BackgroundDespite the sophisticated cross sectional image techniques currently available, a number of biliary stenosis or obstructions remain of an uncertain nature. In these pathological conditions, an "intrinsic" parietal alteration is the cause of biliary obstruction and it is very difficult to differentiate benign from malignant lesions using cross-sectional imaging procedures alone. We evaluated the efficacy of different endoluminal techniques to achieve a definitive pathological diagnosis in these situations.MethodsEighty patients underwent brushing, and or biopsy of the biliary tree through an existing transhepatic biliary drainage route. A subcoort of 12 patients needed balloon-dilatation of the bile duct and the material covering the balloon surface was also sent for pathological examination (balloon surface sampling). Pathological results were compared with surgical findings or with long-term clinical and instrumental follow-ups. Success rates, sensitivity, specificity, accuracy, confidential intervals, positive predictive value and negative predictive value of the three percutaneous techniques in differentiating benign from malignant disease were assessed.The agreement coefficient of biopsy and brushing with final diagnosis was calculated using the Cohen's "K" value.ResultsFifty-six patients had malignant strictures confirmed by surgery, histology, and by clinical follow-ups. Success rates of brushing, balloon surface sampling, and biopsy were 90.7, 100, and 100%, respectively. The comparative efficacy of brushing, balloon-surface sampling, and biopsy resulted as follows: sensitivity of 47.8, 87.5, and 92.1%, respectively; specificity of 100% for all the techniques; accuracy of 69.2, 91.7 and 93.6%, Positive Predictive Value of 100% for all the procedures and Negative Predictive Value of 55, 80, and 75%, respectively.ConclusionsPercutaneous endoluminal biopsy is more accurate and sensitive than percutaneous bile duct brushing in the detection of malignant diseases (p < 0.01).