Cholangiocarcinoma (CCA) arises from intra- and extrahepatic bile duct epithelia, has an incidence of around 2/100,000 and a poor prognosis with a mean survival of 24 months. While CCA is strongly associated with primary sclerosing cholangitis (PSC) and parasitic infections of the biliary tree, risk factors for most CCA remain unclear. CCAs have traditionally been classified according to their anatomic location. They are heterogeneous tumors with differing biological behavior and prognosis that require differentiated therapeutic approaches depending on their location.
Perihilar CCA (pCCA), which accounts for over 50% of all CCAs is located above the cystic duct up into the second order biliary ducts (Fig. 1). This location in the liver hilus with frequent involvement of both main bile ducts and vascular structures make pCCA a particularly dismal tumor.
Diagnosis of pCCA is often difficult. Painless jaundice is the clinical presentation in about 90% of patients, often accompanied by unspecific symptoms such as weight loss, abdominal pain, or nausea. Risk factors, especially PSC and liver flukes, should be assessed. Benign, PSC-related strictures and immunoglobulin subclass 4 (IgG4) related cholangiopathy are important differential diagnoses of pCCA.
Magnetic resonance imaging (MRI) with magnetic resonance cholanigo-pancreatography (MRCP) (Fig. 2A-C) is the imaging technique of choice for diagnosis and staging of pCCA allowing for a 90% sensitivity and 76% specificity in detecting pCCA. Thickening of the bile duct wall, mass lesions, involvement of vascular structures (encasement or invasion of portal vein or hepatic artery), lymph node enlargement and distant metastasis are crucial diagnostic parameters. Hepatic MRI volumetry is valuable prior to surgical resection.
High-quality, contrast-enhanced multidetector computed tomography (CT) shows good sensitivity and specificity in pCCA. Recent three-dimensional CT cholangiography techniques allow for accurate assessment of tumor extent, especially in patients with obstructive jaundice. Positron emission tomography (PET)–CT does not generally improve diagnosis of pCCA.
Sensitivity and specificity of transabdominal ultrasonographic imaging is dependent on lesion size and operator experience with a sensitivity of up to 60% and a specificity of 95%.
Endoscopic retrograde cholangiography (ERC) is a diagnostic cornerstone for pCCA and can be a therapeutic intervention. ERC allows for focused biliary imaging (Fig. 2D) and diagnostic sampling by biopsy or brush cytology. Dilatation and stenting of strictures can be employed to drain cholestatic liver segments. The sensitivity and specificity of ERC are approximately 75% and 70%, respectively. In addition to ERC, endoscopic ultrasound (EUS) with fine needle aspiration (FNA) allows for the assessment of lymph node involvement and cytological confirmation of suspected metastases, thus avoiding aggressive curative treatment protocols in ineligible patients with extrahepatic tumor manifestations. FNA of the primary bile duct tumor bears the risk of tumor seeding with adverse consequences for prognosis and therapeutic options and should not be performed.
Cytological sampling for suspected pCCA is difficult. ERC brush cytology has a low sensitivity (∼20%) with frequent sampling errors due to the desmoplastic nature of the tumor with limited yield of tumor cells. Combination of cytology with fluorescence in situ hybridization (FISH) for chromosomal aberrations increases the sensitivity of brush cytology to 40%-90%. Cytology suspicious for cancer or a dominant biliary stricture in combination with polysomy in FISH are considered diagnostic for cancer.
Elevation of carbohydrate antigen 19-9 (CA 19-9) can result from nonmalignant biliary diseases such as PSC and should be interpreted in conjunction with the other diagnostic tests. Sensitivity and specificity of CA 19-9 (>100 U/mL) to distinguish benign from malignant biliary strictures are reportedly 76% and 92%, respectively. Concomitant elevation of carcinoembryonic antigen supports the diagnosis of CCA. Of note, Lewis antigen–negative patients do not produce CA 19-9. The search for more sensitive and specific CCA biomarkers is ongoing, and analysis of bile for tumor antigens such as carcinoembryonic cell adhesion molecule 6 might be a promising new approach.
Diagnosis of pCCA thus can be made in the following scenarios: 1) cytologic brushings positive or suspicious for adenocarcinoma from a stricture; 2) cytologic brushings positive or suspicious for adenocarcinoma from a stricture in conjunction with a positive FISH for polysomy; and 3) perihilar mass lesion on MRI with CA 19-9 >100.
Although lacking data from randomized controlled clinical trials, gemcitabine as single agent therapy was considered a provisional standard in CCA for many years given its efficacy in advanced pancreatic cancer. Recently, a large scale randomized trial showed that the combination of gemcitabine with cisplatin consistently prolonged overall survival and is an appropriate first-line therapy in patients who are ineligible for curative treatment. Currently, there is no established second-line treatment available, and patients should be encouraged to participate in clinical trials. The impact of adjuvant chemotherapy after tumor resection is also unclear due to the lack of available clinical data. According to the National Comprehensive Cancer Network guidelines, the use of adjuvant chemotherapy or chemoradiation may be considered according to the resection and nodal status if participation in a clinical trial is not possible. Neoadjuvant strategies outside the liver transplantation protocol described below are not established. Few small studies show promising results for neoadjuvant chemoradiotherapy to achieve R0 resectability, signifying a curative complete resection of the tumor. These data warrant further randomized trials.
Adjuvant radiotherapy is controversial, with only few prospective studies available and potentially severe side effects. The data seem to support adjuvant radiation in patients with R1 resection, which indicates the microscopic presence of residual tumor after the resection. Palliative radiotherapy can improve quality of life and overall survival and novel fractioned stereotactic radiation therapy techniques can result in noticeable prolongation of survival.
Surgical resection is the only curative therapeutic approach for pCCA. Unfortunately, only 30% of CCA are resectable at diagnosis, with a 5-year survival of around 45% after curative resection. pCCA is the most challenging form surgically due to its longitudinal spreading along the bile ducts and infiltration of surrounding structures. Improved and aggressive surgical techniques allow for curative treatment of tumors that were previously considered unresectable.
The surgical approach to pCCA usually involves an extended hepatectomy, routine removal of segment 1 shown to reduce local recurrence, bile duct resection, and lymphadenenctomy. Resection for pCCA in PSC has been found to be vain, mainly due to advanced tumor stage at diagnosis and underlying liver disease.
To avoid postoperative liver insufficiency and a “small-for-size syndrome” after extensive liver resection, optimal preparation of the remaining liver is mandatory. This includes preoperative biliary drainage and normalization of liver function. Preoperative portal vein embolization of the diseased liver is frequently required prior to resection to induce hypertrophy of remaining segments and increase the volume of the future remaining liver.[23, 24] Frequent tumor infiltration of neighboring vascular structures has triggered highly challenging approaches that combine extended liver resection with resection and reconstruction of portal vein and hepatic artery—in the case of right-sided tumors, ideally performed using a non-touch technique with en bloc resection of the portal vein.25
The protocol of neoadjuvant radiochemotherapy with subsequent liver transplantation in a strictly selected patient population developed at the Mayo Clinic overcame discouraging historic results for orthotopic liver transplantation in pCCA.26 Diagnostics consist of a chest and abdomen CT, MRI/MRCP, abdominal ultrasound, ERC-pancreatography (ERCP), and endoscopic ultrasound with FNA of all celiac and perihilar lymph nodes. Importantly, transperitoneal needle biopsy of the primary tumor is an exclusion criterion for this protocol. In this protocol, patients with unresectable pCCA measuring <3 cm in radial diameter and no metastasis receive aggressive external beam radiation, brachytherapy (by percutaneous transhepatic cholangiodrainage or ERCP), and chemotherapy (5-fluorouracil or capecitabine) until staging laparotomy. If no extrabiliary tumor is found at laparotomy, the patient may be listed for liver transplantation (Fig. 3). Liver transplantation itself should be performed without hilar dissection, and an arterial interposition graft is regularly needed due to radiation damage of the hepatic artery. In case of involvement of the common bile duct, the procedure is extended to include a pancreato-duodenectomy. Dropout rates of this highly selective protocol reach 30% at 12 months due to tumor progression, but 5-year survival after liver transplantation within these strict criteria is 65%-70%.
Diagnosis and treatment of pCCA is complex. Painless jaundice is the leading clinical presentation. High-quality CT, MRI, and MRCP are the imaging techniques of choice. ERC with biopsy or brush cytology is a key diagnostic technique. Endoscopic ultrasound with FNA of lymph nodes is helpful for disease staging. FISH techniques have improved cytological results. The only curative approach to pCCA is surgical resection or neoadjuvant radiochemotherapy plus liver transplantation in highly selected patients. Chemotherapy with cisplatin and gemcitabine has been shown to improve survival in patients with advanced disease who are not amenable to curative treatment strategies. No adjuvant chemotherapy or radiation treatment has shown clear benefits in overall survival but can be of palliative value.
The author would like to thank Beat Müllhaupt and Panagiotis Samaras for proofreading the manuscript and Thomas Pfammatter and Christoph Gubler for providing MRCP and ERC images.