Comparative analysis of ERCP, IDUS, EUS and CT in predicting malignant bile duct strictures.

ABSTRACT

AIM: To compare endoscopic retrograde cholangio-pancreatography (ERCP), intraductal ultrasound (IDUS), endosonography (EUS), endoscopic transpapillary forceps biopsies (ETP) and computed tomography (CT) with respect to diagnosing malignant bile duct strictures. METHODS: A patient cohort with bile duct strictures of unknown etiology was examined by ERCP and IDUS, ETP, EUS, and CT. The sensitivity, specificity, and accuracy rates of the diagnostic procedures were calculated based on the definite diagnoses proved by histopathology or long-term follow-up in those patients who did not undergo surgery. For each of the diagnostic measures, the sensitivity, specificity, and accuracy rates were calculated. In all cases, the gold standard was the histopathologic staging of specimens or long-term follow-up of at least 12 mo. A comparison of the accuracy rates between the localization of strictures was performed by using the Mann-Whitney U-test and the χ(2) test as appropriate. A comparison of the accuracy rates between the diagnostic procedures was performed by using the McNemar's test. Differences were considered statistically significant if P < 0.05. RESULTS: A total of 234 patients (127 males, 107 females, median age 64, range 20-90 years) with indeterminate bile duct strictures were included. A total of 161 patients underwent operative exploration; thus, a surgical histopathological correlation was available for those patients. A total of 113 patients had malignant disease proven by surgery; in 48 patients, benign disease was surgically found. In these patients, the decision for surgical exploration was made due to the suspicion of malignant disease in multimodal diagnostics (ERCP, CT, or EUS). Fifty patients had a benign diagnosis and were followed by a surveillance protocol with a follow-up of at least 12 mo; the median follow-up was 34 mo. Twenty-three patients had extended malignant disease, and thus were considered palliative. A comparison of the different diagnostic tools for detecting bile duct malignancy resulted in accuracy rates of 91% (ERCP/IDUS), 59% (ETP), 92% (IDUS + ETP), 74% (EUS), and 73% (CT), respectively. In the subgroup analysis, the accuracy rates (%, ERCP + IDUS/ETP/IDUS + ETP; EUS; CT) for each tumor entity were as follows cholangiocellular carcinoma 92%/74%/92%/70%/79%; pancreatic carcinoma 90%/68%/90%/81%/76%; and ampullary carcinoma 88%/90%/90%/76%/76%. The detection rate of malignancy by ERCP/IDUS was superior to ETP (91% vs 59%, P < 0.0001), EUS (91% vs 74%, P < 0.0001) and CT (91% vs 73%, P < 0.0001); EUS was comparable to CT (74% vs 73%, P = 0.649). When analyzing accuracy rates with regard to localization of the bile duct stenosis, the accuracy rate of EUS for proximal vs distal stenosis was significantly higher for distal stenosis (79% vs 57%, P < 0.0001). CONCLUSION: ERCP/IDUS is superior to EUS and CT in providing accurate diagnoses of bile duct strictures of uncertain etiology. Multimodal diagnostics is recommended.