Fluorescent in situ hybridization testing in primary sclerosing cholangitis: Useful testing or a random chance result?


  • William C. Chapman M.D.

    Corresponding author
    1. Department of Surgery, Section of Abdominal Transplantation, Washington University School of Medicine, St. Louis, MO
    • Professor and Chief, Division of General Surgery, Chief, Section of Transplantation, Washington University School of Medicine, 660 South Euclid Avenue, Campus Box 8109, St. Louis, MO 63110
    Search for more papers by this author
    • fax: 314-361-4197

  • See Article on Page 174

  • Potential conflict of interest: Nothing to report.

Primary sclerosing cholangitis (PSC) has long been recognized as a predisposing factor for the development of cholangiocarcinoma, although the causative risk factors remain largely unknown. By definition, patients with PSC have fibrosing inflammatory strictures of the biliary tract, usually multiple and diffuse, that may affect the intrahepatic and/or extrahepatic biliary tree. This process leads to progressive cholestatic liver disease, which may cause unremitting pruritus, cholangitis, and ultimately liver failure. Liver transplantation is the only known curative therapy, and is usually very effective. Approximately 5%-10% of patients undergoing liver transplantation at most centers have PSC as the indication, with reported 10-year survival rates approaching 80% in many reports.1


ERCP, endoscopic retrograde cholangiopancreatography; FISH, fluorescent in situ hybridization; PSC, primary sclerosing cholangitis.

The risk of developing cholangiocarcinoma in the setting of PSC has been well recognized for many years, and is now a dreaded complication feared by patients with PSC, their families, and their physicians. Although the lifetime risk of developing cholangiocarcinoma in patients with PSC is estimated to be more than 1500 times greater than in the non-PSC population,2 it is also equally clear that “early” liver transplantation should not be undertaken, because the risks of this complex procedure and life-long immunosuppression outweigh any potential risk of cancer reduction. The estimated risk of development of cholangiocarcinoma in patients with PCS is approximately 1% per year over 10 years and a cumulative lifetime risk of 9%-31%,3 but the mortality risk after liver transplantation is approximately 10% in the first year alone. The challenge therefore, is how best to perform adequate surveillance testing for the detection of early-stage cholangiocarcinoma in patients with PSC but who do not otherwise need liver replacement therapy. Unlike for non-PSC patients with biliary tract strictures, the development of strictures in the setting of PSC is usually part of the natural history and not unexpected. In addition, confirmation of this malignancy may be one of the most difficult challenges in all solid organ malignancies, because access is highly restricted, direct biopsies are difficult to obtain, and cytologic assessment with brush analysis is frequently falsely negative, probably related to the desmoplastic reaction so frequently associated with such tumors.

Based on the significant diagnostic challenges of standard biopsy techniques and low sensitivity of cytology for hilar cholangiocarcinoma, investigators at the Mayo Clinic have spearheaded efforts to develop additional techniques to improve diagnostic accuracy in this malignancy. In particular, the use of digital image analysis and fluorescent in situ hybridization (FISH) have been investigated for a variety of malignancies of the liver, biliary tract, and pancreas,4–8 and multivariate modeling using these advanced techniques has been proposed for use in indeterminate biliary strictures.9 Of note, the testing and validation of FISH in cholangiocarcinoma including the specific probes and assessment techniques has largely taken place in patients who were selected for investigation based on a high likelihood of cholangiocarcinoma.

In this issue of HEPATOLOGY, Bangarulingam et al. report their experience in a large cohort of patients with PSC, all of whom were undergoing endoscopic retrograde cholangiopancreatography (ERCP) for management of PSC, but many apparently without dominant strictures or mass lesions on cross-sectional imaging to support a diagnosis of cholangiocarcinoma.10 A total of 235 patients with PSC underwent one or more FISH analyses over an almost 5-year time period, among whom 56 (24%) patients had biopsy or cytologic evidence of cholangiocarcinoma during the period of study. Although 120/235 (51%) of patients had a positive FISH test at some point during the period of study, only one-third (40/120, 33%) of these were ultimately diagnosed with cholangiocarcinoma by tissue analysis. Put another way, two-thirds (80/120, 66%) of the patients with a positive FISH did not have cholangiocarcinoma. The authors note the very low sensitivity and specificity of tetrasomy or trisomy FISH results: only 14/73 (19%) FISH-positive tetrasomy or trisomy patients were ultimately diagnosed with cholangiocarcinoma. The sensitivity and specificity were improved with polysomy results, nevertheless, only 26/47 (55%) patients with positive polysomy FISH had cholangiocarcinoma. Thus, the positive predictive value of FISH polysomy was 55%, only slightly better than a coin toss. And it was disappointingly worse at only 19% for trisomy 7 or 3.

Use of FISH analysis for diagnosing cholangiocarcinoma is limited by the technical difficulties of obtaining pathologic material for analysis. In the current report, Bangarulingam et al. used commercially available probes to assess chromosomes 3, 7, 17, and 9p21, and considered the assessment “positive” if ≥ 5 cells showed gains for polysomy or ≥ 10 cells showed gains for tetrasomy or trisomy. The authors' selection of as few as 5 or 10 cells to be considered a positive was likely based on the extensive prior work in a group of patients more likely to harbor malignancy. Whether or not selection of additional chromosomal targets and/or increasing the number of positive cells required for a “positive” diagnosis would increase the sensitivity, specificity and positive predictive value is unknown, but should be an area of further investigation. Aneuploidy could possibly be characterized using high-resolution microarray-based technology for global copy number analysis. Because whole-genome amplification methods can be used, this test has the benefit of availability when material is scarce, however, such considerations are only speculative at this time.

So what exactly are patients with PSC (and their physicians and surgeons) to do in regards to FISH testing? The authors note in the discussion that patients positive for FISH polysomy, but without documented evidence of cholangiocarcinoma (negative histology and cytology), should undergo repeat evaluation at 3-month intervals because “patients with polysomy rarely die within 3 months”. In this regard, the authors appear to be appropriately questioning the low positive predictive value of FISH polysomy. However, from a clinical care provider's point of view, one has to question if it is reasonable to expect that a PSC patient undergoing evaluation for cholangiocarcinoma would comfortably accept having only repeat evaluations to search for positive histology, positive cytology, or significant changes on magnetic resonance study as the next move before more definitive intervention. How will patients who are already terrified about the risk of developing cholangiocarcinoma react to being informed that the “cytologic analysis” suggests malignancy; however, “we recommend continued follow-up visits until more definitive testing becomes positive”?

Historically, results for orthotopic liver transplantation (OLT) for cholangiocarcinoma have been poor and most centers have considered known cholangiocarcinoma a contraindication.11 More recent reports from the University of Nebraska and the Mayo Clinic have demonstrated satisfactory long-term overall and disease-free survival when the indication for OLT is cholangiocarcinoma,12–14 as long as OLT is performed as part of a neoadjuvant chemoradiation protocol and when pre-transplant staging laparotomy demonstrates no evidence of metastatic disease in regional lymph nodes or elsewhere. Amongst patients with unresectable cholangiocarcinoma, those with PSC associated cholangiocarcinoma may represent the “best” candidates for OLT, since the disease in these patients is often “unresectable” because of the PSC-associated fibrosis, and not because of advanced-stage cholangiocarcinoma. In the most recent Mayo series reports, survival following OLT for cholangiocarcinoma, in patients with and without associated PSC, is essentially the same as for liver transplantation for all other indications (i.e., approximately 70%-75% at 5 years).15 Based on these improved results for transplantation for cholangiocarcinoma under a neoadjuvant protocol, the United Network for Organ Sharing (UNOS)16 made recommendations that patients with unresectable hilar cholangiocarcinoma, less than 3 cm in largest diameter, without evidence of metastatic disease to regional nodes or elsewhere, be granted 22 Model for End-Stage Liver Disease (MELD) exception points. The patients must also have documentation of malignancy, successful completion of pretransplant chemoradiation protocol, and pretransplant staging laparotomy with regional lymph node sampling to further exclude metastatic disease. In March 2006, the MELD exception study group convened a consensus conference in Chicago and considered 18 disease specific areas for transplant priority consideration, including cholangiocarcinoma. As a consequence of this consensus conference, the MESSAGE recommendations were formulated and proposed to UNOS as a basis for non-binding recommendations for regional review boards to follow in such cases as “guidelines”. At that time, the authors of the cholangiocarcinoma subgroup, lead by Dr. Gores from the Mayo Clinic, recommended that establishment of the diagnosis of cholangiocarcinoma could be considered “positive” on the basis of one of the following factors: (1) a positive biopsy or brush cytology; (2) an elevated CA19-9 greater than 100 U/mL in the appropriate clinical setting; or (3) aneuploidy on FISH analysis.16 Thus, for most transplant surgeons and physicians (and members of regional review boards), the presence of “aneuploidy” or “positive FISH analysis” has essentially been a “yes/no” question (Fig. 1). On the basis of the data presented by the Mayo group in the current manuscript, however, it is clear that greater care will be required in the interpretation of results of FISH. Whether or not the prior recommendations to grant MELD exception points for a positive FISH analysis should still hold remains to be seen, but should probably be reconsidered. Clearly it is not sufficient to recommend chemoradiation and liver transplantation on the basis of aneuploidy on FISH analysis, if, as this large study has shown, this could mean trisomy or tetrasomy with only a 19% positive predictive value.

Figure 1.

The authors of the current report are to be commended for publishing what appears to be a disappointing set of results. Use of genomic analysis for diagnostic purposes and selection of therapeutic intervention is the way of the future in cancer management and this will likely be true for cholangiocarcinoma as well. However, until we have a better understanding of more accurate predictors of aneuploidy that correlate with cholangiocarcinoma, my recommendations in cases without positive biopsy or cytology and partially informed by the current study, would be as follows:

  • 1Trisomy and tetrosomy analysis should not be reported as part of FISH analysis in patients with PSC and suspected cholangiocarcinoma.
  • 2UNOS granting of exception points using FISH analysis as a basis for transplant listing should require positive polysomy for a diagnosis of aneuploidy.
  • 3Granting of MELD exception points on the basis of an elevated CA19-9 or aneuploidy on FISH analysis should also require the finding of a dominant stricture or an enhancing mass on cross-sectional imaging.


Shashikant Kulkarni, Ph.D., Assistant Professor of Pediatrics, Pathology and Immunology, and Director of CytoGenomics and Molecular Pathology at Washington University School of Medicine, St. Louis, MO, provided valuable assistance in review of technical methods reported in the manuscript under review.