Aug 19, 2011 - GS6 (86%) or GS3+4 (9%) and all had node negative pT2 disease. ...... reads (FPKM) values were computed using Cufflinks in the Galaxy.
190A Sensitivity, specificity, positive/negative predictive value, and the odds ratio of the 2 markers were summarized in Table 1. There was no difference in p16 expression between the DG and control group. Table 1. Summary of sensitivity, specificity, PPV, NPV, and OR. Sensitivity (%) Specificity (%) PPV (%) NPV (%) OR beta-catenin 92.3 76.9 80.0 90.9 40.0 c-met 69.2 61.5 64.3 66.7 3.6 Either 100.0 53.8 68.4 100.0 13.9* Both 61.5 92.3 88.9 70.6 19.2 NPV, negative predictive value; PPV, positive predictive value; OR, odds ratio. *, estimated OR.
Conclusions: The combination of c-met and beta-catenin provides a relatively specific panel to identify patients with BE who are at increased risk for future development of dysplasia. 783 Microfibrillar Associated Protein 5 (MFAP 5): A Marker for Desmoplasia That Facilitates the Distinction between an Invasive Component and Pseudoinvasion in Colonic Adenomas L Zhao, T Antic, S-Y Xiao, C VanSlambrouck, J Hart. University of Chicago, Chicago, IL. Background: Distinction between invasive adenocarcinoma and submucosal displacement of adenomatous mucosa in colonic polyps can be problematic in clinical practice. Recent microarray studies have demonstrated that tumor stroma exhibits reproducible gene-expression changes compared to normal stromal tissue. MFAP5, a 25kD glycoprotein that is involved in elastic microfibril assembly, has been demonstrated to be significantly down regulated in tumorous stroma [Jia et al. Cancer Res 2011]. The aim of this study is to confirm the reduced expression of MFAP5 in tumor stroma by immunohistochemistry and evaluate the utility of MFAP5 as a diagnostic marker. Design: An immunohistochemical stain for MFAP 5 (Sigma) was first performed on ten invasive colon cancer resection specimens to compare the its expression pattern in invasive tumor and in the surrounding lamina propria and submucosa. Then a total of 24 diagnostically challenging adenomatous polypectomy specimens (6 with an invasive focus and 18 with pseudoinvasion) were used to evaluate the diagnostic value of MFAP5. Results: In all 10 colon cancer resection cases there was no reactivity in the desmoplastic stroma surrounding the invasive component, while the uninvolved lamina propria and submucosa exhibited strong diffuse reactivity in stromal cells.
In 16 out of 18 polypectomy specimens with pseudoinvasion, strong reactivity was preserved in the stroma surrounding the displaced adenomatous mucosa (left panel of figure below). Lack of staining in the stroma surrounding small foci of invasive tumor was observed in all 6 malignant polyps (right panel of figure below).
Conclusions: MFAP5 is a useful marker to demonstrate tumor associated desmoplastic stromal proliferation. It facilitates the distinction between pseudoinvasion and true invasive cancer in colonic adenomatous polyps with high sensitivity and specificity. 784 ALK Status in Esophageal Adenocarcinoma and Squamous Carcinoma Studied by DNA Microarray, FISH and Immunohistochemistry Z Zhou, S Bandla, J Ye, L Li, T Godfrey, N Wang. University of Rochester Medical Center, Rochester, NY. Background: The anaplastic lymphoma kinase (ALK) gene encodes a tyrosine kinase receptor that belongs to the insulin receptor superfamily. The 3’-tyrosine kinase domain is fused to a variety of partners such as NPM-ALK or EML4-ALK leading to express fusion proteins that play oncogenic roles in a variety of malignancies. Using mass spectroscopy alone, TPM4-ALK fusion protein was detected in esophageal squamous cell carcinoma (SCC). To verify the possible role of ALK in SCC and esophageal adenocarcinoma (EAC), we use different approaches to study ALK gene arrangement, amplification or expression. Design: Genomic DNA from 116 EAC (95 M and 21 F) fresh tissue was analyzed for copy number aberrations using Affymetrix SNP 6.0 arrays. Tissue microarrays constructed at the University of Rochester between 1997 and 2005 included squamous mucosa (SE), EAC and SCC. ALK amplification and rearrangement were detected by FISH and ALK expression was tested by immunohistochemistry (IHC). Results: By genomic analysis, ALK amplification was found in 7% (8/116) EAC frozen tissue cases by SNP analysis which was confirmed by FISH test in TMA cases (7%; 5/74). However, no ALK expression was identified either in EAC (0/112) or in SCC (0/33) by IHC. No gene rearrangement of ALK was detected in either EAC (0/74) or
ANNUAL MEETING ABSTRACTS SCC (0/33) by FISH. No difference of overall survival time was observed between amplified and non-amplified groups in EAC patients. Conclusions: ALK amplification was present in 7% of EAC cases, but ALK rearrangement and expression was not present. No ALK amplification, rearrangement and expression were detected in SCC. ALK amplification is not associated with EAC patients’ overall survival. 785 Mitosis-Specific Marker PHH3 Immunostain Is a More Sensitive and Efficient Method To Evaluate the Mitotic Activity in Gastrointestinal Stromal Tumor (GIST) S Zhu, F Lin, ZE Chen. Geisinger Medical Center, Danville, PA. Background: Mitotic activity is an important prognostic factor in GIST. The accurate identification of mitotic figures on the H&E stained slides could be challenging due to processing artifact, degeneration, apoptosis, or lymphocytic infiltration. Mitosis-specific marker PHH3 was proven as a sensitive and reliable method to assess the mitotic activity in various tumors. The aim of the current study was to compare the PHH3-stained mitotic counts and the time for counting on immunostained slides with the mitotic counts and the time for counting on H&E stained slides in GIST. Design: Immunohistochemical stain for PHH3 and routine H&E staining were performed on 45 cases (41 non-malignant and 4 malignant at the time of sampling) of GIST. The mitotic counts were assessed on both immunostaining slides and H&E stained slides. The PHH3-stained mitotic counts were compared to the mitotic counts on the H&E stained slides. The time to count the mitosis by two methods was recorded too. Results: For 41 non-malignant GIST cases, the mean mitotic count on the H&E stained slides was 1.9 per 50 high power fields and the mean PHH-stained mitotic count was 4.9 (p < 0.001) per 50 high power fields. For 4 malignant GIST cases, the mean mitotic count on the H&E stained slides was 47.5 per 50 high power fields and the mean PHH-stained mitotic count was 90 per 50 high power fields. The mean time to count the mitotic figures for 50 high power fields on the H&E stained slides was 2 minutes and 50 seconds. The mean time to count the PHH3-stained mitotic figures for 50 high power fields was 40 seconds (p < 0.001). Conclusions: PHH3 immunostain is a more sensitive and efficient method to evaluate the mitotic activity in GIST. The malignant GISTs demonstrate significant higher mitotic counts by both PHH3 immunostain and H&E stained slides. 786 Etiology and Histomorphology of Esophagogastric Junction Intramucosal Adenocarcinoma (IMC): In Comparison to Esophageal IMC H Zhu, Z Li, H Xie, TW Rice, LA Rybicki, JR Goldblum, X Liu. Cleveland Clinic Foundation, Cleveland, OH; Second Military Medical University, Shanghai, China. Background: While it is clear Barrett Esophagus (BE) predisposes to esophageal adenocarcinoma, controversy exists regarding the etiology of esophagogastric junction (EGJ) adenocarcinoma. The aim of this study is to evaluate the clinical and pathologic features of EGJ IMC and compare them to IMC of the distal esophagus, as these tumors are smaller than more deeply invasive ones and allow for a better comparison of the surrounding mucosal changes which are often obscured by larger tumors. Design: 149 cases of IMC from an esophagectomy database (1983 – 2010) were identified by reviewing medical charts and slides from these resection specimens. Clinicopathologic features were compared between IMC arising in the esophagus versus EGJ. Results: Of these 149 cases; 54 cases (36.2%) were EGJ and 95 (63.7%) were esophageal IMC. Mean age and gender were not significantly different between these two groups [(61.8 yrs vs. 64.7, p=0.12; 92.6% male vs. 83.2%, p=0.14)]. Hiatal hernia was present in 87.5% and 90.5% of EGJ and esophageal IMC (p=0.51); hiatal hernia length was shorter in patients with EGJ IMC [median: 3 cm vs. 4 cm, p = 0.007)]. Compared with esophageal IMC, intestinal metaplasia (IM) in the esophagus was absent in 8 of 54 EGJ IMC (14.8% vs. 0%, p8.0 (Bioanalyzer 2100, Agilent Technologies, Santa Clara, CA). mRNA (500ng) underwent cDNA synthesis and in vitro transcription using the Ambion WT Expression assay (Ambion Inc, Austin, TX) followed by fragmentation and hybridization on Human Exon 1.0 ST arrays (Affymetrix Corp.,Santa Clara, CA). Washing, staining and scanning of arrays was performed (Fluidics Station 450, Scanner 3000) after hybridization and signal intensity calculated by Microarray Suite version 5.0. Statistical comparisons (ANOVA) were performed (Partek Genomics Suite, St. Louis, MO) with false discovery rate: q value=0.05 and -2.0>fold-change>2.0. Results: 415 CCRCC transcripts were significantly increased (n=191; 2 to 22x) or decreased (n=224; -2 to -92x)). 137 transcripts involved cancer-related genes. Altered tumor suppressor gene (n=14) and oncogene (n=10) transcripts were seen, with a partial list with greatest fold changes including: EGF variant(-17.5x), SCL481(-10x), MAL(-9.5x), GPC3(-7x), PTGS2(-6.5x), RARB(-2.3x) and CA-IX(10x), IGFBP3(6x), C3(4.6x), VEGF-A(4.6x), MDM2(2.5x). HIF-1α/1β, and VHL transcripts were not significantly altered. 21 transcripts were found in 13 different cancer signaling pathways (e.g. pancreas, bladder, ovary, lung (non-small cell and small cell), melanoma, breast, thyroid, other) with a mean+SD (range) of transcripts per pathway of 4.7+2.3 (1-9) (Ingenuity database). Conclusions: CCRCC showed many mRNA transcript alterations, including increases and decreases of important oncogene and tumor suppressor transcripts. A subset of these transcript alterations is shared amongst a diverse group of other organ cancer signaling pathways, suggesting that CCRCC involves dysfunction of multiple growth regulatory pathways. The absence of significant HIF-1 transcript alterations is consistent with the understanding that HIF-1 is dysregulated at the protein level. 809 SIX1 Expression Correlates with Tumor Grade of Urothelial Carcinoma RC Batiste, R Rong, L-P Wang, Z Bing. Hospital of the University of Pennsylvania, Philadelphia, PA. Background: Six1 is a homeobox transcription factor belonging to the evolutionarilyconserved Six1–Eya–Dach transcriptional regulatory network that is critical during embryonic development. Six1 requires interaction with the Eya family of proteins to activate the transcription of genes involved in nephrogenesis, neurogenesis and myogenesis. Mutations in either Six1 or Eya2 can lead to branchio-oto-renal syndrome. Mutations in either gene can disrupt individual protein function, Six1–DNA binding or Six1–Eya2 binding. While both genes are commonly expressed during development, their overexpression has been observed in various neoplastic states. SIX1 is aberrantly expressed in numerous cancers including those arising from breast, ovary, cervix, and liver, as well as Wilms tumor and alveolar rhabdomyosarcoma. Eya2 is required to mediate the pro-metastatic functions of Six1. Design: The following is an immunohistochemical study of Six1 and Eya2 expression in high grade and low grade urothelial carcinoma. In the appropriate setting, the expression of Six1 and Eya2 in invasive carcinoma has also been investigated. Tumor cell nuclear staining intensity is scored in a blinded fashion using the Allred scoring system. Results: Eighteen cases of high grade and 10 cases of low grade urothelial carcinoma were retrieved and stained for Six1 and Eya2. Twelve out of the eighteen cases of high grade urotehlial carcinoma had invasive components. The scores for non-invasive and invasive high grade urothelial carcinoma for Six1 were 6.3 +/- 2.4 and 5.9 +/- 2.3 respectively. There was no statistical difference between these two groups. The average score for low grade urothelial carcinoma for Six1 was 1.9 +/-2.5. There were significant statistical differences between low grade and high grade non-invasive and invasive carcinomas (p
