Epidemiology and etiology of pancreatic cancer

Abstract

Pancreatic cancer is one of the most devastating malignancies with an extremely high fatality, resulting that its mortality rate almost equals to incidence rate. Although primary prevention is of upmost importance, the underlying etiology of this cancer remains largely unknown. Pancreatic cancer is a heterogenetic disease, and the accumulated genetic alterations play an important role in pancreatic pathogenesis. Recent advances in next-generation sequencing have enabled comprehensive cancer genomic studies. However, clinical pancreatic cancer samples are characterized as having low tumor cellularity, as a result of an abundance of stroma in the tumor microenvironment, and this presents a big challenge for direct genomic sequencing for clinical pancreatic cancer samples. In this thesis, we aimed to enrich our knowledge of the etiology of pancreatic cancer with regard to several infectious agents and poor oral hygiene. Of note, we took the challenge to directly sequence clinical pancreatic cancer samples with a broad range of tumor cellularities, and attempted to depict its variant profile.

In Study I, we retrieved all hepatitis C virus (HCV) and hepatitis B virus (HBV) infection notifications in Sweden from records in a national surveillance database at the Swedish Institutet for Infectious Disease Control (SMI) from 1990 to 2006, and followed them for pancreatic cancer occurrence by the end of 2008. The pancreatic cancer risk in the exposed population was compared with that in a matched reference population. Hazard ratios (HRs) were derived from Cox proportional hazards regression models. The main finding in this study is that the subjects with HCV infection had a 60% increased risk after adjustment for potential confounders. Therefore, the finding implied that HCV infection may be associated with a higher pancreatic cancer risk but further studies are warranted to confirm the observed association. The point estimate in this study also suggested an excessive risk among subjects with HBV infection, however, without statistical significance due to a lack of study power.

In Study II, we took advantage of the population-based prevalence study of oral mucosal lesions conducted in Uppsala County in central Sweden during 1973-74. The study population was followed through linkages with the Swedish population and health registers. A total of 19 924 participants were included in the final analysis, with 126 pancreatic cancer ascertained during an average of 28.7 years of follow-up. Among all tested indicators of poor oral hygiene, we found that fewer teeth at baseline appeared to increase pancreatic cancer risk, although the relative risk estimates were not statistically significant. Among the subjects with more than 10 teeth, subjects with unacceptable dental plaque had a doubled risk of pancreatic cancer compared with those without dental plaque after controlling for potential confounding factors. Subjects with Candida-related or denture-related oral mucosal lesions, or tongue lesions, compared with those without any of the three lesions, showed a 70%, 30% and 80% increased pancreatic cancer risk, respectively.

In Study III, we carried out a nested case-control study within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, including 448 pancreatic cancer cases and their individually matched control subjects. We measured serum antibodies against Helicobacter pylori (H. pylori) and pepsinogens I and II (markers for presence of chronic corpus atrophic gastritis) by enzyme-linked immunosorbent assays. Conditional logistic regression models were used to estimate odds ratios (ORs). Overall, our results demonstrated that pancreatic cancer risk was neither associated with H. pylori seropositivity nor CagA seropositivity. On the other hand, our findings showed that presence of chronic corpus atrophic gastritis was non-significantly associated with an increased pancreatic cancer risk. Although based on small numbers, the association was particularly prominent among individuals seronegative for both H. pylori and CagA (OR=5.66; 95% confidence interval: 1.59, 20.19; p value for interaction <0.01).

In Study IV, we conducted a case-only study that sourced from a population-based case-control study of pancreatic cancer in Stockholm, Sweden. This study included patients with pancreatic ductal adenocarcinoma (PDAC) who underwent resection surgery between 2007 and 2012 (n=73). Patients were followed from diagnosis until death or the end of the study. We used an Anchored Multiplex Polymerase chain reaction (AMP)-based method for profiling variants in a panel of 65 selected genes. Our findings suggested that the AMP-based next-generation sequencing method can detect variants with allelic frequencies as low as 1% given sufficient sequencing depth. KRAS G12 mutations were completely confirmed by Sanger sequencing for high-allele-frequency samples (>5%), and also fully confirmed by allele-specific PCR and digital PCR for low-allele-frequency samples (1%-5%). The results demonstrated that KRAS mutant subtype G12V is related to a worse prognosis in PDAC patients, and transversion variants are more common among smokers.

In conclusion, we found that HCV, as an infectious agent, may be associated with a higher pancreatic cancer risk. Our findings also support the hypothesis that poor oral hygiene plays a key role in the development of pancreatic cancer. On the other hand, we observed a null association between H. pylori infection and pancreatic cancer risk in the western European populations, but a suggested positive association between chronic corpus atrophic gastritis and pancreatic cancer risk based on a small sample size. Further studies are warranted to verify whether severe gastric atrophy contributes to pancreatic carcinogenesis. AMP-based next generation sequencing is a sensitive and accurate method for profiling tumor variants in PDAC. Future studies with larger sample sizes are needed to explore the role of tumor variants in PDAC prognosis and the impact of environmental risk factors on tumor mutational profile.