“FISHing” for an Accurate Diagnostic Test for Cholangiocarcinoma

Abstract

Navaneethan U, Njei B, Venkatesh PG, et al. Fluorescence in situ hybridization for diagnosis of cholangiocarcinoma in primary sclerosing cholangitis: a systematic review and meta-analysis. Gastrointest Endosc 2014;79:943–950. Primary sclerosing cholangitis (PSC) is a chronic fibroinflammatory condition epitomized by progressive obliteration of the biliary tree. Although the natural history is variable, secondary biliary cirrhosis and end-stage liver failure occur in ≤50% of patients. In addition to managing the symptoms of PSC (predominantly pruritus and recurrent acute cholangitis), clinicians and patients remain vigilant for the interval development of cholangiocarcinoma (CCA), the lifetime incidence of which is 6%–20% (Am J Gastroenterol 2007;102:107–114). With improvements in cross-sectional imaging, PSC diagnosis and surveillance can be accomplished without the routine use of endoscopic retrograde cholangiopancreatography (ERCP). However, individuals with symptoms or signs of disease progression are referred for ERCP to palliate dominant strictures via dilation with or without short-term stent placement, and tissue sampling to evaluate for CCA. In this context, traditional intraductal tissue sampling techniques include brushings for cytology and forceps biopsies for histology. These modalities, alone or in combination, have limited (20%–60%) sensitivity for confirming CCA owing to the desmoplastic nature of these lesions, so clinicians cannot definitively “rule out” CCA with a negative result (Gastrointest Endosc 1995;42:565–572). Because cytopathology and histopathology are essentially qualitative tests that rely on a subjective assessment of cellular morphology, fluorescence in situ hybridization (FISH) has been studied as a supplementary assay. FISH is performed on intraductal brushing samples and uses fluorescently labeled probes that target the centromeres of chromosome 3, 7, and 17, as well as the 9p21 band (P16), evaluating for aneuploidy (abnormal number of chromosomes) by counting the number of affected cells; therefore, it is a quantitative test that requires theoretically fewer cells than cytopathology and removes the subjectivity associated with cytopathologic interpretation. Navaneethan et al (Gastrointest Endosc 2014;79: 943–950. e3) performed a structured metaanalysis of studies assessing the diagnostic utility of FISH in diagnosing CCA among patients with PSC. The authors included studies that allowed construction of a 2 × 2 contingency table with true-positive, false-negative, false-positive, and true-negative results. The authors used standard metaanalytic techniques to calculate composite sensitivity, specificity, likelihood ratios, and diagnostic odds ratio. There was moderate statistical heterogeneity among included studies, so the authors used a random effects model. A positive FISH for malignancy was defined as either polysomy (≥5 cells with gains of ≥2 of 4 probes) or trisomy/tetrasomy (trisomy, ≥10 cells with 3 copies of chromosome 7 or 3; tetrasomy, ≥10 cells with 4 copies of all probes). Eight studies of moderate to high methodologic quality (n = 828 individuals with PSC) met eligibility criteria. The pooled analysis from these studies demonstrated that FISH had an overall sensitivity of 68% (95% CI, 61%–74%), and specificity of 70% (95% CI, 66%–73%) for diagnosing CCA. The pooled positive likelihood ratio was 2.69 (95% CI, 1.84–3.97), and the negative likelihood ratio was 0.47 (95% CI, 0.39–0.58). The pooled odds ratio to detect CCA was 7.24 (95% CI, 3.93–13.36). The positive predictive value was 0.42 (95% CI, 0.36–0.47), and the negative predictive value was 0.87 (95% CI, 0.84–0.90). Stated alternatively, an individual with PSC and positive FISH had a 42% posttest probability of CCA and an 87% probability of being cancer free after negative FISH. Two of the included studies described the accuracy of FISH in the setting of negative cytology, reporting sensitivities of 57% (Gastroenterology 2006;131:1064–1072) and 60% (Am J Gastroenterol 2008;103:1263–1273) and specificities of 87% and 71%. This is the context in which FISH would be considered most valuable: when CCA is present, but routine cytology is negative for malignancy, FISH will be positive in ≤60% of cases. However, there remains a possibility that FISH is falsely positive, as reflected its lower specificity. The specificity of FISH seemed to be optimized (93%) by defining a positive test using polysomy alone (6 of 8 included studies), but this compromised the technique’s sensitivity (51%). The authors systematically removed 1 dataset at a time and recalculated their outcomes; no single dataset carried sufficient weight to influence significantly the pooled performance of FISH for diagnosing CCA in the setting of PSC.