Biliary stricture measurement and stent selection.

Abstract

Because of variations in magnification, errors in stent selection may occur when stricture location is determined from fluoroscopic images or x-ray film. An ERCP catheter with measurement markings was inserted 5, 7, and 10 cm into the bile duct in 30 patients without bile obstruction. Film measurements obtained at each depth were converted to actual distances using standard conversion as well as endoscope ratio conversion formulas. The site of obstruction in 52 patients with an obstructing lesion was measured with a ruled catheter, by a wire withdrawal technique, and using x-ray film. After a stent was selected on the basis of the catheter measurement, the accuracy of stent selection was determined for each method of measure. Conversion of x-ray measurements to ruled catheter measurements obtained by insertions of 5, 7, and 10 cm using standard conversion factors yielded measurements of 4.9 +/- 0.9, 6.8 +/- 0.3, and 9.5 +/- 1.9 cm, respectively, with a correlation coefficient of r = 0.80. Ratio conversion yielded measurements of 0.5 +/- 0.8, 7 +/- 1.0, and 9.9 +/- 1.4, respectively, with a correlation coefficient of r = 0.88. Measurement of stricture location with the ruled catheter and then by wire withdrawal yielded a correlation coefficient of 0.98. When ruled catheter measurements were compared with the x-ray ratio conversions, the concordance dropped to 0.79. The ruled catheter and wire withdrawal were more accurate in predicting the location of the stent tip than x-ray film measurements (p < 0.001, Wilcoxon matched pairs). Of 52 stents selected, no errors in stent selection occurred when the ruled catheter was used (p < 0.001, Fisher's exact test), two errors occurred when wire withdrawal was used (p < 0.004), and 14 errors occurred when film measurements were used. Use of a ruled catheter or wire withdrawal is much more accurate for selecting stents than use of x-ray film measurements.