Cancer epidemiology and Cancer registry Lausanne-Part 1-2010 Cancer epidemiology and Cancer registry C. Sauvaget MD, PhD Scientist, Screening Group (SCR)
Definition of Epidemiology Lausanne-Part 1-2010 Definition of Epidemiology Study of the distribution and determinants of diseases in human populations Epidemiological studies have been categorized into 2 major groups: Focus on distribution: descriptive epidemiology – burden, pattern Focus on determinants: analytical epidemiology – causes, risk factors Application: prevention, control
Descriptive Epidemiology Lausanne-Part 1-2010 Descriptive Epidemiology Can be conducted with or without the knowledge of exposure Pure descriptive epidemiology—no exposure data Present rates by calendar year, cohort year of birth stratified by age, gender, race, disease subtype, geographic regions Analytical ecological study-with exposure data Correlation or regression at aggregate level
Example of ecological study Lausanne-Part 1-2010 Example of ecological study Estimated daily consumption of fat per caput in 1964-66 and age-adjusted breast cancer incidence rates in women aged 35-64 years in 1972-77 in 24 countries (reproduced, by permission of Oxford University Press, from Armstrong & Mann, 1988-5). Dos Santos I. Cancer epidemiology, 1999, IARC publications
Role of Descriptive Epidemiology-1a Lausanne-Part 1-2010 Role of Descriptive Epidemiology-1a Can provide tremendous insights into the aetiology – Different racial groups living in same area with wide differences in disease rate -- possibility of genetic susceptibility Testicular cancer (Whites: 6/100,000; Blacks: 0.7/100,000) Same racial group living in different areas with similar disease rate -- possibility of genetic susceptibility Testicular cancer (Blacks: 0.6-1.0/100,000)
Role of Descriptive Epidemiology-1b Lausanne-Part 1-2010 Role of Descriptive Epidemiology-1b Can provide tremendous insights into the aetiology Same racial group living in different areas with wide differences in disease rate — possibility of environmental exposures Different occupational groups have major differences in disease rate — importance of occupational exposures Different religious groups having different rates — importance of lifestyle factors Chinese Han, and other minorities
Example of environmental factor Lausanne-Part 1-2010 Example of environmental factor Age-adjusted incidence rates of female breast cancer for USA residents (1972-85) by birthplace and age at migration, and for Jordan (1973-81) (reproduced, by permission of Churchill Livingston, from Shimizu et al., 1991) Dos Santos I. Cancer epidemiology, 1999, IARC publications
Example of occupational factor Lausanne-Part 1-2010 Example of occupational factor Men occupation class and risk of incident cancer (1990-2002, n= 14 853) Melchior et al. Cancer causes and control 2005;16:515-524
Role of Descriptive Epidemiology-2 Lausanne-Part 1-2010 Role of Descriptive Epidemiology-2 Disease incidence trends over time Point to an altered influence of risk factors in the population: Increase exposure to UVR due to: »Increase outdoor activities »Change in cloth patterns »Increase use of suntan lamps »Increase depletion of ozone layer Increase in skin cancer
Time trend of cigarettes sales and lung cancer mortality Lausanne-Part 1-2010 Number of cigarettes per adult Annual Age-std mortality rate from lung cancer Kubik et al. Cancer 1995
Lausanne-Part 1-2010 Annual age-adjusted incidence rates of cancers at selected primary sites in Finland: actual rates from 1953 to 1979 and predictions up to the year 2000 based on a statistical model which included age, period and cohort effects (reproduced with permission from Läärä, 1982) Dos Santos I. Cancer epidemiology, 1999, IARC publications
Lausanne-Part 1-2010 Age-standardised incidence of invasive cervical cancer and screening coverage rate: England, 1975-2002 National call-recall introduced Incidence of invasive cervical cancer Screening coverage rate Incidence rate/1000 000 Percentage Sources: Cancer Research UK; Quinn et al., 1999; Willoughby et al., 2006
Proportion of cancers attribution to various risk factors Lausanne-Part 1-2010 Proportion of cancers attribution to various risk factors Risk factors Best estimate Range Tobacco 30 25-40 Alcohol 3 2-4 Hazardous diet, physical inactivity 35 10-70 Medicines & medical procedures 1 0.5-3 Infection* 10 1-? Reproductive factors 7 1-13 Occupation 4 2-8 Pollution 2 1-5 Geophysical factors * HBV, HCV, HPV, EBV, Helicobacter Pylori, HIV, schistosomiasis… Doll and Peto. JNCI 1981,66:1191-1308
Effect of Hepatitis B vaccination on incidence of liver cancer in Taiwan Lausanne-Part 1-2010 Chang et al. N Eng J Med 1997
Measures of occurrence of disease Lausanne-Part 1-2010 Measures of occurrence of disease Denominator Time period Incidence Prevalence Risk
Measures of occurrence of disease Lausanne-Part 1-2010 Measures of occurrence of disease Population at risk: The part of a population which is susceptible to a disease Incidence: The number of new health events occurring in a defined population during a specified period of time New cases /population at risk /time Used to measure current disease activity Allows comparison between areas with different populations Ex. 100 subjects are followed for 1 year and 20 developed the disease => the incidence rate is 20 cases /100 person-years of observation Prevalence: The proportion of individuals in a population at risk with a disease at a single point in time
Relationship between Incidence and Prevalence Lausanne-Part 1-2010 Prevalence is the number of new and existing cases divided by the total population (can be during a period of time or at a given point) Prevalence = (New cases + existing cases) / Total population Can be expressed as a percent Can give a picture of disease burden within a population Prevalence rate = incidence rate x average duration of disease Incidence Duration Prevalence Dead, cured, missed, etc.
Using routine data to measure disease occurrence Lausanne-Part 1-2010 Using routine data to measure disease occurrence Dos Santos I. Cancer epidemiology, 1999, IARC publications
Lausanne-Part 1-2010 Exposure and outcome-1 Definition Measurement
Lausanne-Part 1-2010 Exposure and outcome-1 * Does alcohol intake increase the risk of lung cancer? Alcohol Lung cancer (exposure) (outcome) * Does hepatitis B vaccination protect against liver cancer? Hepatitis B vaccine liver cancer
Lausanne-Part 1-2010 Exposure and outcome-2 * Does alcohol intake increase the risk of lung cancer? Alcohol Lung cancer (exposure) (outcome) Smoking (confounder)
Exposure Nature Dose Time Source: Questionnaire Records Diaries Lausanne-Part 1-2010 Exposure Nature Dose Time Source: Questionnaire Self-administered Personal interview Records Diaries Biological measurement
Outcome Hospital records Cancer registration Death registry Lausanne-Part 1-2010 Outcome Hospital records Cancer registration Death registry
Cancer registries http://ci5.iarc.fr/ http://globocan.iarc.fr/ Lausanne-Part 1-2010 Cancer registries http://ci5.iarc.fr/ http://globocan.iarc.fr/
Cancer Registry: Registration Lausanne-Part 1-2010 Cancer Registry: Registration Cancer Registry The office or institution which is responsible for the collection, storage, analysis and interpretation of data on persons with cancer. Cancer registration The process of continuing systematic collection of data on the occurrence, characteristics, and outcome of reportable neoplasm with the purpose of helping to assess (prevent) and control the impact of malignant disease in the community.
Cancer Registry 1. Population-based Cancer Registry Lausanne-Part 1-2010 Cancer Registry 1. Population-based Cancer Registry All cases in a DEFINED population are registered True (unbiased) profile of cancer in the community incidence, stage distribution, survival, etc. Calculation of incidence rates (because population at risk is quantified) The main interest is for epidemiology and public health
Cancer Registry 2. Hospital-based Cancer Registry Lausanne-Part 1-2010 Cancer Registry 2. Hospital-based Cancer Registry Records all cases of cancer treated in a given hospital The population from which the cases come is not defined The main interest is clinical care hospital administration
Basic information required and variable definitions Lausanne-Part 1-2010 Basic information required and variable definitions Item no. Item The person The tumour Personal identification Incidence date Name Most valid basis of diagnosis Sex Date of birth or age Topography (site) Demographic Morphology (histology) Address Behavior Ethnic group Source of information
Most valid basis of diagnosis of cancer Lausanne-Part 1-2010 Most valid basis of diagnosis of cancer Microscopic 1. Autopsy with concurrent or previous histology 2. Histology of primary 3. Histology of metastasis 4. Cytology or haematology Non-microscopic 5. Specific biochemical and/or immunological tests 6. Exploratory surgery 7. Clinical investigation (X-ray, US…) 8. Clinical only 9. Death Certificate Only (DCO) ICD-Oncology
Incidence and survival data Lausanne-Part 1-2010 Incidence and survival data Provided by cancer registries thru the International Association of Cancer Registries (IACR) Mostly regional (except Nordic countries) Not always recent (generally 5-year delay): request time to be compiled and published. Detailed information (site, histology, laterality, grade, stage)
Lausanne-Part 1-2010 Coverage of cancer registration worldwide % of the population covered (around 2000) 40.0 99.0 19.0 7.1 7.9 When considering data of good quality (included in the latest volume for the Cancer Incidence in Five Continents series), these percentages are even lower: only 8% of the world is covered by “good, reliable” cancer registries. 13.0 82.0 16.5% total
Number of deaths from cancer Lausanne-Part 1-2010 Mortality data Number of deaths from cancer National level Provided by the WHO Recent and available for long time periods (1950 to 2008) Limited number of cancers Quality can be poor (under-reporting, incomplete coverage, high percentage of ill-defined causes of deaths)
Mortality data (WHO databank) Lausanne-Part 1-2010 Mortality data (WHO databank) % of the population covered (around 2005) 98.0 100.0 100 9.0 13.8* 95.0 The coverage is better for mortality. However, the detail and the quality of the data (accuracy of the recorded cause of death and the completeness of registration vary substantially). Note that there is quite no mortality information for Africa except in South Africa and Egypt (13% of total African pop) 76.5 33% total *Egypt and South-African Republic
Lausanne-Part 1-2010 Estimated age-standardised incidence rate per 100,000 Lung: male, all ages 1,095,186 cases in 2008 GLOBOCAN 2008 Ferlay J, Shin HR, Bray F, Forman D, Mathers C and Parkin DM. GLOBOCAN 2008, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10 [Internet]. Lyon, France: International Agency for Research on Cancer; 2010. Available from: http://globocan.iarc.fr
Estimated age-standardised incidence rate per 100,000 Breast: all ages Lausanne-Part 1-2010 Estimated age-standardised incidence rate per 100,000 Breast: all ages 1,383,523 cases in 2008 GLOBOCAN 2008 Ferlay J, Shin HR, Bray F, Forman D, Mathers C and Parkin DM. GLOBOCAN 2008, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10 [Internet]. Lyon, France: International Agency for Research on Cancer; 2010. Available from: http://globocan.iarc.fr
Population (2008): 640 million population Less developed countries Lausanne-Part 1-2010 Population (2008): 640 million population Males 319 million Population 3.7 million cases 2.7 million deaths Females 321 million population 3.5 million cases 2.1 million deaths Lung Breast Stomach Cervix uteri Liver Colorectum Oesophagus Prostate Corpus uteri Ovary Bladder Leukaemia Lip, oral cavity Incidence Mortality (Thousands) GLOBOCAN 2008
Lausanne-Part 1-2010
Cancer epidemiology studies Lausanne-Part 1-2010 Cancer epidemiology studies Descriptive studies Observational studies Experimental studies
Prevalence estimate Prevalence at time t1 = 2/10 = 0.20 = 20% Lausanne-Part 1-2010 Prevalence estimate Prevalence at time t1 = 2/10 = 0.20 = 20% Prevalence at time t2 = 3/8 = 0.38 = 38% Incidence case Dos Santos I. Cancer epidemiology, 1999, IARC publications
Person-years and incidence rate Lausanne-Part 1-2010 Person-years and incidence rate Consider a hypothetical group of nine persons who were followed up from the beginning of 1980 to the end of 1984. Subjects joined the study at different points, as shown in Figure. Three subjects, (2), (6) and (7), developed the disease of interest during the study period and one, (4), was last contacted at the end of 1983. Incidence rate = 3/32 = 0.094 per person-year or 9.4/100,000 P-Yrs Dos Santos I. Cancer epidemiology, 1999, IARC publications
Main types of epidemiological studies Lausanne-Part 1-2010 Intervention (experimental) studies Clinical trials Filed trials Individual level Aggregated level (community trials) Observational (non-experimental) studies Cohort studies Case-control studies Cross-sectional surveys Routine-data-based studies Aggregated level (ecological studies)
Randomized Control Trials Lausanne-Part 1-2010 Randomized Control Trials Intervention outcome+ Research cohort outcome- outcome+ Randomization outcome- No intervention Research direction: What will happen in the future?
Prospective Cohort Study Lausanne-Part 1-2010 Prospective Cohort Study Exposed Disease Cohort No disease Disease No disease Not exposed Present Future Time Research direction: What will happen in the future?
Case-control Study Case Control past present Time Lausanne-Part 1-2010 Interview Review of medical records Review of biological exams How many people were exposed? Case How many people were not exposed? How many people were exposed? Control Selection of controls is one of the most difficult problems in epid. How many people were not exposed? past present Time Research direction: What happened in the past?
Cancer Registry Types 1. Population-based cancer registry (PBCR) Lausanne-Part 1-2010 Cancer Registry Types 1. Population-based cancer registry (PBCR) 2. Hospital-based cancer registry 3. Pathology registry
Cancer Registry 3. Pathology Tumor Registry Lausanne-Part 1-2010 Cancer Registry 3. Pathology Tumor Registry Collects information from one or more laboratories on histologically diagnosed cancers The population from which the tumour tissue has come is not defined The information - has high diagnostic quality - but is difficult to generalize
HOSPITAL CANCER REGISTRY Additional variables Lausanne-Part 1-2010 HOSPITAL CANCER REGISTRY Additional variables Contact details Admission + Discharge dates Hospital referred from , to Physicians (treating, following) Diagnostic procedures Extent of disease Treatment details (first, subsequent) Outcome (recurrence) Follow - up
Accessible thru the Internet or using a Windows-based PC software Lausanne-Part 1-2010 The results are presented for 170 countries of the world, plus build-in areas (six WHO regions, more and less developed countries and the world) Data available for 27 major cancers, for men and women, and for 5 age groups: 0- 14,15-44,45-54,55-64,65+ Accessible thru the Internet or using a Windows-based PC software http://globocan.iarc.fr/ If there are no data, the country specific rates are those of the corresponding region (calculated from the other countries for which estimates could be made).
Cancer – worldwide burden Lausanne-Part 1-2010 Cancer – worldwide burden 7 million Deaths 11 million New Cases 25 million Living with Cancer
New cancer cases and deaths, World 2002 (10.8 million cases/6.7 million deaths) Lausanne-Part 1-2010 Males 5.8 million cases 3.8 million deaths Females 5.0 million cases 2.9 million deaths Lung (1.35/1.18) Breast (1.15/0.41) Colon/Rectum (1.0/0.53) Stomach (0.93/0.70) Prostate (0.68/0.22) Liver (0.63/0.60) Cervix uteri (0.49/0.28) Oesophagus (0.41/0.34) Bladder (0.35/0.14) Non-Hodgkin lymphoma (0.30/0.16) Leukaemia (0.28/0.20) Oral cavity (0.27/0.12) Pancreas (0.23/0.22) Kidney (0.21/0.11) Ovary (0.20/0.12) Incidence Mortality 1000 800 600 400 200 0 200 400 600 800 1000 (Thousands)