Epidemiological, Clinical and Pathological Characteristics of Gastric Tumors. The Population Based Specialized Registry in Northern Italy

Background: Gastric cancer (GC) is the 5th most common cancer and the 3rd leading cause of cancer-related deaths worldwide. GC incidence and mortality rates vary widely across different geographical areas. In Italy the province of Cremona is characterized by an high incidence, compared with the national one. For these reason a specialized population based registry was set up. Methods: Up to now the collection encompasses all the GC cases diagnosed in Cremona province from January 1st 2010 until December 31st 2015. The main data sources were represented by the pathological records and the patient clinical charts. Data were collected following AIRTum (Associazione Italiana Registri Tumori) and IARC (International Agency for Research on Cancer) cancer registration recommendations. Territory of Cremona province comprises 3 main public hospitals and 3 private structures. All of them refer to 2 pathology services. All these facilities were involved in the project. A minority of patients usually ask for different surgery and move to neighboring provinces, mainly Milan and Brescia. An official collaboration with these facilities was signed in order to have the access to the data of their patients who were inhabitants of the province of Cremona. A multidisciplinary team of clinicians (oncologists, gastroenterologists, general practitioners, surgeons), pathologists, geneticists and methodologists (epidemiologists, hygienists) was involved in the project. The registration of an incident case usually began with the pathological confirmation of diagnosis of cancer. Subsequently, medical records were obtained. For each case the following variables were registered: personal and familiar data; imaging studies performed; details on surgery and other treatments received; host genetic background and biomolecular characteristic, social and environmental factors. All data were collected, recorded, protected and processed in accordance to AIRTum-IARC (International Agency for Research on Cancer) cancer registration recommendations and national privacy laws. Any age at diagnosis was included. At the time of diagnosis patients had to be inhabitants of the province of Cremona (districts of Crema, Cremona or Casalmaggiore). Only diagnoses of primary gastric neoplasms were considered. Precancerous lesions and relapsed tumor diagnoses were not considered. For the specific purposes of our study, we didn't record cases diagnosed based on Death Certificate Only (DCO)"in situ" tumors were also included. Site of the tumor at diagnosis was stomach or gastro-esophageal junction (GEJ), ref. cod. ICD-X C16, according to the UICC, 7th ed. All gastric malignancies were considered: gastric cancer; lymphoma; sarcoma and GIST. Gastric cancer was classified according to the Lauren's classification system, which distinguishes two main histological types: "intestinal" and "diffuse" . "Mixed" gastric carcinomas composed of intestinal and diffuse components have also been identified. The primary tumor location was divided into three groups in consideration not only of the oncological but also of the surgical approach. Such groups were 1) GEJ-cardia; 2) body-fundus; 3) antrum-pylorus-angulus. The TNM classification was recorded and the corresponding pathological stage was determined according to the 7th edition of the Union for International Cancer Control (UICC) and hereditary cases according to International GC Linkage Consortium guidelines. Evaluation of the infection of Helicobacter pylori (HP) was performed by immunohistochemistry (IHC) in health gastric mucosa using the GIEMSA stain method. The HER-2 oncogene overexpression was evaluated in tumour gastric mucosa by the IHC method Dako Hercept TestTM. Results were confirmed by Fluorescent in Situ Hybridization (FISH) when IHC positivity score was 2. To individuate Hereditary Diffuse Gastric Cancer (HDGC) cases, criteria by International GC Linkage Consortium 2010 guidelines were followed. Incidence, standard errors, and 95% confidence intervals (CI) were calculated according to the International Agency for Research on Cancer (IARC) general guidelines. As for CIs, the method that considered the approximation to the Poisson distribution was chosen. Age-specific incidence rates were stratified into 18 subgroups by 5-year age interval (0 - 4, to 85+ years). Raw age-specific (5-years classes) and age-standardised (age-adjusted) rates were calculated per 100,000 inhabitants. For raw rates the denominator was the resident population census at 31st December of each year, available online at the official site of Cremona province. An age-standardized rate (ASR) is a weighted average of the age-specific (raw) rate, where the weight is the proportion of individuals in the corresponding age group of a standard population. Age-standardised rates were calculated using the standard age-structure of the European (EU) and World (W) standard population. This corrects the potential confounding effect derived from the differences in age between different populations. The trend of incidence, expressed as "Annual Percent Change" (APC) was evaluated using Join Point (National Cancer Institute, Bethesda, MD) in order to identify significant changes. Overall survival analysis was carried out by Kaplan-Meier methods and statistically significant differences were evaluated by Log-Rank Test . Survival hazard ratio was evaluated by Cox regression model, subsequently the testing of the proportion of hazard. Cumulative risks and other variables, including age at diagnosis, anatomical subsite, morphology of the tumour, TNM stage at diagnosis, presence of HP infection, HER-2 amplification status, 5-year survival and 1-year mortality rate were reported too. Descriptive statistics were used to summarize the data and parametric and non-parametric tests were used to evaluate differences between groups. Statistical analysis was carried out by STATA 13 software package (Texas, USA). A p-value less than 0.05 vas considered as statistically significant.