Search

Background: Pancreatic cancer is a rapidly fatal disease with very few 5-year survivors even after aggressive surgical treatment. Our objective was to determine the actual 5-year survival rate of patients with pancreatic adenocarcinoma who underwent a resection with curative intent in 5 teaching hospitals within the University of Toronto system. We then sought to determine clinical and histopathologic features of 5-year survivors to determine factors associated with a favorable prognosis., Study Design: A retrospective chart review was performed using surgeon and hospital databases to identify patients who had a surgical resection for pancreatic adenocarcinoma between January 1, 1988, and December 31, 1996., Results: One hundred twenty-three patients who had a resection and a pathologic diagnosis of pancreatic adenocarcinoma with complete followup were identified from seven surgical practices. Mean survival (+/- standard error) in this series was 31.7 +/- 3.5 months (median 13.6 months). There were 18 5-year survivors (14.6%), including 5 patients (4.1%) who survived longer than 10 years. The survivors included 13 patients who had undergone a Whipple resection, 4 who had undergone a distal pancreatectomy, and 1 who had undergone a total pancreatectomy. Tumor size, lack of jaundice at presentation, negative nodal disease, low tumor grade, and a low tumor stage were all significant predictors of survival in univariate analysis (all p < 0.05). Only tumor stage (hazard ratio [95% confidence interval]: stage IIA 1.5 [0.8 to 2.8], stage IIB 2.6 [1.4 to 4.7], stage III 1.8 [0.8 to 4.3]) and tumor grade (hazard ratio [95% confidence interval]: moderately differentiated 1.6 [0.9 to 2.9], and poorly differentiated 3.1 [1.6 to 6.2]) were independently associated with survival differences in a multivariate Cox proportional hazards model., Conclusions: We conclude that longterm survival from pancreatic adenocarcinoma is possible if the disease is identified in its early stages. These and other similar data should provide further stimulus for the development and evaluation of novel screening strategies to improve early detection of this disease.

ST7 is a putative tumor suppressor gene on chromosome 7q31. However, the role of ST7 as a tumor suppressor is uncertain as somatic mutations have been difficult to demonstrate. In order to investigate the genetic role of RAY1/ST7 in tumorigenesis, we have screened 135 colorectal cancers for loss of heterozygosity (LOH) at chromosome 7q31. The entire RAY1/ST7 gene, including intron/exon boundaries and alternate 5' and 3' sequences of 15/124 (12%) informative cancers with LOH were characterized. No somatic mutations of the RAY1/ST7 gene were observed. Our results do not support a role for RAY1/ST7 as a colorectal cancer tumor suppressor gene.

Inactivation of deoxyribonucleic acid (DNA) mismatch repair genes, most commonly human mutL homologue 1 (hMLH1) or human mutS homologue 2 (hMSH2), is a recently described alternate pathway in cancer development and progression. The resulting genetic instability is characterized by widespread somatic mutations in tumor DNA, and is termed high-frequency microsatellite instability (MSI-H). Although described in a variety of tumors, mismatch repair deficiency has been studied predominantly in colorectal carcinoma. Most MSI-H colorectal carcinomas are sporadic, but some occur in patients with hereditary nonpolyposis colorectal cancer (HNPCC), and are associated with germline mutations in mismatch repair genes. Until now, the identification of MSI-H cancers has required molecular testing. To evaluate the role of immunohistochemistry as a new screening tool for mismatch repair-deficient neoplasms, the authors studied the expression of hMLH1 and hMSH2, using commercially available monoclonal antibodies, in 72 formalin-fixed, paraffin-embedded tumors that had been tested previously for microsatellite instability. They compared immunohistochemical patterns of 38 MSI-H neoplasms, including 16 cases from HNPCC patients with known germline mutations in hMLH1 or hMSH2, with 34 neoplasms that did not show microsatellite instability. Thirty-seven of 38 MSI-H neoplasms were predicted to have a mismatch repair gene defect, as demonstrated by the absence of hMLH1 and/or hMSH2 expression. This included correspondence with all 16 cases with germline mutations. All 34 microsatellite-stable cancers had intact staining with both antibodies. These findings clearly demonstrate that immunohistochemistry can discriminate accurately between MSI-H and microsatellite-stable tumors, providing a practical new technique with important clinical and research applications.

Some colorectal tumors with wild-type adenomatous polyposis coli gene have activating mutations in beta-catenin (encoded by CTNNB1) that result in decreased phosphorylation by GSK-3beta and increased signaling through the Tcf/Lef transcription factors. To investigate the relationship between CTNNB1 mutations and underlying pathways of genomic instability, we examined 80 colorectal cancers stratified by the presence or absence of microsatellite instability (MSI). CTNNB1 mutations were identified in 13 (25%) of 53 cancers with high frequency MSI (MSI-H), including 12 point mutations at exon 3 phosphorylation sites (codons 41 and 45) and one deletion of the entire exon 3 degradation box. No CTNNB1 mutations were identified in 27 microsatellite stable or low frequency MSI (MSI-L) colorectal cancers (P < 0.01). In contrast, CTNNB1 mutations were identified in 3 of 9 (33%) MSI-H and 10 of 20 (50%) MSS/MSI-L endometrial carcinomas, suggesting a more generalized involvement in these tumors. Only six (46%) of the endometrial carcinoma CTNNB1 mutations occurred at residues directly phosphorylated by GSK-3beta, and only one of these was at either codon 41 or 45. All point mutations in MSI-H cancers were transitions, whereas 64% of those in MSS/MSI-L cancers were transversions (P < 0.01). The differences in the mutation profiles suggest that there may be molecular fingerprints of CTNNB1 mutations, determined by biological factors related to both tumor type and underlying pathways of genomic instability.