Interobserver variability in the interpretation of unenhanced helical CT for the diagnosis of ureteral stone disease

Purpose: The purpose of this study was to analyze interobserver agreement in the interpretation of unenhanced helical CT (UHCT) for the evaluation of ureteral stone disease and obstruction. Method: One hundred three UHCT examinations were independently and retrospectively reviewed by five readers including attending radiologists, a radiology resident, and an attending urologist. Examinations were interpreted as positive, negative, or indeterminate for ureteral stone disease and obstruction. The Cohen κ test was used to measure interobserver agreement. The accuracy of the readers was also assessed. Results: The κ value ranged from 0.67 to 0.71 among the three attending radiologists and from 0.65 to 0.67 among the radiology attending physicians and radiology resident. Although the urologist tended to agree less well with the other readers (κ range: 0.33-0.46), there was no statistically significant difference (p < 0.05) in the accuracy among all five readers. The percentage of cases interpreted as indeterminate ranged from 8 to 25% and almost invariably involved difficulty distinguishing phleboliths from minimally obstructing distal ureteral calculi. The percentage of UHCT scans correctly interpreted as positive and correctly interpreted as negative ranged from 73% (n = 27) to 86% (n = 32) and 63% (n = 22) to 86% (n = 30), respectively. Conclusion: Interobserver agreement was very good among the radiology attending physicians and resident and moderate with the urologist. The examination is an accurate technique in the evaluation of ureteral stone disease, although limitations exist, particularly in the diagnosis of minimally obstructing distal ureteral calculi.