Diagnosis of CCA can be challenging. A dominant stricture in a patient with PSC is a stenosis with a diameter of ≤1.5 mm in the common bile duct or ≤1 mm in the hepatic ducts [9]. It is often difficult to distinguish a benign dominant stricture from PSC from a malignant stricture; thus, one should have a high index of suspicion for CCA when a patient develops evidence of biliary obstruction (jaundice, cholestasis, pruritus, cholangitis), unexplained weight loss, or abdominal pain. A multidisciplinary approach is often needed to diagnose CCA including laboratory studies, cross-sectional imaging, cholangioscopy, and pathology.

A variety of imaging modalities are used in the diagnosis of CCA including ultrasound (US), computerized tomography (CT), and magnetic resonance imaging (MRI) with concurrent magnetic resonance cholangiopancreatography (MRCP) (see Chap. 13). The positive predictive value is nearly 100 % if a characteristic lesion is found on US, CT, or MRI (Table 2.1). Characteristic lesions, however, are not commonly seen, especially in early-stage CCA. The overall positive predictive value for US, CT, and MRI are 48 %, 38 %, and 40 %, respectively [19].

The most commonly used laboratory test besides routine liver enzymes to detect CCA is CA 19-9. CA 19-9 is an antibody that binds to the tumor surface marker Sialyl-Lewis A. CA 19-9 is found to be elevated (normal typically up to 35 U/ml) in multiple other diseases and bile duct conditions including ascending cholangitis, hepatocellular carcinoma, alcoholic liver disease, primary biliary cirrhosis, chronic viral hepatitis, autoimmune hepatitis, and pancreatitis. Levy et al. found that in PSC, a CA 19-9 of ≥129 U/mL had a sensitivity of 79 %, a specificity of 98 %, and a positive predictive value of 79 % for CCA [40]. A change in CA 19-9 of ≥63.2 U/mL had a sensitivity of 90 %, specificity of 98 %, and a positive predictive value of 42 %.

Endoscopic retrograde cholangiopancreatography (ERCP) is often used in patients with PSC to further investigate and characterize biliary strictures and to manage biliary obstruction with balloon dilation and stenting. Tissue sampling of dominant strictures is often achieved through bile duct brushings for cytology. Routine biliary cytology alone has been found to be highly specific (95–100 %) but to have lower sensitivity (36–83 %) [42]. The broad range in sensitivity cited in literature is due to the definition of a positive cytology results. Studies that defined a positive finding as both high-grade and low-grade dysplasia had a higher sensitivity than those that only defined high-grade dysplasia as a positive result.

Fluorescence in situ hybridization (FISH) can be used in addition to cytology to increase sensitivity for malignancy. Fluorescence in situ hybridization uses fluorescently labeled DNA probes to detect chromosomal aneuploidy (losses or gains of chromosomes). Abnormalities are characterized as trisomy, tetrasomy, and polysomy of chromosomes 3 and/or 7. Trisomy refers to ≥10 cells with three copies of chromosome 3 and 7, tetrasomy refers to ≥10 cells with four copies of all probes, and polysomy refers to ≥5 cells with ≥3 signals in two or more of the four probes [3]. Trisomy and tetrasomy of chromosomes 3 and 7 have low specificity for PSC as these findings are frequently found in biliary tree inflammation without malignancy. In contrast, polysomy has a specificity of 88 % for CCA [3]. It is difficult to interpret positive FISH polysomy in the setting of negative cytology. Patients with positive polysomy on serial brushings are significantly more likely to be diagnosed with cholangiocarcinoma than those with subsequent nonpolysomy results [4]. The presence of both polysomy and CA 19-9 ≥ 129 U/mL was a significant predictor for developing CCA (hazard ratio of 20.4 (95 % CI 7.94–52.63)) for polysomy and CA 19-9 ≥ 129 U/mL versus nonpolysomy and CA 19-9 < 129 U/mL [4]. If a patient with PSC is found to have negative cytology and polysomy, they should be followed up closely with repeat ERCP and biliary brushings for cytology and FISH especially if there is a non-resolving dominant stricture and/or elevated CA 19-9. Compared with other prognostic features, multifocal (multiple areas of the biliary tree) polysomy carries the highest risk for cholangiocarcinoma compared to unifocal polysomy HR 82.4 (95 % CI 24.5–277.3) vs. 13.27 (95 % CI 3.32–53.1), respectively, on univariate analysis [24]. Multifocality remains a stronger predictor of CCA even when adjusting for CA 19-9, cytology, and prior abnormal FISH. Patients with unifocal polysomy with suspicious cytology remain at increased risk. If serial polysomy is detected in a malignant appearing stricture, even in the setting of negative cytology, liver transplantation should be considered. Figure 2.2 summarizes the approach to managing a dominant stricture in patients with PSC.

Cholangioscopy allows for direct visualization of the biliary tree and theoretically improves sampling as it allows for directed bile duct biopsies. Visual characteristics suspicious for malignancy are exophytic lesions, ulcerations, papillary mucosal projections, dilated tortuous vessels, and raised lesions [20, 60]. A meta-analysis showed that cholangioscopy with targeted biopsies of dominant strictures was able to detect CCA with a sensitivity and specificity of 66.2 % and 97 %, respectively [37].

Endoscopic ultrasound (EUS) with fine needle aspiration (FNA) of a biliary stricture has also been used for additional tissue sampling in the setting of indeterminate biliary brushings and FISH. However, this method carries a risk of tract seeding and peritoneal metastasis and should be avoided, especially in patients potentially eligible for liver transplantation. In one study, 83 % of individuals who underwent a transperitoneal or transluminal biopsy of biliary strictures had peritoneal metastasis compared to 8 % peritoneal metastasis in those who did not undergo biopsy [32]. EUS with FNA may be useful to sample lymph nodes to evaluate for metastatic disease in those being considered for liver transplantation and is often done prior to exploratory laparotomy.

Surgical resection is an option for localized lesions with otherwise normal hepatic parenchyma. Contraindications to surgical resection of hilar CCA include bilateral tumor extension involving the left and right secondary biliary radicles, unilobar involvement with encasement of contralateral portal vein or hepatic artery, bilateral vascular involvement, distant metastases, underlying liver disease (advanced fibrosis or cirrhosis), future liver remnant <25–30 % with no or poor response to portal vein occlusion, and severe comorbidities [33, 55]. Due to the diffuse nature of PSC and risk for advanced hepatic fibrosis, PSC patients with CCA are often not candidates for resection.

Most patients with PSC and the diagnosis of hilar CCA will need to be considered for liver transplantation (LT) as means for a definitive cure. Liver transplantation is not generally considered a treatment for intrahepatic or distal bile duct tumors. The management of the latter is a Whipple procedure which in a patient with severe end-stage liver disease may require concurrent liver transplantation. Historically, LT for CCA has been associated with very poor outcomes. In 2000, The Mayo Clinic developed a protocol for both patient selection and treatment of patients with CCA undergoing LT [23]. Patients fulfilling the so-called Mayo criteria showed superior outcomes with LT compared to historical controls. One study found a median survival of 3.3 years after LT prior to the publication of the Mayo results in May 2000 compared to a median survival of 7.8 years for LTs done after May 2000 [58].

The Mayo protocol employs neoadjuvant therapy followed by LT as a definitive therapy for patients with hilar CCA. The criteria include patients with biliary duct obstruction and cytologically proven CCA or a mass lesion seen on cross-sectional imaging with biliary obstruction (Table 2.2). The protocol utilizes external and intraductal radiation therapy followed by chemotherapy (capecitabine) until the patient undergoes LT. All patients undergo exploratory surgery prior to LT to exclude extrahepatic disease, either after completing radiation or just prior to transplant. Using this protocol, Rea et al. found that LT with neoadjuvant chemoradiation had significantly improved 5-year survival when compared to conventional resection (82 % vs. 21 %) and had fewer recurrences (12 % versus 27 %) [56]. Overall survival of patients with PSC is approximately 70 % at 5 years. This approach has been externally validated at centers outside Mayo having nearly identical outcomes (65 % 5-year survival) [21]. Currently the United Network for Organ Sharing allows model for end-stage liver disease (MELD) exception points for patients meeting the criteria outlined in the Mayo protocol.

Contributing to the excellent outcomes of this protocol are the strict selection criteria. Predictors of pre-LT dropout include CA 19-9 ≥ 500 U/mL, mass lesion ≥3 cm, malignant brushing or biopsy, and biological lab MELD score ≥20. Predictors of post-LT recurrence include elevated CA 19-9, portal vein encasement, and residual tumor on explant [22]. Finally, it is important to note that this protocol does not require the diagnosis of CCA but includes the presence of polysomy alone or elevation in CA 19-9 > 100 with a concurrent malignant appearing dominant stricture. It is possible that excellent outcomes with this protocol are further explained by the fact that patients simply did not have cancer. This is supported by the external validation of this protocol at 12 large volume transplant centers which found that patients without residual CCA on explant did better and had a significantly lower chance of recurrence than those with residual tumor tissue on explant [22]. It is impossible to determine whether these individuals never had CCA to begin with or that their CCA was effectively treated with neoadjuvant chemoradiation.