By Frank G. Gress, MD
Cholangiocarcinoma and other malignant bile-duct tumors are often difficult to diagnose because they commonly present late and are therefore associated with a poor prognosis. The majority of these tumors present as biliary strictures. Up to 20 percent of proximal biliary strictures, however, can be due to non-malignant causes.
While noninvasive imaging modalities (i.e. transabdominal ultrasound, CT, MRCP and PET) can provide important clues regarding the differential diagnosis and even staging of these tumors, a definitive diagnosis requires tissue sampling.
Endoscopic techniques can be very useful for distinguishing benign from malignant biliary strictures using tissue acquisition combined with modalities such as ERCP, endoscopic ultrasound-guided fine needle aspiration (EUS-FNA), intraductal ultrasound (IDUS) and cholangioscopy.
The only treatment approach that offers long-term survival for cholangiocarcinoma is surgical resection. Although, surgical results are poor, even with radical resection techniques the five-year survival ranges between 20-40 percent with surgical mortality of 8-10 percent and a surgery complication rate of 64 percent.
Unfortunately, the majority of patients with cholangiocarcinoma will not be surgical candidates due to advanced disease at the time of presentation and diagnosis. These patients are usually referred for chemotherapy and/or radiotherapy after a positive tissue diagnosis is obtained. The dilemmas faced by clinicians are twofold; how to best select those patients that are most likely to benefit from surgery, and for those patients not deemed surgical candidates based on imaging, what is the best approach for obtaining a histological diagnosis?
ERCP is used for diagnostic evaluation as well as palliation in patients with a known stricture or suspected pancreatic or biliary malignancies. ERCP sampling is commonly performed with brush cytology, which has a sensitivity ranging between 33-57 percent for the diagnosis of malignant biliary strictures and has also been reported to have a higher yield for diagnosing cholangiocarcinoma (sensitivity of 63-80 percent).
The sensitivity of ERCP sampling can be improved with the use of endobiliary forceps or endoscopic fine needle aspiration techniques. In addition, the combination of all three techniques yields an improved sensitivity ranging between 62-82 percent. Despite the use of advanced sampling techniques, however, the negative predictive value remains low (39 percent).
ERCP sampling can also be challenging due to the significant time and expertise required to master these techniques and the risk of losing access during multiple sampling. In addition, when using a forceps biopsy or fine needle biopsy, a biliary sphincterotomy is initially performed to improve access. This, however, can increase the risk of complications.
When ERCP sampling is unable to make a definitive diagnosis, alternative sampling techniques should be considered. Studies have shown that EUS-FNA has a sensitivity of 85 percent and a specificity of 100 percent for the diagnosis of pancreatic adenocarcinoma.
Previous studies have examined the role of EUS-FNA in the evaluation of biliary strictures, including proximal biliary strictures. From these studies, the diagnostic sensitivity of EUS-FNA has been reported to be between 43-86 percent for all biliary strictures and 25-83 percent for proximal biliary strictures.
A study by Lee, et al reported that EUS features of biliary strictures can accurately differentiate benign from malignant strictures.
In the study, the positive predictive value of finding a pancreatic mass or irregular outer border of the bile duct was 100 percent with a negative predictive value of 84 percent. Furthermore, malignancy was strongly associated with a bile duct wall thickness ≥3 mm.
EUS and EUS-guided FNA provide several advantages in the evaluation of biliary strictures.
EUS can evaluate the pancreas for the presence of a mass or changes consistent with chronic pancreatitis, both of which can cause biliary strictures. EUS-FNA can also be performed during the same procedure in order to provide a definitive diagnosis. EUS can also identify suspicious regional lymphadenopathy which can be sampled with FNA, thereby providing additional staging information. In the evaluation of biliary strictures, the presence of biliary stents may lead to thickening and asymmetry of the bile-duct wall.
These stent-related changes should be recognized at the time of EUS, but probably do not significantly impact the role of EUS-FNA where cytology is used to establish a definitive diagnosis.
The role of intraductal ultrasound (IDUS) for the evaluation of indeterminate biliary strictures continues to evolve. IDUS has been used for both the diagnosis and staging of biliary tumors. Increased resolution due to the development of higher frequency probes also gives IDUS some advantages over transduodenal EUS for the evaluation of biliary strictures.
Several studies have reported sonographic features suggestive of malignancy including: eccentric wall thickening with an irregular surface, disruption of the normal 3-layer sonographic pattern, a hypoechoic mass with irregular margins, heterogeneous echo-poor areas invading surrounding tissue, continuation of the main hypoechoic mass into adjacent structures, presence of lymph nodes, and evidence of vascular invasion.
The accuracy rate of IDUS has been reported as 76 percent for the diagnosis of cholangiocarcinoma.26 Furthermore, compared to ERCP with tissue sampling, IDUS has been shown to significantly increase the diagnostic accuracy (90 percent vs. 67 percent, p=0.04) in patients with indeterminate bile duct strictures. IDUS also improved diagnostic accuracy (89 percent vs. 76 percent, p<0.002) compared to EUS without FNA in 56 patients who underwent surgical resection.
Cholangioscopy has also been reported to be useful in the evaluation of indeterminate biliary strictures. Peroral cholangioscopy (POCS) utilizing a thin, flexible, catheter-type endoscope allows for direct visualization and better sampling of bile ductal pathology. When compared to ERCP brush cytology, in a recent single-center Japanese study, POCS improved the diagnostic accuracy for biliary strictures from 78 percent to 93 percent.
As we see improvements in the design, optical resolution, and maneuverability of these miniaturized endoscopes, cholangioscopy can be expected to assume an increasingly important role in the diagnosis of indeterminate biliary strictures.
DeWitt et al, reported that EUS-FNA is a sensitive method for the diagnosis of proximal biliary strictures following negative or unsuccessful ERCP brush cytology. In the second-largest series to examine patients with proximal biliary strictures, the sensitivity, specificity, positive predictive value, negative predictive value and accuracy of EUS-FNA were 77 percent (95 percent CI, 54-92 percent), 100 percent (CI, 15-100 percent), 100 percent (CI, 83-100 percent), 29 percent (CI, 4-71 percent) and 79 percent (CI, 58-93 percent), respectively.
Although the sensitivity is slightly less than that reported in previous studies, it still compares favorably with the sensitivity of ERCP sampling techniques.
Furthermore, the study reported a negative predictive value (NPV) of 29 percent for EUS-FNA of proximal billiary strictures. This is consistent with the NPV of 40-60 percent found by other authors and argues that malignancy cannot be reliably excluded by a negative result.
In addition, the study reports that EUS visualized a mass in 13 patients with previously negative imaging and correctly predicted resectability in 4/7 (57 percent) patients who underwent surgery. No complications were observed following EUS-FNA for indeterminate biliary strictures, which is in keeping with previous reports. Importantly, the findings of this study further support a role for EUS and EUS-FNA in the evaluation of biliary strictures, especially for those patients with negative imaging or non-diagnostic sampling with ERCP.
In summary, the evaluation of biliary strictures can be challenging and presents the clinician with several choices. Non-invasive imaging modalities can identify lesions that are suspicious for malignancy, however imaging findings are nonspecific and up to 20 percent of patients who undergo surgery have diagnoses other than cholangiocarcinoma. Improved preoperative diagnosis and staging is essential to identify those patients who are most likely to benefit from surgical resection for malignant biliary obstruction. Furthermore, histological diagnosis is required for those patients who are not surgical candidates prior to commencing chemotherapy and/or radiotherapy. In the presence of jaundice it is our practice to recommend ERCP as the first endoscopic procedure, since the cause of the biliary obstruction can be evaluated, sampling undertaken and a biliary stent inserted to obtain biliary drainage. In the absence of jaundice or when ERCP sampling is non-diagnostic, we perform EUS-FNA for further evaluation and sampling to identify those patients who are more likely to benefit from surgery and to establish a histological diagnosis in those who are not candidates for surgery. In view of the low negative predictive value for EUS-FNA in this setting, further tissue sampling should be pursued when a strong index of clinical suspicion for exists.
Frank Gress, MD, is a professor of medicine in the Division of Gastroenterology and Hepatology at State University of New York (SUNY), College of Medicine and SUNY Downstate Medical Center, in Brooklyn, N.Y.