Purpose: We assessed unenhanced helical computerized tomography (CT) secondary findings as predictors of renal obstruction as determined by diuretic scintirenography, and determined their reproducibility.
Materials and methods: We performed a retrospective review of the records of 77 consecutive patients with unenhanced helical CT findings (stones and secondary findings, including renal parenchymal edema, hydronephrosis, hydroureter, perinephric fat stranding, periureteral fat stranding and extravasation) of urinary lithiasis who had also undergone concomitant diuretic scintirenography during the initial emergency room evaluation during a 1-year period. Unenhanced helical CT films were independently reviewed by 2 attending radiologists (blinded to clinical outcome) to determine interobserver variability. The results were compared to those of diuretic scintirenography.
Results: Considerable interobserver variability, as evidenced by kappa values ranging from 0.26 to 0.60, existed for the diagnosis of secondary findings associated with urinary lithiasis on unenhanced helical CT. There was no significant difference in terms of CT findings between patients diagnosed by diuretic scintirenography as having high grade/complete obstruction and those with partial obstruction (p values 0.24 to 0.85 for the 6 unenhanced helical CT findings analyzed). Analyses of variance followed by Tukey's pairwise comparisons showed no significant difference in average number of unenhanced helical CT findings between patients with high grade/complete obstruction (mean plus or minus standard deviation 4.4 +/- 1.31), partial obstruction (4.4 +/- 1.30), and decompression/no obstruction (4.2 +/- 1.16). However, the mean number of unenhanced helical CT findings for patients with normal scintirenography/no obstruction (1.9 +/- 1.41) was significantly different from each of the other 3 diuretic scintirenography groups. Separate logistic regression analyses showed that each unenhanced helical CT finding, except for renal parenchymal edema and urinary extravasation, was a significant predictor of "any degree of obstruction" (high grade and partial obstruction groups) compared to "no obstruction" (decompressed and no obstruction groups). Odds ratios (95% confidence interval) ranged from 6.15 (2.25, 16.82) for perinephric fat stranding to 3.41 (1.30, 8.97) for hydroureter. When these analyses were repeated after exclusion of 8 patients with bladder/passed stones, only perinephric fat stranding and periureteral fat stranding remained significant predictors of "any degree of obstruction," with respective odds ratios of 4.21 (1.49, 11.91) and 4.08 (1.31, 12.65).
Conclusions: Measures of agreement between trained, independent radiologists with respect to unenhanced helical CT secondary findings show considerable variability. The average number of CT consensus findings is not helpful in differentiating patients with variable degrees of obstruction, except for those with normal scintirenography/no obstruction. Unenhanced helical CT findings, except for renal parenchymal edema and urinary extravasation, are significant predictors of "any degree of obstruction" (high grade or partial obstruction) compared to "no obstruction" (decompressed or no obstruction). Therefore, unenhanced helical CT findings may be useful for identifying patients with any degree of obstruction but do not differentiate between those with high grade and partial obstruction.