Plasma AM levels are elevated in PaC, especially with DM. Further studies in a larger cohort of patients are planned to validate AM as a potential biomarker of PaC associated DM.
Risk of Pancreatic Cancer Development in Cancer Survivors in the United States - A Population-Based Epidemiological Study Using the SEER Cancer Database Sathya Jaganmohan, Ramu P. Raju, Manoop S. Bhutani, Gottumukkala S. Raju, John R. Stroehlein, Jeffrey H. Lee Pancreatic cancer (PC) is the 4th most common cancer in the U.S with dismal 1 and 5year survival rates of 24% and 5% respectively. Tumors of the exocrine pancreas (TEP) have an overall survival rate of 4% compared to 48% for tumors of the endocrine pancreas. Knowledge of etiological factors is evolving and screening may be an effective strategy to reduce mortality. Screening for PC is controversial and is being evaluated in patients with strong family history or hereditary cancer syndromes. Objective: To perform a population based study to evaluate subsequent TEP risk in patients with history of cancer compared to the general population. Methods: Using SEER*Stat MP-SIR software (http://seer.cancer.gov/ seerstat; version 6.5.2), we extracted and analyzed data from the original 9 SEER registries, Nov. 2008 submission (1973-2006). We analyzed the standardized incidence ratios (SIRs) for PC in cohorts with each index primary cancer (IPC) using the observed (O) divided by the expected number (E) of cases at 1,5 and 10 years of follow up. The expected PC incidence rates were derived from a reference population. IPCs included in the analysis were lung, breast, oropharyngeal, colorectal, prostate, urogential, endometrial/ovary, lymphoma or leukemia. Nonmalignant tumors, PC diagnosed within 6 months of IPC and any cancer diagnosis based only on death certificate or autopsy reports were excluded. TEP were selected using International classification of Diseases for oncology (ICD- O) histological codes (80008149,8156-8239,8247-9970). Results: In a total of 1,810,487 patients with history of various IPC, 5324 patients developed PC. The 1 year, 5 year and cumulative O/E ratio for patients with history of lung cancer was 1.39, 1.30 and 1.29 respectively (p≤ 0.05). A statistically significant reduction in the cumulative O/E ratios for TEP development was noted for patients with colorectal (0.89), prostate (0.93) and breast (0.91) probably related to aggressive surveillance and improving survival in these patients with treatment. Conclusion: In this population based epidemiological study, a statistically significant increase in standaradized incidence ratio for pancreatic cancer was identified in patients with previous history of lung cancer and a significant decrease seen in patients with colorecta, prostate and breast cancer.
T1384 Smoking Induces Pancreatic Fibrosis in Humans Erwin-Jan M. van Geenen, Mark M. Smits, Tim C. Schreuder, Donald L. van der Peet, Elisabeth Bloemena, Chris J.J. Mulder Introduction Smokers are at risk for pancreatic cancer and other pancreatic diseases. Cigarette smoking also aggravates the risk of pancreatic cancer in patients with hereditary and chronic pancreatitis and results in a higher incidence of acute pancreatitis and relapses in chronic pancreatitis. Both pancreatic cancer and chronic pancreatitis are characterized by a progressive fibrosis. Recently, two studies on rats reported that tobacco smoking is associated with chronic pancreatic inflammation with fibrosis, and scarring of pancreatic acinar structures. In this study we aim to confirm a relation between cigarette smoking and pancreatic fibrosis (PF) in humans. Methods In this retrospective study, pancreatic tissue acquired during autopsy was collected and revised. Pancreatic fibrosis (PF) was gradually scored by analyzing intra-lobular, extra-lobular and total PF: grade 0 (normal or mild; 0-25% PF), grade 1 (moderate; 25-50% PF), grade 2 (severe; >50%). Information on smoking habits was extracted from (electronic) medical files. Results Of 900 autopsies, performed from January 2005 till December 2007, the minority of available histology material (n=111) was of significant quality to be included for analysis. Grade 2-3 total PF and intra-lobular PF was significantly more present in “smokers” versus “never smokers” (total: 42.9% vs 26.5%, p=0,027 and intra-lobular: 39.3% vs 15.6%, p=0.013), whereas no differences could be found between “never smokers” and “abstinent smokers” and “abstinent smokers” and “smokers”. When taking into account interlobular PF, no differences between all groups were observed. Conclusion To date no human study studied the effect of tobacco smoking on pancreatic tissue. We demonstrate for the first time that current cigarette smoking is associated with total pancreatic fibrosis and more specific intra-lobular pancreatic fibrosis, compared to non-smokers.
T1387 Idiopathic Chronic Pancreatitis, Diabetes, and Smoking are Significant Risk Factors for Pancreatic Cancer: Results of Case-Control and Cohort Studies Shallu Midha, Vishnubhatla Sreenivas, Tushar K. Chatterjee, Madhulika Kabra, Yogendra K. Joshi, Pramod K. Garg Background: The putative risk factors for pancreatic cancer include increasing age, smoking, diabetes, and chronic pancreatitis (CP). Previous studies regarding risk factors for pancreatic cancer were inconsistent due to their retrospective nature, no adjustment for confounders, and study of predominantly alcoholic CP. Objective: To identify risk factors for pancreatic cancer. Methods: A case-control study was performed for risk factors for pancreatic cancer that included CP, diabetes, family history of cancer, weight, smoking, alcohol, and dietary factors. The target sample size for the case-control study was 230 patients with pancreatic cancer and 920 healthy controls (1:4 ratio). Odds ratio (OR) was calculated for risk factors by multivariate logistic regression. Population attributable risk was calculated for significant risk factors. A cohort study was also conducted to study CP as a risk factor for pancreatic cancer. The target sample size was 400 patients with CP having >2 years of disease in the cohort study. K-ras mutation was studied in the pancreatic juice/tissue of patients with CP by direct sequencing. Standardized incidence ratio was calculated as a measure of relative risk (RR) from the observed incidence of pancreatic cancer in CP patients and the incidence of pancreatic cancer in the general population. Results: We included 249 pancreatic cancer patients and 1000 healthy controls in the case-control study. Multivariate analysis showed chronic pancreatitis (OR 123.94, 95% CI 15.51-990.09), diabetes (>4 years duration) (OR 2.99, 95% CI 1.67-05.36), smoking (OR 2.08, 95% CI 1.45-2.99), family history of cancer (OR 3.35, 95% CI 1.66-6.78), non-vegetarian diet (OR 1.54, 95% CI 1.09-2.19), and higher intakes of carbohydrates (OR 2.59, 95% CI 1.67-4.02) and proteins (OR 1.77, 95% CI 1.14-2.77) as significant independent risk factors for pancreatic cancer. Of the 249 patients with pancreatic cancer, 24 had underlying idiopathic CP and none had alcoholic pancreatitis. The population attributable risk was 17.76, 13.73, and 10.94 for smoking, diabetes, and CP respectively. We included 402 patients (mean age 29.29 years; 309 males) with CP in the cohort study. During 3879 person-years of follow-up, 5 patients (4 idiopathic CP, 1 hereditary CP) developed pancreatic cancer after a mean interval of 16.60±3.51 yrs from the onset of CP with a RR of 124. K-ras mutation was found only in 2 of 55 CP patients and had no correlation with the development of pancreatic cancer. Conclusion: Idiopathic (not alcoholic) CP, diabetes of >4 years duration, smoking, family history of cancer and non-vegetarian diet were important risk factors for pancreatic cancer.
T1385 Adrenomedullin: A Biomarker of Pancreatic Cancer-Associated Diabetes? Gaurav Aggarwal, George Klee, Martin E. Fernandez-Zapico, Eric W. Klee, Thomas C. Smyrk, William Bamlet, Mariza de Andrade, Gloria M. Petersen, Suresh T. Chari Background: Pancreatic cancer (PaC)-associated new onset Diabetes Mellitus (DM) is likely a paraneoplastic phenomenon caused by tumor secreted products. Since onset of DM precedes diagnosis of PaC by up to 3 years, identification of the diabetogenic secretory products of PaC could be useful for early diagnosis of PaC. We examined the role of adrenomedullin (AM), a 52 amino acid peptide that inhibits insulin secretion and promotes PaC growth, as a biomarker of PaC. Methods: We performed immunohistochemistry (IHC) for AM on 30 resected PaC and 4 histologically normal pancreata. We generated gene expression profiles in PaC cell lines using the Affymetrix Human U133 Plus 2.0 microarray. We tested the ability of supernatant from PaC cell lines to inhibit glucose-induced insulin secretion in a β cell line (INS-1). We measured total plasma AM (Shionogi) in 55 subjects without PaC (28 with normal fasting glucose and 27 with new onset type 2 DM) and 61 matched subjects with resectable PaC (31 with normal FBG and 30 with new onset DM). Results: By IHC, AM expression in normal pancreata was evident only in the periphery of islets. In contrast, all 30 PaCs showed moderate to marked AM expression throughout the cancer and in neighboring islets. Gene expression studies showed marked overexpression of AM in 2 PaC cell lines (PANC1 and SU86.86) in comparison to the control cell line (HPDE). Interestingly, conditioned media from these PaC cell lines impaired glucose stimulated insulin secretion by β cells In Vitro. Mean plasma levels (fmol/l) of AM were higher in patients with PaCDM (22.9+/-10.7) compared to PaC patients with normal FBG (18.3 +/- 7, p=0.057), non-cancer subjects with new onset type 2 DM (14.8 +/- 10.7, p<0.001) and non-cancer subjects with normal FBG (12.9 +/- 6.6, P<0.001). Healthy controls with normal fasting glucose and subjects with new-onset type 2 DM had similar AM levels (p=0.216). At a cutoff of 15.9 fmol/l, AM had a sensitivity of 69% and a specificity of 81% for PaC. Conclusion: AM is overexpressed in PaC. Cell lines with AM overexpression impair glucose signaling in β cells.