Quantification of asymptomatic kidney stone burden by computed tomography for predicting future symptomatic stone events.

ABSTRACT

OBJECTIVE: To find the optimal characterization of asymptomatic radiographic stone burden on computed tomography (CT) scans. METHODS: A survey was sent to stone formers who underwent a CT scan while asymptomatic during a stone clinic evaluation. Symptomatic stone passage events after CT scan were detected by survey and medical record review. Radiographic stone burden was quantified by number of stones, largest stone diameter, automated total stone volume (TSV), and bilateral stones and then compared as predictors of stone events. RESULTS: There were 550 stone formers; 43% had a stone event for a median of 4.7 years after the CT scan. Stone burden by quartiles was 0-1, 2-3, 4-6, and ≥7 for number of stones; 0-2, 3-4, 5-7, and ≥8 mm for largest stone diameter; and 0-8, 9-78, 79-280, and ≥281 mm(3) for TSV; 48% had bilateral stones. The hazard ratios (HRs) for symptomatic event was 1.30 (P <.001) for the number of stones per quartile, 1.26 (P <.001) for largest stone diameter per quartile, 1.38 (P <.001) for TSV per quartile, and 1.80 (P <.001) for bilateral stones. On multivariate analysis, only TSV was an independent predictor of symptomatic events (HR, 1.35 per quartile; P = .01). This risk of events with TSV was also independent of demographics, urine chemistries, and stone composition. Among the 53 patients with interim events between CT scans, a rapid increase in TSV between CT scans (>570 mm(3) per year) predicted subsequent events (HR, 2.8; P = .05). CONCLUSION: Automated TSV is more predictive of symptomatic events than manual methods for quantifying stone burden on CT scan.