Swedish Oesophago-Gastric Study

Swedish OEsophageal and GAstric tumour Study (SOEGAS)

Background Globally, oesophageal and gastric tumours are common with a high mortality. Oesophageal cancer and gastric cancer combined are often labelled upper gastrointestinal (GI) cancer. These tumours share many characteristics, including anatomy, risk factors, treatment and prognosis.

The main known risk factors for adenocarcinoma of the oesophagus and gastro-oesophageal junction are gastro-oesophageal reflux disease and obesity. The main risk factors for the other main histological type of oesophageal cancer, squamous cell carcinoma, are tobacco smoking and alcohol. The strongest known risk factor for gastric cancer is infection with Helicobacter pylori. Shared risk factors include dietary factors, tobacco smoking and heredity.

The curative treatment of upper GI cancers builds heavily on demanding surgery, where most of the oesophagus or stomach is resected. Except for surgery, most tumours are also treated with pre-operative, and sometimes postoperative, chemotherapy or chemoradiotherapy (6). The palliative treatment is also very important since the majority of patients are not eligible for curative treatment.

The survival is poor in patients diagnosed with upper GI cancer. This is mainly explained by late diagnosis since the symptoms are usually late and aggressive tumour biology, with early spread and recurrence. If detected early, the prognosis is excellent, but only a fraction of patients is detected at early tumour stages. The overall 5-year survival in oesophageal cancer is less than 20% and in gastric cancer it is less than 30% in most populations, including Sweden. The rate of patients found eligible for curative treatment including surgery is only 20-35%. Among operated patients the postoperative 5-year survival is 30-50%. It is crucial to conduct research that can improve the lives also of palliative patients.

It is of great importance to identify factors that can prevent upper GI cancer. Means to facilitate the detection at a premalignant or early and curable tumour stage would improve the prognosis and survivorship. It is also highly relevant to identify factors that can improve the curative and palliative treatment.

Resources Opportunities Sweden offers internationally unique possibilities to link high quality registry data with clinical data from medical records, which make up a fantastic resource for valid research in aetiology, prevention, detection, and treatment of upper GI cancer. The personal identity numbers assigned to all Swedish residents enables individual data to be linked.

Experience Our research group has a long track record in addressing risk factors, preventive measures and treatment of upper GI cancer. The research group contains expertise in clinics, epidemiology and biostatistics.

Overarching aims

The overarching aim of this research project is to reduce death and suffering from upper GI cancer. This aim can be accomplished by a broad research approach that aims to identify:

1. Risk factors and prevention Preventive actions can decrease the incidence of these tumours.
2. Early detection Early detection provides great opportunities for local treatment (endoscopic resection). Detection at an early invasive stage before metastases have occurred also provides good chances of cure, but requires more extensive surgery.
3. Improved treatment The mainstay treatment for cure in upper GI cancer is surgical resection. The research assessing how the surgery can become optimised is limited.

Specific aims

1. Evaluate the recent incidence trends These analyses will separately assess differences in incidence over time across various patient and tumour characteristics, including:

• Sex
• Age
• Tumour stage
• Tumour histology
• Tumour site within the oesophagus or stomach

1. Identify risk factors

• Medications
• Lifestyle factors
• Reproductive factors
• Perinatal factors
• Other diseases and co-morbidities
• Surgical procedures for other indications than upper GI tumours
• Hereditary factors
• Socio-economic factors
• Dental and oral health
• Other factors

1. Identify factors that facilitate early detection

• Lifestyle factors
• Other diseases and co-morbidities
• Hereditary factors
• Socio-economic factors

1. Evaluate trends in prognosis Treatment Curative Surgery Neo-adjuvant therapy Palliative Patient characteristics Sex Age BMI Socio-economic factors Tumour characteristics Stage Histology Site within the oesophagus or stomach
2. Identify factors that can improve prognosis and survivorship

• Tumour characteristics
• Surgical factors
• Neoadjuvant therapy
• Socio-economic factors
• Lifestyle factors
• Other diseases and co-morbidities
• Previous surgical procedures
• Reproductive factors
• Medications
• Dental and oral health
• Other factors

Methods Study cohort The cohort is entitled the Swedish OEsophageal and GAstric tumour Study (acronym SOEGAS), and will include patients diagnosed with oesophageal or gastric tumour from year 2000 onwards. The selection of the cohort will be based on a diagnosis code representing oesophageal or gastric tumours recorded in: Swedish Cancer Registry, Swedish Patient Registry or Cause of Death Registry.

Data collection

This cohort needs to be linked with additional data for us to be able to evaluate the specific research aims listed above. The resources and data are needed for all cohort members:

1. The Swedish Patient Registry ('Patientregistret') since 1987 when this registry became nationwide complete. This registry is managed by the National Board of Health and Welfare (SoS by Swedish acronym). These data are needed for evaluation of surgical procedures, re-operations, and all comorbidities.
2. The Swedish Cancer Registry ('Cancerregistret') since 1958 when this registry was initiated. This registry is managed by SoS. Detailed data are needed for the oesophageal or gastric tumours, including date of diagnosis, exact tumour site, histological type of tumour, and tumour stage. Data are also needed on previous cancers before the start date of the cohort (year 2000) as well as any concurrent or later (secondary) tumours.
3. The Swedish Prescribed Drug Registry ('Läkemedelsregistret') since July 2005 when this registry started. This registry is managed by SoS. We need these data to assess medications as risk factors for these tumours, to assess how medications influence the outcomes in patients diagnosed with these tumours, and also to assess comorbidities.
4. The Swedish Dental Health Registry ('Tandhälsoregistret') since this registry was started in 2008. This registry is held by SoS. We need these data to assess whether dental and oral health influences the risk of developing oesophageal and gastric tumours as well as the prognosis and survivorship in these patients.
5. Swedish Medical Birth Registry ('Medicinska födelseregistret') since it started in 1973. This registry is held by SoS, and from this registry we can assess data on paternity and other key variables in women who have given birth, e.g. BMI. We also need data on the children, e.g. gestational age, small for gestational age, large for gestational age, preterm birth.
6. The Swedish Causes of Death Registry ('Dödsorsaksregistret') since 1st January 2000. This registry is managed by SoS, and is required to assess date of death and causes of death. These data are needed for the prognostic studies and also for censoring of patients no longer at risk of various outcomes under study.
7. The Swedish Multi-Generation Registry ('Flergenerationsregistret') since its start. This registry is held by SCB, and these data are needed to assess the role of heredity in the aetiology of these tumours.
8. The Swedish Longitudinal integration database for health insurance and labour market studies (LISA by Swedish acronym) ('Longitudinell integrationsdatabas för sjukförsäkrings- och arbetsmarknadsstudier'). This registry is held by SCB and is needed for data on socio-economic factors, including highest level of education, cohabitation, civil status, income, country of birth and parental countries of birth, latest year of immigration and emigration, and place of residence.
9. The Swedish Registry of the Total Population ('Registret över totalbefolkningen') since 1968. These data are needed to calculate the expected risk of oesophageal and gastric tumours in the corresponding background population. This registry is also needed for updated information on civil status, divorce, emigration, immigration, cohabitation and updated information about date of death.
10. The Swedish Defence Recruitment Agency ('Mönstringsregistret') since its initiation. We need data on BMI and physical capacity from this registry.
11. Medical records. We need data from the hospitals regarding clinical factors including tumour stage, treatment, risk factors, and various treatment factors. This means that we need to have personal identity numbers for all cohort members to be able to find the relevant medical records.
12. Total population data. Finally, we need data from the Swedish Registry of the Total Population ('Registret över totalbefolkningen') and the Swedish Cancer Registry from the entire Swedish population from year 2000 onwards to calculate the expected incidence and mortality rates in the corresponding background population.

Personal identity numbers and key codes We need the personal identity numbers of the selected cohort members to retrieve the medical records from the healthcare services. We also need the National Board of Health and Welfare ('Socialstyrelsen') to keep the key codes to all cohort members in order for us to ask for up-dates of the cohort regarding potential needs for additional time, follow-up, cohort size and further data.

Statistical analysis The statistical analysis will be done by one of our senior biostatisticians employed in the group. All data management and analyses will be done according to laws and regulations. There will be various statistical methods used and these need to be tailored for each individual study based on the cohort. However, studies evaluating risk factors will typically be analysed using multivariable logistic regression, providing odds ratios and 95% confidence intervals adjusted for potential confounding factors. Studies evaluating mortality as the outcome will be analysed using multivariable Cox regression, providing hazard ratios and 95% confidence intervals adjusted for potential confounding factors.

Statistical power The cohort will be powerful enough to study the main research questions. We cannot expand the study further while maintaining its high internal validity. It is already nationwide complete and the study period from 2000 onwards mirrors a modern era of treatment. Thus, we cannot increase the sample size.

Specific study protocols The study protocol for each study included in this project will be specified with detailed study protocols before initiated. This project plan does not allow detailed presentation of all future studies based on the data collection, but it gives an overview of the topic.

Significance

SOEGAS has all prerequisites to contribute substantially with new knowledge that might reduce deaths and suffering in oesophageal and gastric tumours on a global scale. All of the studies included in the project will be based on Swedish data and we have good reasons to believe that outputs from the project will have a global influence as well. The quality and the contents of the Swedish registries might have changed over calendar periods but from 2000 onwards only limited changes have been made, which improves the quality of the research and facilitates interpretations and reduces various sources of bias. Taken together, SOEGAS might bring extremely valuable information to an area of research which is clearly understudied which is the aetiology, prevention and treatment of tumours of the oesophagus and stomach.