.572).

Independent of other scanning parameters reconstruction section width influences the ability to detect small renal calculi. It must be considered when creating computerized tomography protocols.