Automated Computer Software Compared with Manual Measurements for CT-Based Urinary Stone Metrics: An Evaluation Study

Stone size guides treatment decisions, yet there is no standard method for measuring stone size. Prior work has shown significant variability in manual stone measurements. We tested a novel stone software program designed to provide an automated and objective comprehensive CT-based stone profile.

Urinary stones identified on CT imaging were manually measured to obtain linear size and maximal stone density (in Hounsfield unit [HU]). Manual stone volume was calculated using the formula 0.52 × length × width × height. The same stones were assessed with computer software capable of automatically providing stone length, density, and volume. Computer measurements were compared with manual measurements.

Eighty-five stones were identified on 42 CT scans from 17 patients. Manual measurements showed an average length of 8 mm (range 1.9-21 mm), average maximal density of 686 HU (126-1492 HU), and average stone volume of 192 mm³ (2.9-2555 mm³). Automated computer measurements did not differ from manual measurements for density (755 HU vs 686 HU, p = 0.18) and volume (183 mm³ vs 192 mm³, p = 0.86. Automated length was slightly longer then manual length (10 mm vs 8 mm, p < 0.003). The mean percent differences between manual and automated metrics were 14.3% for density, 21.0% for volume, and 25.2% for length.

Automated stone measurements can be accomplished quickly and precisely with dedicated software that can assess stones of varying size as well as stones with complex geometry. This software eliminates interobserver variability and offers a comprehensive stone profile with which to make clinical decisions.