Long-term outcomes of positive fluorescence in situ hybridization tests in primary sclerosing cholangitis


  • Potential conflict of interest: Dr. Lindor received grants from Axcan and Orphan. Dr. Halling received grants and holds intellectual property rights for Abbott Molecular Inc.


Patients with primary sclerosing cholangitis (PSC) are at increased risk for developing cholangiocarcinoma (CCA). Fluorescence in situ hybridization (FISH) is a cytological test designed to enhance early CCA diagnosis. The long-term outcome of PSC patients with a positive FISH test (polysomy, trisomy/tetrasomy) are unclear. All PSC patients with at least one FISH test were identified and defined to have CCA if they had a positive tissue biopsy, positive cytology, or evidence of cancer in the explant after liver transplantation. A total of 235 PSC patients had at least one FISH test performed, and 56 patients had CCA on histopathology (n = 35) or cytology (n = 21). Overall, 120 of 235 (51%) of PSC patients tested for FISH were positive, but only one third of these positive patients had CCA. Sensitivity and specificity for FISH polysomy were 46% and 88%, and for trisomy/tetrasomy they were 25% and 67%, respectively. Survival analysis showed that patients with FISH polysomy had an outcome similar to patients with CCA; whereas FISH trisomy/tetrasomy patients had an outcome similar to patients with negative FISH tests. The FISH polysomy patients without cancer compared with those with CCA had lower serum bilirubin, lower carbohydrate antigen 19-9 (CA 19-9), lower Mayo risk score, and lower occurrence of dominant strictures. Conclusion: In PSC patients, the presence of a dominant stricture plus FISH polysomy has a specificity of 88% for CCA. Patients with FISH showing trisomy or tetrasomy have a similar outcome to patients with negative FISH. FISH testing should be used selectively in patients with other signs indicating CCA and not as a screening tool in all PSC patients undergoing endoscopic retrograde cholangiopancreatography (ERCP). (HEPATOLOGY 2009.)

Primary sclerosing cholangitis (PSC) carries an increased risk of hepatobiliary malignancy, especially cholangiocarcinoma (CCA), which also is the most lethal complication of PSC.1 There are no specific clinical features that predict the diagnosis of CCA.2–9 The lifetime occurrence of CCA in PSC patients varies from 5% to 36%.2–6 The factors that predispose PSC patients to develop CCA are unknown and unpredictable, which makes it very challenging to diagnose CCA early.

The combination of an annual ultrasound and the tumor marker carbohydrate antigen 19-9 (CA 19-9) was found to be useful in early detection of CCA.9 If there are any suspicious findings, these patients are subjected to endoscopic retrograde cholangiopancreatography (ERCP). During ERCP, brush cytology and biopsy specimens are obtained from dominant strictures to confirm or exclude the diagnosis of CCA.8–10 A CCA may be present but not seen on cross-sectional imaging studies because these cancers often grow longitudinally along the bile duct rather than radially away from the bile duct.4, 5, 7, 8 The distinction between malignant and inflammatory strictures is confounded in PSC because the inflammation associated with PSC complicates cytological assessment, and access to the bile duct is limited for cytological and tissue acquisition.8–10 CCA are frequently desmoplastic, resulting in acellular sampling, further complicating the diagnosis.8–10

Up to 80% of biliary malignancies exhibit chromosomal abnormalities in the form of structure or the number of chromosomes within the cell.8 Fluorescence in situ hybridization (FISH) uses fluorescently labeled DNA probes to detect aneusomy of individual cells (abnormal loss or gain of chromosomes or chromosomal loci).10–15

Previous studies evaluating the utility of FISH tests in patients with indeterminate biliary strictures were all limited by the number of PSC patients and the duration of follow-up.10, 13, 16–19 None of these studies addresses the long-term outcomes of the tests focused on PSC patients. Selection bias from assessing patients referred for liver transplantation or surgery for suspected malignancy may have influenced results in some series10, 13, 16–19 because of the high pretest probability of CCA.

The aim of the current study was to analyze the outcome of FISH testing in PSC patients and to assess the diagnostic utility of this test in this group of patients in detecting CCA. We also wanted to formulate follow-up guidelines for patients with a positive test result based on our findings.


CA 19-9, carbohydrate antigen 19-9; CCA, cholangiocarcinoma; CEP, peri-centromeric regions of chromosomes; ERCP, endoscopic retrograde cholangiopancreatography; FISH, fluorescence in situ hybridization; HGD, high-grade dysplasia; PSC, primary sclerosing cholangitis.

Patients and Methods


Our patient population comprised consecutive patients with well-defined PSC who were identified using a computerized master diagnosis index of the Mayo Clinic, Rochester, MN. The diagnosis of PSC was based on cholangiographic findings of multifocal strictures and beading of the intrahepatic or extrahepatic bile ducts with compatible biochemical abnormalities and exclusion of secondary causes.20, 21 These patients had FISH testing performed for further evaluation of clinical, laboratory, or radiological abnormalities suspicious for CCA in the setting of PSC at the Mayo Clinic, Rochester, MN, between October 2003 and June 2008. Some of these tests were obtained in PSC patients during an ERCP as a part of the diagnostic workup of cholestatic liver disease during the study period and were not done to address a clinical suspicion of cancer.

The study was approved by the Mayo institutional review board, and written informed consent was obtained from all patients for participation in medical research.

Routine Cytology and FISH.

Dedicated gastrointestinal cytopathologists with particular expertise for each diagnostic test independently interpreted the routine cytology and FISH specimens. They had no knowledge of the other test results or patient's clinical history. Cytology specimens were interpreted as either positive for malignancy, suspicious for malignancy, atypical, negative for malignancy, or with inadequate cellularity for interpretation.

FISH uses fluorescently labeled DNA probes to peri-centromeric regions of chromosomes or unique loci to detect cells that have numerical or structural abnormalities indicative of malignancy. The probe set used for FISH (UroVysion; Abbott Molecular, Inc., Des Plaines, IL) targets the peri-centromeric regions of chromosomes 3 (CEP3), 7 (CEP7), and 17 (CEP17), and band 9p21 (P16/CDKN2A gene). Slides were processed and hybridized with the probe set using the manual method as described previously.11, 22, 23 The slides were assessed by scanning for cytologically atypical cells and by determining the number of CEP3, CEP7, CEP17, and 9p21 signals in those cells. To scan for atypical cells by FISH, the cells are assessed for patchy and lighter nuclear 4′-6-diamidino-2-phenylindole staining, nuclear enlargement, and irregular nuclear contour. Three general types of chromosomal abnormalities were observed by FISH in this study: polysomy, tetrasomy, and trisomy of chromosome 7 or 3. A patient's specimen was reported as follows: polysomy if five or more cells showed gains of two or more of the four probes; tetrasomy if 10 or more cells showed four copies of all probes; trisomy if 10 or more cells showed three copies of chromosome 7 (or 3), and two or fewer copies of the other three probes.

Definition of Cancer.

The patients were considered to have cancer if they had a positive tissue biopsy or positive cytology or evidence of cancer in the liver explant.

The patients were considered to have possible CCA if there was evidence of a mass lesion on the imaging study. Patients with high-grade dysplasia suspicious for malignancy were not counted as definite for cancer.


Patients with positive FISH (polysomy, trisomy/tetrasomy) results were actively followed-up with cross-sectional imaging, liver function tests, and cholangiography with repeat tissue sampling every 3 to 6 months.

Statistical Analysis.

Continuous variables were presented as mean ± standard deviation or median (range) and compared using standard parametric and nonparametric methods where appropriate. Frequency data were presented as number and percentage and compared using the chi-squared test or Fisher's exact test where appropriate. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy, with their exact 95% confidence intervals, were obtained based on the binomial distribution. All statistical testing was done at the conventional two-tailed level of 0.05.


A total of 235 PSC patients who underwent at least one FISH test between October 2003 and June 2008 were identified. The demographics, symptomatology, and laboratory values at the time when the patients had their first test if their results were negative, or at the time when their first FISH test was positive, or at the time when they had positive cytology or biopsy are presented in Tables 1 and 2. Most of the patients had more than one FISH test during repeated ERCP. Some patients had up to 13 test results, with some of the results varying from positive to negative and vice versa. For the sake of analysis, we considered patients with any positive test result as positive even though some of them on subsequent testing had a negative result. Clinicians tend to follow these patients with repeat ERCPs to validate the consistency of the cytology results rather than to pursue definitive treatment in the form of orthotopic liver transplantation when the diagnosis is in question.

Table 1. Demographics and Laboratory Values at Entry (n = 235)
Age48 ± 15 years
Sex (F/M)76/159
Abdominal pain64 (27%)
Jaundice100 (43%)
Weight loss56 (24%)
Smoking65 (28%)
GI bleed/portal hypertension/splenomegaly20 (9%)/92 (39%)/83 (35%)
Family history of cancer150 (64%)
Aspartate aminotransferase (AST) (U/L)83 (48–124)
Alanine aminotransferase (ALT) (U/L)83 (51–147)
Alkaline phosphatase (U/L)305 (197–471)
Bilirubin (mg/dL)2.4 (0.9–6.6)
Albumin (g/dL)3.8 ± 0.6
Duration of PSC (in years)7 (4–11)
Fever/itching48 (20%)/86 (37%)
Total protein (mg/dL)7 ± 0.8
Ulcerative colitis/Crohn174 (74%)/9 (4%)
Table 2. Diagnostic Tests Performed and Positive Results
  • *

    Twelve patients had both positive cytology and biopsy results.

FISH-trisomy (N = 235)65 (28%)
FISH-polysomy (N = 235)47 (20%)
FISH-tetrasomy (N = 235)8 (3%)
Cytology positive (N = 234)*33 (14%)
Cytology suspicious (N = 234)27 (12%)
Cytology atypical (N = 234)41 (18%)
Biopsy (N = 181)35 (18%)
MRCP (N = 183)33 (17%)
CT (N = 180)21 (12%)
US (N = 223)11 (5%)
EUS (N = 111)9 (8%)
CA 19-9 (U/L) (N = 213)30 (16–124)

FISH Testing.

Overall, 51% (120/235) of PSC patients tested for FISH were positive, of which only 33% (40/120) of the patients had CCA. A total of 47 of 120 (39%) patients had a positive polysomy FISH test, of which 26 (55%) patients had CCA. Patients with dominant strictures in the FISH polysomy-positive group more likely had CCA (19/26 [73%] versus 9/21 [43%]) than those without dominant strictures (P = 0.02). Importantly, 73 of 120 patients (61%) had a positive tetrasomy or trisomy 7 or 3 test, of which only 14 (19%) had CCA.

Cholangiocarcinoma Patients.

Thirty-five PSC patients had a positive histological diagnosis for CCA, and 21 had positive cytology for CCA. These were subgrouped as patients with gold standard diagnosis of CCA. Patients with the diagnosis of gallbladder cancer (n = 4) were not included in this group. The performance of FISH testing and cross-sectional imaging in this group is depicted in Table 3. No CCA was verified by histological or cytological methods in the remaining 179 PSC patients during follow-up. The demographics, symptomatology, and laboratory values between these two groups were compared (Table 4). The clinical symptoms of weight loss and jaundice along with male sex had significant associations with the presence of cancer. The presence of splenomegaly and portal hypertension had a negative association in patients with CCA. The CA 19-9 level was higher in patients with CCA, and the total protein level was lower in patients with CCA. No difference was observed in the levels of alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and albumin between the two groups (data not shown). There was no significant difference among the two groups in terms of the presence or absence of ulcerative colitis or Crohn disease (data not shown). Using a backward stepwise analysis, portal hypertension (P = 0.03; odds ratio [OR], 0.48) was less likely in those who had CCA, and weight loss (P < 0.0001; OR, 6.2) was more likely in those with CCA.

Table 3. PSC Patients with Cancer
 + Biopsy N = 35+ Cytology N = 21
Death23 (66%)6 (29%)
OLT15 (43%)14 (67%)
Dominant stricture24 (69%)12 (57%)
Suspicious cross-sectional imaging (MRI/CT/US/EUS)23 (66%)7 (33%)
FISH & Cytology +10 (29%)NA
FISH +14 (40%)16 (76%)
Cytology +2 (6%)NA
All negative (FISH/cytology)8 (23%)NA
FISH–NA5 (24%)
Table 4. Characteristics and Laboratory Values of PSC Patients With and Without Cholangiocarcinoma
CharacteristicWith Cancer (N = 56)Without Cancer (N = 179)P Value
Age50 ± 12 years47 ± 15 years0.1
Sex (F/M)9/4767/1120.003*
Abdominal pain15 (27%)49 (27%)1.0
Jaundice32 (57%)68 (38%)0.01*
Weight loss27 (48%)29 (16%)<0.0001*
Smoking21 (37%)44 (25%)0.04*
GI bleed3 (5%)17 (9%)0.4
PHT15 (27%)77 (43%)0.04*
Splenomegaly13 (23%)70 (39%)0.03*
Family history of cancer32 (57%)118 (66%)0.5
Fever9 (16%)39 (22%)0.4
Itching24 (43%)64 (35%)0.3
Duration of PSC (in years)7 (3–12)yrs8 (4–11) years0.3
CA 19-9 (U/L)71 (20–566)28 (15–99)0.0004*
Total protein (g/dL)7 ± 0.87.3 ± 0.80.02*
Bilirubin (mg/dL)4 (0.9–11)2 (0.9–5)0.03*
Dead29 (52%)19 (10%)<0.0001*
OLT29 (52%)42 (23%)<0.0001*


Survival illustrated by Kaplan-Meier analysis in patients with CCA and those without CCA is shown in Fig. 1. Furthermore, Fig. 2 shows a Kaplan-Meier survival curve in patients with negative FISH testing, patients with trisomy/tetrasomy, and those with FISH polysomy. Patients with polysomy on FISH have a poor outcome, whereas patients with trisomy/tetrasomy have a similar outcome to patients with negative tests.

Figure 1.

Kaplan-Meier survival for patients with and without cholangiocarcinoma.

Figure 2.

Kaplan-Meier curve comparing survival of FISH polysomy, trisomy 3/7/tetrasomy, and FISH-negative PSC patients.

The 179 PSC patients without cytology or tissue biopsy evidence of cancer were analyzed for the number of deaths, liver transplants, active follow-up, presence of dominant strictures, results of cross-sectional imaging, cause of deaths, and reason for liver transplantation (Table 5). Four patients died of underlying infection causing septic shock and multiorgan failure. There was no confirmatory evidence of CCA at the time of autopsy in one patient. No autopsy was performed in the other patients.

Table 5. Outcomes of FISH Testing
 − FISH (N = 99)+ FISH (N = 80)
 7 (7%)12 (15%)
Cause of death1. Advanced PSC (n = 3)1. Sepsis (n = 4)
2. End-stage renal disease2. Myocardial infarction (n = 1)
3. Metastatic carcinoid3. CCA (n = 3)
4. Gallbladder cancer4. Pancreatic cancer (n = 1)
5. CCA5. Gallbladder cancer (n = 2)
6. Advanced PSC (n = 1)
 20 (20%)22 (27%)
OLT1. Advanced PSC (n = 19)1. For CCA (n = 13)
2. For CCA (n = 1)2. For advanced PSC (n = 6)
One patient had incidental GB cancer in the explant.3. For IPMN done outside Mayo (n = 1)
 4. HCC (n = 1)
 5. Performed outside Mayo (n = 1)


In patients with histological evidence of CCA, orthotopic liver transplantation was performed in 29 patients, of whom 13 (45%) had evidence of CCA in the liver explants and 16 did not. The explant specimens typically showed extensive fibrosis and necrotic changes caused by intrabiliary radiation seed implantation, particularly around the hilar region, as expected with the PSC/CCA liver transplant protocol.24

A total of 22 orthotopic liver transplants were performed in patients with positive FISH but no evidence of cancer on histology or cytology. In 13 of these patients, the main indication for the transplant was CCA.24 Residual tumor was found in none of these, but small tumors are often treated with the intense radiation of this protocol.

Diagnostic Performance.

We evaluated the performance of FISH tests in the diagnosis of CCA. The proportions of patients with and without CCA in patients with positive FISH tests are shown in Table 6. The sensitivity, specificity, positive predictive value, and negative predicative values are depicted in Table 7. We also analyzed the clinical and laboratory characteristics of the patients with positive FISH polysomy who are actively being followed, without any evidence of CCA despite a median follow-up of 11 months (range, 7-34) with the patients who had a definitive diagnosis of CCA. The most important differences between these two groups showed that those who had CCA had significantly higher bilirubin and Mayo risk score, and most had dominant strictures, whereas none of those without CCA had dominant stricture (Table 8).

Table 6. FISH Test Results
 FISH (N = 235)
Trisomy/Tetrasomy (n = 73)Polysomy (n = 47)Neg (n = 115)
With cancer (N = 56)14 (18%)26 (55%)16 (14%)
Without cancer (N = 179)59 (82%)21 (45%)99 (86%)
Table 7. Diagnostic Performance of FISH for Detecting Cholangiocarcinoma in PSC
 Sensitivity (95% CI)Specificity (95% CI)PPV (95% CI)NPV (95% CI)
Polysomy46% (0.34–0.59)88% (0.83–0.92)55% (0.41–0.69)84% (0.78–0.89)
Trisomy 7 or 325% (0.15–0.38)67% (0.60–0.73)19% (0.11–0.30)74% (0.67–0.80)
Polysomy or trisomy72% (0.59–0.82)57% (0.49–0.64)34% (0.26–0.43)86% (0.79–0.92)
Table 8. FISH Polysomy
 FISH Polysomy with CCA (N = 26)FISH Polysomy Actively Follow-up No CCA (N = 7)P Value
  1. The 14 patients without CCA with positive polysomy result either underwent transplantation or died during follow-up.

Total bilirubin8 (1.8–12)0.9 (0.5–2.6)0.0002*
Mayo Risk Score1.9 (0.3–2.3)0.11 (−0.4–1.2)0.01*
Dominant stricture1900.0002*
CA 19-990 (23–1639)22 (17–286)0.3


FISH tests have been shown to play a potential role in the diagnosis of indeterminate biliary and pancreatic strictures. The results of our study seem to clarify some of the ambiguity associated with FISH tests in PSC and point to new directions in the evaluation of suspected CCA in PSC patients.

As observed in our study and in agreement with other reports, FISH tests increase the sensitivity of cytology at the cost of a decrease in specificity. In agreement with Levy et al.19 regarding the questionable utility of trisomy 7 or 3 positive FISH results in PSC patients, our results indicated that FISH trisomy 7 or 3 has a very low sensitivity and limited specificity for CCA. Our results indicate that most patients with a positive trisomy 7 or 3 do not have CCA at the time of testing, and less than 20% manifest CCA on short-term (1-year) follow-up. Thus, we question the need for repeat ERCP to follow-up on the results of a positive trisomy 7 or 3, especially in the absence of a dominant stricture. Although analysis of cost effectiveness has not been performed, ERCP has drawbacks in terms of complications. Chromosome 7 contains genes for the epidermal growth factor, c-Met, and interleukin-6, which have been implicated with bile duct carcinogenesis,25 so that cancers may develop later in these patients, and further study is needed.

DeHaan et al.,26 in a study of paraffin-embedded cholangiocarcinoma from PSC patients, observed polysomy not only in CCA but also in areas that had been interpreted as high-grade dysplasia (HGD).26 HGD of the bile ducts of PSC patients is the morphologic precursor to frank CCA. HGD has been observed to have a level of genetic abnormality by FISH that is similar to in situ and invasive carcinoma in other settings such as Barrett's esophagus.27, 28 It is likely that the development of CCA in PSC patients is preceded by one or more foci of HGD. It may take months or years for areas of HGD in PSC patients to progress to CCA, and in some cases this progression may not occur. The finding of polysomy in HGD in PSC patients indicates that this genetic alteration is not absolutely specific for CCA in PSC patients. We believe that when polysomy is observed in patients with other concerning findings (such as a dominant stricture), it has a high positive predictive value for the presence of CCA. However, when such additional clinical findings are not present, the positive predictive value of polysomy for CCA is significantly lower. Polysomy in PSC patients without additional concerning clinical findings should be interpreted more cautiously. Its occurrence in such patients may indicate that they are at higher risk of developing CCA but may not actually have frank CCA.

Our results indicate that FISH testing should not be used as a screening modality in unselected PSC patients undergoing ERCP. However, in patients with clinical or laboratory suspicion of CCA, such as weight loss, abdominal pain, dominant stricture, or high CA 19-9, these tests can be helpful. The analysis of our findings suggests the following guidelines: If a positive trisomy or tetrasomy are obtained without evidence of CCA on imaging, cross-sectional imaging should be repeated 3 months later. If other features such as dominant stricture, prominent CA 19-9 elevation, or mass are present, cross-sectional imaging and ERCP should repeated at 3 months. These patients should thereafter be followed clinically as are other PSC patients with CA 19-9 levels and ultrasound at 6 months and then annually, as recently shown to be effective.9 The presence of FISH trisomy or tetrasomy does not indicate a high likelihood of CCA.

If patients with positive polysomy are not found to have CCA at the initial examination, we would repeat the evaluation after 3 months. According to our Kaplan Meier analysis, patients with positive polysomy very rarely die within 3 months. In these patients, it is appropriate after 3 months to perform magnetic resonance imaging/magnetic resonance cholangiography, measure CA 19-9, and characterize stricture development on ERCP and repeat cytology and FISH testing. Approximately half of the patients with positive polysomy develop CCA, but not all do. Follow-up testing with ERCP should depend on occurrence of mass on MRI, stricture development on MRC, presence of prominent CA 19-9 elevation, and whether other clinical symptoms that might be associated with CCA are present.

In conclusion, the results of FISH tests need to be interpreted with caution in PSC patients. FISH trisomy/tetrasomy-positive results have very limited implications in PSC patients. A positive FISH polysomy test result does enhance the sensitivity of cytology testing for CCA, especially if a dominant stricture is present. Results of FISH should be interpreted in association with patient's clinical, laboratory, and cholangiographic features.