The epidemiology of cancer is the study of the factors affecting cancer, as a way to infer possible trends and causes. The study of cancer epidemiology uses epidemiological methods to find the cause of cancer and to identify and develop improved treatments.
This area of study must contend with problems of lead time bias and length time bias. Lead time bias is the concept that early diagnosis may artificially inflate the survival statistics of a cancer, without really improving the natural history of the disease. Length bias is the concept that slower growing, more indolent tumors are more likely to be diagnosed by screening tests, but improvements in diagnosing more cases of indolent cancer may not translate into better patient outcomes after the implementation of screening programs. A similar epidemiological concern is overdiagnosis, the tendency of screening tests to diagnose diseases that may not actually impact the patient's longevity. This problem especially applies to prostate cancer and PSA screening.[2]
Some cancer researchers have argued that negative cancer clinical trials lack sufficient statistical power to discover a benefit to treatment. This may be due to fewer patients enrolled in the study than originally planned.[3]
Contents |
State and regional cancer registries are organizations that abstract clinical data about cancer from patient medical records. These institutions provide information to state and national public health groups to help track trends in cancer diagnosis and treatment. One of the largest and most important cancer registries is Surveillance Epidemiology and End Results (SEER), administered by the US Federal government.[4] Health information privacy concerns have led to the restricted use of cancer registry data in the United States Department of Veterans Affairs[5][6][7] and other institutions.[8]
Observational epidemiological studies that show associations between risk factors and specific cancers mostly serve to generate hypotheses about potential interventions that could reduce cancer incidence or morbidity. Randomized controlled trials then test whether hypotheses generated by epidemiological studies and laboratory research actually result in reduced cancer incidence and mortality. In many cases, findings from observational epidemiological studies are not confirmed by randomized controlled trials.
The most significant risk factor is age. According to cancer researcher Robert A. Weinberg, "If we lived long enough, sooner or later we all would get cancer."[9] Essentially all of the increase in cancer rates between prehistoric times and people who died in England between 1901 and 1905 is due to increased lifespans.[9] Since then, some other factors, especially the increased use of tobacco, have further raised the rates.[9]
Over a third of cancer deaths worldwide are due to potentially modifiable risk factors. The leading modifiable risk factors worldwide are:
Men with cancer are twice as likely as women to have a modifiable risk factor for their disease.[10]
Other lifestyle and environmental factors known to affect cancer risk (either beneficially or detrimentally) include the use of exogenous hormones (e.g., hormone replacement therapy causes breast cancer), exposure to ionizing radiation and ultraviolet radiation, and certain occupational and chemical exposures.
Every year, at least 200,000 people die worldwide from cancer related to their workplace.[11] Millions of workers run the risk of developing cancers such as pleural and peritoneal mesothelioma from inhaling asbestos fibers, or leukemia from exposure to benzene at their workplaces.[11] Currently, most cancer deaths caused by occupational risk factors occur in the developed world.[11] It is estimated that approximately 20,000 cancer deaths and 40,000 new cases of cancer each year in the U.S. are attributable to occupation.[12]
In the United States, cancer is responsible for 25% of all deaths with 30% of these from lung cancer. The most commonly occurring cancer in men is prostate cancer (about 25% of new cases) and in women is breast cancer (also about 25%). Cancer can occur in children and adolescents, but it is uncommon (about 150 cases per million in the U.S.), with leukemia the most common.[13] In the first year of life the incidence is about 230 cases per million in the U.S., with the most common being neuroblastoma.[14] Data from 2004-2008 in the United States indicates that the overall age-adjusted incidence of cancer was approximately 460 per 100,000 men and women per year.[15]
In the U.S. cancer is second only to cardiovascular disease as the leading cause of death;[13] in the UK it is the leading cause of death.[16] In many Third World countries cancer incidence (insofar as this can be measured) appears much lower, most likely because of the higher death rates due to infectious disease or injury. With the increased control over malaria and tuberculosis in some Third World countries, incidence of cancer is expected to rise; in the Eastern Mediterranean region, for example, cancer incidence is expected to increase by 100% to 180% in the next 15 years due to increases in life expectancy, an increasing proportion of elderly people, and the successful control of childhood disease.[17] This is termed the epidemiologic transition in epidemiological terminology.
Cancer epidemiology closely mirrors risk factor spread in various countries. Hepatocellular carcinoma (liver cancer) is rare in the West but is the main cancer in China and neighbouring countries, most likely due to the endemic presence of hepatitis B and aflatoxin in that population. Similarly, with tobacco smoking becoming more common in various Third World countries, lung cancer incidence has increased in a parallel fashion.
Cancer is responsible for about 25% of all deaths in the U.S., and is a major public health problem in many parts of the world. The statistics below are estimates for the U.S. in 2008, and may vary substantially in other countries. They exclude basal and squamous cell skin cancers, and carcinoma in situ in locations other than the urinary bladder.[13] As seen, breast/prostate cancer, lung cancer and colorectal cancer are responsible for approximately half of cancer incidence. The same applies for cancer mortality, but with lung cancer replacing breast/prostate cancer as the main cause.
| Male | Female | |||
|---|---|---|---|---|
| most common (by occurrence)[13] | most common (by mortality)[13] | most common (by occurrence)[13] | most common (by mortality)[13] | |
| prostate cancer (25%) | lung cancer (31%) | breast cancer (26%) | lung cancer (26%) | |
| lung cancer (15%) | prostate cancer (10%) | lung cancer (14%) | breast cancer (15%) | |
| colorectal cancer (10%) | colorectal cancer (8%) | colorectal cancer (10%) | colorectal cancer (9%) | |
| bladder cancer (7%) | pancreatic cancer (6%) | endometrial cancer (7%) | pancreatic cancer (6%) | |
| non-Hodgkin lymphoma (5%) | liver & intrahepatic bile duct (4%) | non-Hodgkin lymphoma (4%) | ovarian cancer (6%) | |
| skin melanoma (5%) | leukemia (4%) | thyroid cancer (4%) | non-Hodgkin lymphoma (3%) | |
| kidney cancer (4%) | esophageal cancer (4%) | Skin melanoma (4%) | leukemia (3%) | |
| oral and pharyngeal cancer (3%) | bladder cancer (3%) | ovarian cancer (3%) | uterine cancer (3%) | |
| leukemia (3%) | non-Hodgkin lymphoma (3%) | kidney cancer (3%) | liver & intrahepatic bile duct (2%) | |
| pancreatic cancer (3%) | kidney cancer (3%) | leukemia (3%) | brain and other nervous system (2%) | |
| other (20%) | other (24%) | other (22%) | other (25%) | |
In the developed world, one in three people will develop cancer during their lifetimes. If all cancer patients survived and cancer occurred randomly, the normal lifetime odds of developing a second primary cancer (not the first cancer spreading to a new site) would be one in nine.[18] However, cancer survivors have an increased risk of developing a second primary cancer, and the odds are about two in nine.[18] About half of these second primaries can be attributed to the normal one-in-nine risk associated with random chance.[18] The increased risk is believed to be primarily due to the same risk factors that produced the first cancer, such as the person's genetic profile, alcohol and tobacco use, obesity, and environmental exposures, and partly due, in some cases, to the treatment for the first cancer, which might have included mutagenic chemotherapeutic drugs or radiation.[18] Cancer survivors may also be more likely to comply with recommended screening, and thus may be more likely than average to detect cancers.[18]
Cancer can also occur in young children and adolescents, but it is rare (about 150 cases per million yearly in the US). Leukemia (usually acute lymphoblastic leukemia) is the most common cancer in children aged 1–14 in the U.S., followed by the central nervous system cancers, neuroblastoma, Wilms' tumor, and non-Hodgkin's lymphoma.[13] Statistics from the SEER program of the US NCI demonstrate that childhood cancers increased 19% between 1975 and 1990, mainly due to an increased incidence in acute leukemia. Since 1990, incidence rates have decreased.[19]
The age of peak incidence of cancer in children occurs during the first year of life, in infants. The average annual incidence in the United States, 1975–1995, was 233 per million infants.[19] Several estimates of incidence exist. According to SEER,[19] in the United States:
Teratoma (a germ cell tumor) often is cited as the most common tumor in this age group, but most teratomas are surgically removed while still benign, hence not necessarily cancer. Prior to the widespread routine use of prenatal ultrasound examinations, the incidence of sacrococcygeal teratomas diagnosed at birth was 25 to 29 per million births.
Female and male infants have essentially the same overall cancer incidence rates, a notable difference compared to older children.
White infants have higher cancer rates than black infants. Leukemias accounted for a substantial proportion of this difference: the average annual rate for white infants (48.7 per million) was 66% higher than for black infants (29.4 per million).[19]
Relative survival for infants is very good for neuroblastoma, Wilms' tumor and retinoblastoma, and fairly good (80%) for leukemia, but not for most other types of cancer.