Automated Volumetric Assessment by Noncontrast Computed Tomography in the Surveillance of Nephrolithiasis

Abstract

To evaluate the use of automated volumetric assessment for stone surveillance and compare the results with manual linear measurement. We retrospectively reviewed patients seen in our stone clinic who had undergone 2 noncontrast computed tomography (NCCT) scans without stone intervention during the interval between scans. Thirty patients met our inclusion criteria and underwent longitudinal assessment for urolithiasis via NCCT (mean interval 583.2 days, range 122-2030). Fifty-two discrete calculi were analyzed. Three board certified radiologists measured maximal linear stone size in the axial plane using electronic calipers on soft tissue (ST) and bone windows (BWs). Automated stone volume was also obtained by each reader using a dedicated prototype software tool for stone evaluation. Mean stone linear size and volume was 4.9 ± 2.8 mm (ST), 4.5 ± 2.6 mm (BW), and 116.2 ± 194.6 mm(3) (window independent), respectively. Mean interobserver variability for linear size measurement was 16.4 ± 10.5% (ST) and 20.3 ± 13.8% (BW). Interobserver variability for volumetric measurement was 0%. Of the 52 persistent stones, the mean percent change in linear stone size between CT studies was 39.3 ± 46.7% (ST) and 42.9 ± 53.1% (BW) growth, compared with 171.4 ± 320.1% (window independent) growth for automated volume measurement over a mean of 583.2 days. However, discordant results for increased vs decreased interval size was seen between linear and volumetric assessment in 19/52 stones (36.5%). Automated volumetric measurement of renal calculi via NCCT is independent of specific reader and window settings. Volumetric assessment amplifies smaller linear changes over time, whereas as much as one third of cases show linear-volume measurement discordance. Volumetric assessment is therefore preferable, particularly for longitudinal surveillance of renal calculi.