Interinstitutional variability and effect of tissue fixative on the interpretation of a Barrett cytokeratin 7/20 immunoreactivity pattern in Barrett esophagus.

A unique pattern of cytokeratin (CK) 7/20 immunostaining (diffuse staining with CK7 and surface and superficial crypt staining with CK20) has been reported to be useful in differentiating Barrett esophagus (BE) from intestinal metaplasia of the stomach. However, there are conflicting results regarding the prevalence of a BE CK7/20 staining pattern in BE between different studies. Therefore, this study was performed to determine the degree of variability in interpretation of a BE CK7/20 pattern and to determine the reasons for variability when present. Esophageal and gastric mucosal biopsies from 67 patients with BE and antral intestinal metaplasia at 2 institutions were immunostained for CK7/20. All cases were evaluated for the presence of a BE CK7/20 pattern by 2 gastrointestinal pathologists from each institution, and the degree of agreement between institutions was determined. To determine the effect of tissue fixation and staining methods on the pattern of CK7/20 staining, unstained slides were exchanged between institutions, stained separately by each institution, and reexamined by all pathologists. There was excellent agreement on the presence of a BE CK7/20 staining pattern between pathologists at the same institution but only moderate agreement between pathologists at different institutions (71% overall, kappa = 0.58). Among BE cases, a BE CK7/20 staining pattern was identified in 50 (96%) of 52 cases by Cleveland Clinic Foundation pathologists but only 35 (67%) of 52 cases by Brigham and Women's Hospital pathologists. The major source of disagreement related to the interpretation of weak or variable CK7 staining of deep intestinalized mucosa in BE biopsies that were fixed in Hollande, but not those that were fixed in formalin. After the creation of a new set of criteria for a positive BE CK7/20 staining pattern, which took into account the effects of Hollande's fixative, the degree of agreement between pathologists at each of the 2 institutions was excellent (100%, kappa value = 1.0). Therefore, the CK7/20 staining pattern is influenced by the type of fixative used. Only a moderate level of interobserver agreement among pathologists regarding a BE CK7/20 pattern can be achieved if one is not aware of these effects. Nevertheless, specific criteria for interpretation of CK7/20 staining can be successfully applied between institutions and need to be developed before use of this technique in clinical practice.

In vivo colonoscopic optical coherence tomography for transmural inflammation in inflammatory bowel disease.

BACKGROUND & AIMS: Transmural inflammation, a distinguishing feature of Crohn's disease (CD), cannot be assessed by conventional colonoscopy with mucosal biopsy. Our previous ex vivo study of histology-correlated optical coherence tomography (OCT) imaging on colectomy specimens of CD and ulcerative colitis (UC) showed that disruption of the layered structure of colon wall on OCT is an accurate marker for transmural inflammation of CD. We performed an in vivo colonoscopic OCT in patients with a clinical diagnosis of CD or UC using the previously established, histology-correlated OCT imaging criterion. METHODS: OCT was performed in 40 patients with CD (309 images) and 30 patients with UC (292 images). Corresponding endoscopic features of mucosal inflammation were documented. Two gastroenterologists blinded to endoscopic and clinical data scored the OCT images independently to assess the feature of disrupted layered structure. RESULTS: Thirty-six CD patients (90.0%) had disrupted layered structure, whereas 5 UC patients (16.7%) had disrupted layered structure (P < .001). Using the clinical diagnosis of CD or UC as the gold standard, the disrupted layered structure on OCT indicative of transmural inflammation had a diagnostic sensitivity and specificity of 90.0% (95% CI: 78.0%, 96.5%) and 83.3% (95% CI: 67.3%, 93.3%) for CD, respectively. The kappa coefficient in the interpretation of OCT images was 0.80 (95% CI: 0.75, 0.86, P < .001). CONCLUSIONS: In vivo colonoscopic OCT is feasible and accurate to detect disrupted layered structure of the colon wall indicative of transmural inflammation, providing a valuable tool to distinguish CD from UC.

Ex vivo histology-correlated optical coherence tomography in the detection of transmural inflammation in Crohn's disease.

BACKGROUND AND AIMS: Distinguishing Crohn's disease (CD) from ulcerative colitis (UC) can be difficult. Transmural inflammation, a key feature of CD, cannot be assessed by conventional colonoscopy with biopsy. Optical coherence tomography (OCT) provides high-resolution, cross-sectional images of the gut wall and might become a new diagnostic tool. The aims of this study were to perform histology-correlated OCT on surgical specimens of CD and UC and to determine its diagnostic accuracy. METHODS: Colectomy specimens from patients with a preoperative diagnosis of CD (N = 24) or UC (N = 24) were studied with OCT in the operating room. OCT and histopathology were assessed blindly, and diagnostic accuracy of OCT was assessed. RESULTS: Eight preoperatively identified UC patients (33%) with transmural inflammation on postoperative histology were diagnosed with CD, and all 8 had a disrupted layered structure on OCT, a characteristic feature of transmural disease. Sixteen UC patients (67%) had superficial inflammation on histology; of them, 13 (81%) had an intact layered structure on OCT. All 24 preoperative CD patients had transmural inflammation on histology, and 23 (96%) had a disrupted layered structure on OCT. Of 585 histology-OCT image sets from the 48 patients, 152 sets (26%) had transmural inflammation on histology. The sensitivity and specificity for OCT to detect transmural disease were 86% and 91%, respectively. CONCLUSIONS: Transmural inflammation, as characterized by disruption of the layered structure of colon wall on OCT, is an accurate marker for the diagnosis of CD. Ex vivo OCT predicted transmural inflammation on postoperative histopathology.

Fluorescence in situ hybridization of cytologic specimens from Barrett's esophagus: a pilot feasibility study.

BACKGROUND: Endoscopic brush cytology is a promising surveillance technique for Barrett's esophagus. However, there is a need for ancillary biomarkers to increase the sensitivity of cytology and to allow identification of patients at increased risk for disease progression. The aims of this study were to evaluate the feasibility of fluorescence in situ hybridization of endoscopic brush cytology specimens and to determine if there are specific chromosomal changes in cytologic specimens from patients with cancer that are not present in patients without dysplasia. METHODS: Archival cytology slides from 16 patients with Barrett's esophagus were studied: 8 negative for dysplasia and 8 positive for adenocarcinoma. Fluorescence in situ hybridization was used to detect two alterations: HER-2 gene (17q11.2-q12) and 20q13.2 region amplification. OBSERVATIONS: For 7 of 8 adenocarcinoma cases, there was amplification/aneusomy of at least one of the two analyzed regions by fluorescence in situ hybridization. None of the samples negative for dysplasia were abnormal for either of the two genomic regions studied. CONCLUSIONS: Fluorescence in situ hybridization is feasible by using routine Barrett's esophagus cytologic specimens. Differences in genomic makeup can be detected in cells from patients negative for dysplasia and in those with adenocarcinoma.

Chromosomal gains and genomic loss of p53 and p16 genes in Barrett's esophagus detected by fluorescence in situ hybridization of cytology specimens.

Endoscopic brush cytology is a promising surveillance technique for Barrett's esophagus. Ancillary markers are sought to increase the sensitivity of cytology and allow identification of patients at increased risk for disease progression. To determine if there are specific genetic changes in Barrett's esophagus with associated high-grade dysplasia/intramucosal adenocarcinoma compared to those without dysplasia, we performed fluorescence in situ hybridization (FISH) on cytologic specimens using probes to chromosomes and genomic regions previously described as altered in this disease. We studied archival brush cytology slides from 40 Barrett's esophagus patients: 21 with biopsy-proven high-grade dysplasia/carcinoma and 19 with no dysplasia and a minimum 5 years of negative follow-up. Centromeric enumeration probes (CEP) for chromosomes 6, 7, 11, and 12, and locus-specific probes (LSI) for 9p21 (p16 gene), and 17p13.1 (p53 gene) loci along with their corresponding CEP (9 and 17, respectively) were used in this study. A positive FISH result was defined as the presence of cells with >2 CEP signals or with a loss of the LSI signals relative to their corresponding CEP. p53 locus loss and/or aneusomy of chromosomes 6, 7, 11, and 12 abnormalities could be detected by FISH in routinely processed endoscopic brush cytology specimens from 95% of biopsy-positive cases with a specificity of 100%. Interestingly, all five cases with cytologic changes classified as indefinite for dysplasia from patients with a positive biopsy showed changes by FISH. Loss of the p16 locus was seen commonly in patients both with and without dysplasia/carcinoma. Selected biomarkers from this study merit further investigation to determine their potential to detect genetic changes in patients with Barrett's esophagus prior to the development of high-grade dysplasia.

p53 expression in low grade dysplasia in Barrett's esophagus: correlation with interobserver agreement and disease progression.

OBJECTIVES: The frequency of progression from low grade dysplasia (LGD) to high grade dysplasia/carcinoma (HGD/ CA) in Barrett's esophagus (BE) varies among studies. Current assessment is made more difficult because of pathologists' interobserver variability in diagnosing LGD. We recently conducted an interobserver study on LGD and reported a positive correlation between the extent of agreement among GI pathologists and progression of LGD. In the current study, we analyzed the immunohistochemical staining for p53 in patients diagnosed with LGD with known clinical outcome and interobserver agreement data. METHODS: Fixed, paraffin-embedded endoscopic biopsy specimens from 16 patients diagnosed with LGD in BE were immunostained for p53 (DO-7, Dako, Carpinteria, CA). Hematoxylin and eosin-stained and immunostained sections were examined in tandem to determine whether the LGD areas in question stained for p53. The p53 immunoreactivity was correlated with clinical progression and with the interobserver agreement among three GI pathologists. RESULTS: The overall mean follow-up was 23 months (range 2-84 months). LGD areas in seven of eight patients (88%) who progressed to HGD/CA stained positively for p53 compared to only two of eight nonprogressors (25%). A correlation with clinical progression was seen for p53 positivity (p = 0.017; log-rank test), and for either p53 positivity or complete agreement among three GI pathologists on LGD diagnosis (p = 0.014; log-rank test). The p53 staining demonstrated 88% sensitivity and 75% specificity for progression of LGD to HGD/CA. Adding complete interobserver agreement on LGD among three experienced GI pathologists to p53 positivity resulted in improved sensitivity with no change in specificity (100% and 75%, respectively). CONCLUSIONS: In conjunction with histological evaluation by GI pathologists for a diagnosis of LGD, immunohistochemical staining for p53 can be used as an adjunctive test, as it correlated with progression to HGD/CA in this series.

Cytokeratin expression patterns in noncardia, intestinal metaplasia-associated gastric adenocarcinoma: implication for the evaluation of intestinal metaplasia and tumors at the esophagogastric junction.

BACKGROUND: Barrett esophagus (BE)/Barrett adenocarcinoma and distal gastric intestinal metaplasia (IM)/adenocarcinoma are similar histologically, but they differ in their clinical presentation, epidemiology, and pathogenesis. Differentiating BE from gastric IM and Barrett adenocarcinoma from gastric adenocarcinoma is difficult, especially when IM is short or tumors are large and involve both sides of the esophagogastric junction. Previously, the authors identified unique cytokeratin (CK) immunoreactivity patterns that were associated strongly with BE and Barrett adenocarcinoma. The specificity of CK7 and CK20 (CK7/20) expression patterns in patients with IM-associated gastric adenocarcinoma, which is distinct epidemiologically from BE/Barrett adenocarcinoma, has not been evaluated. The objective of the current study was to evaluate the CK7/20 expression patterns in noncardia, IM-associated gastric adenocarcinoma in a Chinese population with a low risk for BE and esophageal adenocarcinoma and a high risk for Helicobacter pylori infection and gastric carcinoma. METHODS: Endoscopic biopsy specimens of gastric IM and adjacent tumor from 50 consecutive patients with advanced noncardia gastric carcinoma were immunostained for CK7 and CK20. Clinical and endoscopic features and H. pylori status were documented. Two gastrointestinal pathologists, blinded to clinical and endoscopic data, independently assessed CK7/20 immunohistochemistry. RESULTS: H. pylori infection was present in 43 of 50 patients (86%). In the area of IM, patchy CK7 staining was seen in 9 patients (18%), and diffuse CK20 staining was seen in all 50 patients (100%). The BE CK7/20 pattern characterized by CK7 staining in superficial and deep glands and the CK20 staining in surface epithelium was not seen in any of the 50 patients. Only one patient (2%) demonstrated a CK7 positive/CK20 negative immunophenotype characteristic of Barrett adenocarcinoma. The remaining 49 patients (98%) showed non-Barrett adenocarcinoma patterns of CK7/20 staining, i.e., a CK7 positive/CK20 positive pattern was seen in 33 patients (66%), a CK7 negative/CK20 positive pattern was seen in 12 patients (24%), and a CK7 negative/CK20 negative pattern was seen in 4 patients (8%). CONCLUSIONS: In a patient population without risk factors for the development of BE/esophageal adenocarcinoma, the CK7/20 pattern characteristic of BE was not present in gastric IM adjacent to adenocarcinoma, and the CK7/20 pattern characteristic of Barrett adenocarcinoma also was extremely rare. These results support the hypothesis that, despite the presence of intestinalized mucosa in both disorders, BE/Barrett adenocarcinoma and gastric IM/adenocarcinoma are two distinct clinical entities with unique demographic, clinical, and CK immunophenotypic findings. These results may have application to the evaluation of patients with IM and adenocarcinoma arising at the esophagogastric junction.

Thin-layer Pap test vs. conventional Pap smear. Analysis of 400 split samples.

OBJECTIVE: To analyze our experience with 400 Thin-Prep (TP) split samples (Cytyc Corp., Boxborough, Massachusetts) as an initial assessment of this new technology's effect in our laboratory. STUDY DESIGN: Three gynecologic oncologists and two general gynecologists obtained the 400 split samples using a broom sampling device. Following conventional smear (CS) preparation, they rinsed the broom in Preservcyt solution (Cytyc) for subsequent TP processing. The paired samples were separated, independently analyzed and classified by the Bethesda System. All available follow-up surgical pathology material was reviewed and compared to the cytologic diagnoses. RESULTS: TP had significantly more abnormal results (22% vs. 16%, P = .007), including more atypical squamous cells of undetermined significance (ASCUS) (9.5% vs. 6.3% P = .07) and low grade squamous intraepithelial lesion (LSIL) (7.8% vs. 5.3%, P = .03). Both methods had 3.3% high grade squamous intraepithelial lesion (HSIL). For TP, ASCUS/squamous intraepithelial lesion (SIL) = 0.86 and for CS, ASCUS/SIL = 0.74. Ten TP SILs had a paired negative CS, including LSIL (nine cases) and HSIL (one case). Consensus review of these 10 TP slides confirmed the HSIL and four LSILs. No CS SILs had a paired negative TP. Only 36 (9%) cases had surgical pathology follow-up. The surgical specimens included 17 cervical intraepithelial neoplasia (CIN) 2 or above. The TP method had no false negatives, while the CS method had 3 false negatives among the 17 confirmed cases of CIN 2 or above. CONCLUSION: TP appears to be superior to CS for detecting SILs.