Radiologic evaluation of urolithiasis.

Interobserver variations between three radiologists with different backgrounds (duration of practice, and departments) were studied in connection with the diagnosing of urolithiasis. Films from 152 consecutive patients (examined with conventional abdominal films, urography/tomography) were analyzed independently. With respect to the number of patients with or without calculi there was agreement between the three observers in 129 of the 152 cases, which was taken to represent very good agreement beyond chance. The total number of stones was interpreted as 832, 460 and 570 respectively, and the median stone size as 6, 9 and 11 mm2. There was a fair interobserver agreement rate regarding whether stones were localized to the renal parenchyma or calyces, and whether stenosis/stricture or hydronephrosis was present. Tomography appeared to reduce the interobserver variability significantly. Adequate bowel cleaning, oblique projections, and lack of clinical information apparently did not influence the agreement rate. It was concluded that such interobserver variations are factors of importance in the reliability of the radiologic evaluation of urolithiasis.