19: Occupational Cancer

The majority of cancers are multifactorial in etiology, the result of a combination of genetic and nongenetic factors. Genetic factors alone are estimated to cause only about 5% of cancers. Nongenetic factors, sometimes referred to as environmental factors, account for the majority of cancers. They include lifestyle factors such as tobacco use, alcohol consumption, poor diet, obesity, physical inactivity, and occupational and consumer exposures to myriad chemicals and product formulations, which collectively contribute to the occurrence of a substantial proportion of cancers. Millions of US workers are exposed to substances that are known to cause cancer in humans, with 125 documented chemical/exposure circumstances, and more that cause cancer in animal studies. Unfortunately, however, less than 5% of chemicals manufactured or processed in the United Sates have been tested for carcinogenicity in animal bioassays. Based on associations between occupational exposures and cancer, it is estimated that 4–10% of US cancers are caused by occupational exposures.

The identification of occupational carcinogens is important in part because most occupational cancers are completely preventable with appropriate exposure controls, personnel practices, and strict protective legislation. Various agencies have classified chemicals and other agents as to their potential carcinogenicity to humans and animals, including the International Agency for Research on Cancer (IARC), an independent scientific institution within the World Health Organization (WHO) (see http://www.iarc.fr/). Further classifications of carcinogenic agents may be found on the Web site of the U.S. National Toxicology Program (http://ntp.niehs.nih.gov/) and in its most recent report on carcinogens (12th Report on Carcinogens).

Evidence suggests that cancers arise from a single abnormal cell. The initial stage in development of the abnormal cell appears to result from an alteration or mutation in the genetic material, deoxyribonucleic acid (DNA). This alteration may occur spontaneously or may be caused by exogenous factors, such as exposure to carcinogenic chemicals or radiation. Whether a tumor develops from this altered cell may depend on a variety of factors, such as the ability of the cell to repair the damage, the presence of other endogenous or exogenous agents that foster or inhibit tumor development, and the integrity of the immune system.

Stages in Tumor Development

A variety of evidence indicates that cells undergo multiple heritable changes in the process of becoming a “cancer cell”; this process is termed carcinogenesis. Early animal studies investigating the etiology of cancer cell growth hypothesized that tumor development involved at least two distinct stages: initiation and promotion. A classic example of this process is the mouse skin tumor model. In this model, a small dose of a carcinogen, known as the initiator (in this case, typically a polycyclic aromatic hydrocarbon [PAH]) is applied to the skin. Although large doses of PAHs alone readily induced skin tumors, the smaller doses alone did not. However, application of a promoter, such as croton oil, following application of the initiator did result in tumor development. Interestingly, application of the promoter alone or prior to administration of the initiator did not result in skin tumors. A similar process has been implicated in the development of tumors in other organs, such as mouse liver and lung and rat trachea. Clearly, there are limitations to the application of animal models of tumorigenesis to the etiology of common human tumors. However, these data have provided a framework to understand basic toxin-induced carcinogenesis (Figure 19–1).

From a functional point of view, it now appears helpful to view carcinogenesis as a multistep process including initiation, promotion, and progression. Initiation is thought to result from an irreversible change in the genetic material (DNA) of the cell arising from interaction with a carcinogen that is a necessary, but not sufficient, condition for tumor development. It is this somatic mutation that sets the stage for tumor development and is the basis for the somatic mutation theory of carcinogenesis.

Promotion consists of those processes subsequent to initiation that facilitate tumor development, presumably by stimulating proliferation of the altered cell. The mechanisms of promotion, sometimes referred to as epigenetic mechanisms (as opposed to the genotoxic or mutational effects of initiators), are incompletely understood. Promotion classically does not result from binding to and alteration of DNA but may result from production of or suppression of proteins that alter the way that DNA is transcribed. In the case of the mouse skin tumor model, croton oil appears to interact with membrane receptors to affect cellular growth and differentiation. Promotion typically yields a benign tumor or group of preneoplastic cells that do not have the ability to invade stroma or metastasize; progression then creates those additional heritable changes necessary for the development of a malignant tumor.

This process is referred to as the multistep or multistage model of carcinogenesis. Most common adult cancers, such as colorectal cancer, are the results of multiple genetic and protein mutational changes. While it is clearly a simplistic categorization, carcinogens are often divided into initiating agents, or genotoxic (DNA-reactive) “early stage” carcinogens, and promoting agents, or epigenetic “late stage” carcinogens. Table 19–1 lists the distinguishing features of initiating and promoting agents. Some agents (eg, cigarette smoke or asbestos) that seem to possess both initiating and promoting properties are termed complete carcinogens. However, it is also clear that damage from cigarette smoking may set the stage for a multiplicative relationship to carcinogenesis with exposure to carcinogens such as asbestos or nickel. Given the complexity of the multistage model and increasing experimental evidence that tends to blur the distinction between these categories, the value of categorizing specific agents probably is limited. In addition, owing to the time from exposure to an initiating agent to the subsequent development of visible cancer, identification of possible causal agents or contributing factors may be extremely difficult.

The mechanism by which carcinogen-induced alteration in DNA leads to initiation and ultimately to tumor development is related at least in part to mutations in protooncogenes and tumor-suppressor genes. Protooncogenes contain DNA sequences that, when altered by a mutational event into an oncogene, stimulate transformation and proliferation of an altered potentially neoplastic cell. There are a number of protooncogenes in human and animal cells that are responsible for normal cellular differentiation and maturation. In contrast, tumor-suppressor genes function as negative regulators of cell growth. A genetic change in one or more tumor-suppressor genes, which results in inactivation of a specific gene, may allow unregulated growth of the altered cell. The observation in experimental animals that tumors develop only after activation of one or more oncogenes and inactivation of one or more tumor-suppressor genes provides a mechanistic explanation for the multistep model of carcinogenesis.

There are a number of mechanisms that result in genetic alterations, including point mutations, chromosome translocation or rearrangement, gene amplification, and the induction of numerical chromosome changes (aneuploidy), each of which theoretically could be induced by a chemical exposure. For example, point mutations in ras protooncogenes (so called because they were identified originally in rat sarcomas) have been observed in human and rodent tumors. The ras gene codes for a protein product known as p21, differing by only one amino acid from the normal protein. This protein may function as a direct cell-transforming agent, conferring malignant potential to the cell.

For most toxic effects, the persistence or progression of damage requires the continued presence of the offending chemical agent. For cancer initiators, however, relatively short-term exposure in humans may, based on the results of animal studies, induce genetic damage in the cell sufficient to ultimately result in tumor development years after exposure has ceased. An example of this is the development of mesothelioma more than 20 years after relatively brief exposures to asbestos, seen after as little as one day in both animals and humans.

Induction-Latency Period

In both experimental animal models of cancer and human cancers with known causes, a significant interval of time is required from first exposure to the responsible agent to the development of malignancy. This interval is referred to as the induction-latency (or sometimes just latency) or incubation period. The requirement for multiple heritable changes in the cell may be at least partly responsible for prolonged latency intervals.

For humans, the length of the induction-latency period may be as short as 2 years for radiation-induced leukemia (as in survivors of atomic bomb explosions) to as long as 40 or more years for some cases of asbestos-induced mesothelioma. For most solid tumors, however, the latency interval is approximately 12–25 years. Obviously, this long period of time may obscure the relationship between a remote exposure and a newly found tumor.

Many toxic agents result in known adverse effects only when the exposure is above a certain threshold dose or duration. If this threshold dose is not exceeded, there are no demonstrable consequences to the health of the animal or human. With carcinogens, it is much more difficult to determine if such a threshold exists. Theoretically, if no threshold exists, there is no dose (other than zero) at which the incremental risk of cancer is zero.

There are differing scientific opinions as to the existence of threshold doses for carcinogenic agents, with accompanying arguments on both sides. Regulators generally accept the nonthreshold model. Given that a single alteration (mutation) in DNA in one cell may set the stage for tumor development, it is theoretically possible that exposure of the cell to only one molecule of a carcinogen ultimately could lead to tumor formation. Although the probability of tumor formation tends to increase with increasing frequency and magnitude of exposure to the carcinogen, a single small exposure might be sufficient. For example, in the mouse skin tumor model described earlier, a single high-level exposure to a PAH has been shown to be capable of inducing a tumor.

On the other hand, there are counterarguments that support the concept of a threshold dose for carcinogens. Although exposure to a single molecule of a particular offending agent might induce a tumorigenic change in a cell, the likelihood that the molecule will reach its target cell is lowered when exposure to smaller doses occurs. Mechanisms that might potentially result in a nonzero threshold dose include the “first pass” effect, DNA repair processes, immunologic surveillance, and, in some cases, the observed phenomenon that preceding cellular damage from high doses of a toxin is required to induce subsequent cancers. If the carcinogen is subject to rapid metabolic deactivation, as in a “first pass” effect in the liver after ingestion, the ability of a small or single dose to contact the susceptible cell would be reduced. DNA repair mechanisms (eg, excision of altered DNA nucleotides) may allow repair of an induced mutation before a clone of tumor cells develops. In this regard, some researchers have noted that most of the large number of spontaneous mutations in human and animal cells induced by endogenous oxidants, such as superoxide anion (O₂⁻) and hydrogen peroxide (H₂O₂), are “corrected” by effective DNA repair mechanisms. To the extent that most of the mutations caused by exogenous mutagens, occurring with lower frequency, are repaired, their impact on carcinogenesis would be minimal. Furthermore, immunologic mechanisms may be capable of destroying transformed cells before a tumor develops. Finally, some cancers are induced by prior tissue damage. For example, alcohol and probably some chlorinated hydrocarbon solvents can induce liver inflammation with subsequent cirrhosis, following which there is a higher risk of liver tumor development.

If any of these phenomena pertain for a given carcinogen, there may be a threshold dose below which no carcinogenic effect will occur. These mechanisms may fail to take into account the concept of individual susceptibility, for example, in an individual who is immunocompromised or has genetic defects in DNA repair. As an example, there is substantial variability in the chronic dose of alcohol required to induce hepatic damage and thus carcinogenesis in individuals, which appears to be, at least in part, a result of differences in alcohol metabolism. In practice, it is essentially impossible to prove or disprove the notion that a nonzero threshold exists, because of limitations in the sensitivity of available scientific methodologies (ie, epidemiologic or animal experimental studies) to detect an effect at very low doses. Although public health authorities typically evaluate risks with the assumption that there are no demonstrable thresholds for carcinogens, they also recognize that the theoretical increment in risk at low doses may be negligible, especially in comparison to background risks for the cancer unrelated to exposure. For example, the U.S. Environmental Protection Agency, through its Integrated Risk Information System (IRIS) has, through a quantitative risk assessment process, calculated that the approximate increment in risk for skin cancer over background risks due to low concentrations of arsenic in drinking water would be, as an example, 1 in 1,000,000 for a concentration of 0.02 μg/L (or 0.02 ppb) over a lifetime of exposure. If one considers that skin cancer is the most common type of cancer in the United States, occurring in approximately one in five Americans over the course of a lifetime, the theoretical risk associated with ingestion of this low concentration of arsenic in drinking water is so low that it would be viewed as having little real health significance. Needing to make a judgment as to the importance of these calculated risks, many agencies, such as the U.S. EPA, set an allowable level of exposure at lifetime doses that theoretically would increase risk by an “acceptable” level, usually stated to be at 1 in 10,000 to 1 in 1,000,000.

Dose-Response Relationships

Although thresholds may or may not exist, there is strong evidence for a dose-response effect for carcinogens that have been studied adequately, as exists for most items that cause human diseases of all kinds. In other words, exposure to larger doses of a specific agent results in a higher risk of developing cancer than do smaller doses. Both animal and human epidemiologic studies support this concept.

Unfortunately, the observed data points defining the dose-response relationship from these studies tend to come from relatively high doses of exposure. The shape of the dose-response curves at lower doses, in the range of typical human exposures, must be extrapolated based on assumptions about the unobservable “behavior” of the curve at low doses. At these low doses, the curve might theoretically be linear, concave (sublinear), or convex (supralinear) and may or may not have a threshold. These uncertainties lead to an inability to definitively predict the impact of low doses from the effects observed at much higher doses. Absent information about the effects at low doses, the conventional public health view is that there is linearity at low doses.

Evidence to support the carcinogenicity of a chemical for humans may be derived from three basic types of studies: human epidemiologic studies, experimental studies in animals, and mechanistic studies. Epidemiologic studies, especially cohort studies, provide the strongest evidence for human carcinogenicity; they evaluate effects on human subjects and usually involve a large number of individuals. However, such studies are subject to a number of limitations, including potential types of epidemiologic bias, confounding factors, and inadequate power, which may limit their ability to detect and confirm carcinogenic effects. Well-conducted animal bioassays can provide strong support for carcinogenicity in the animal species tested. However, the findings for humans are less clear than with epidemiologic studies, although a positive animal study should raise potential concern for humans. Other potentially relevant classes of data with a lesser role in assessing causation are termed by IARC as mechanistic and other relevant data, which derive from different types of studies, such as toxicokinetic data and data on mechanisms of carcinogenesis. Mechanistic data include data regarding changes at the molecular level, typically evaluated through short-term in vitro assays, and data regarding structure-activity relationships. These data may provide evidence for a plausible mechanism by which an agent may increase the risk for cancer, but they cannot alone provide causal evidence for carcinogenicity. Positive results of short-term assays (eg, for bacterial mutation) raise the possibility that an agent may be a carcinogen. Structural similarity of a chemical to a known carcinogen would also raise the possibility that it could be a carcinogen. Although the presence of such similarity has, in fact, predicted the carcinogenicity of some previously untested compounds, the overall utility of structural analysis in predicting carcinogenicity has not been established. The sources for information regarding the results of all of these types of studies may include IARC monographs, NTP reports, U.S. EPA publications (eg, in the Integrated Risk Information System [IRIS] database), and comprehensive review publications, including meta-analyses, the latter having significant limitations.

Epidemiologic Studies

Evidence for causality of an association is usually derived from analytic epidemiologic studies or, in other words, cohort or case-control studies, although case reports and descriptive epidemiologic studies may provide suggestive evidence for a possible association that merits further study. In some cases, such as asbestos causing mesothelioma or vinyl chloride causing angiosarcoma, these may be sufficient. Well-conducted cancer epidemiologic studies with positive results provide strong evidence in support of carcinogenicity. Epidemiologic studies are difficult to conduct and often are not feasible because of a number of factors, including high cost, the need to study a large number of individuals, and the requirement for long periods of observation.

Various authorities, beginning with Sir Austin Bradford Hill in 1965 and more recently including the IARC, have established criteria for evaluating causation from epidemiologic studies and assessing the degree of evidence supporting the designation of an agent as a carcinogen. These criteria are used to help decide whether a positive association in epidemiologic studies likely indicates a causal relationship or, alternatively, whether it is likely to be due to chance, bias, or confounding. Not all criteria must be met. The most important or useful criteria are:

1. The strength of an association is the magnitude of the relative risk in the exposed group compared with that in the control group. Strong statistically significant associations are more likely to be causal, because it is less likely that chance, biases (such as recall bias), or confounding (eg, the effects of differences in smoking rates) would account for the observed association.

2. Consistency of an association is the extent to which it is reported from multiple studies conducted in different populations and using different study methods. Consistency can be evaluated quantitatively in some cases using a meta-analysis, a technique for combining the findings from multiple independent studies

3. The biologic gradient of an association is the degree to which it exhibits a dose-response relationship (ie, the observation that higher doses result in a higher frequency of adverse effects).

4. The biologic plausibility of a study is based on the assessment that it makes sense in light of what is known about the mechanism of production of the adverse effect.

5. A study's temporality rests on the conclusion or observation that the cause (ie, exposure) preceded the effect in time.

Despite the utility of these criteria in making assessments of causality, only temporality is absolutely required. A weak association that does not exhibit a dose-response effect and does not (as yet) fully make sense nevertheless may be causal. Furthermore, fulfillment of some of the criteria may occur when the association is, in fact, a result of chance or bias. Statistical significance alone does not prove causation. Thus, one must carefully consider potential alternative explanations for observed associations and the likely impact of chance, bias, or confounding factors in studies. If most of these criteria are met, the likelihood that an association is causal is greater. In practice, development of scientific consensus as to whether an agent is likely to be a human carcinogen may take many years, as evidence accumulates from multiple epidemiologic studies, animal cancer bioassays, in vitro assays, and mechanistic studies. One cannot properly conclude that the association between exposure to an agent and the development of a cancer is a causal one, based upon the finding of a statistically significant association in a single study.

IARC, in the most recent Preamble to its Monograph series, has defined categories of evidence regarding carcinogenicity in humans:

• Sufficient evidence of carcinogenicity. The Working Group considers that a causal relationship has been established between exposure to the agent and human cancer. That is, a positive relationship has been observed between the exposure and cancer in studies in which chance, bias, and confounding could be ruled out with reasonable confidence. A statement that there is sufficient evidence is followed by a separate sentence that identifies the target organ(s) or tissue(s) where an increased risk of cancer was observed in humans. Identification of a specific target organ or tissue does not preclude the possibility that the agent may cause cancer at other sites.

• Limited evidence of carcinogenicity. A positive association has been observed between exposure to the agent and cancer for which a causal interpretation is considered by the Working Group to be credible, but chance, bias, or confounding could not be ruled out with reasonable confidence.

• Inadequate evidence of carcinogenicity. The available studies are of insufficient quality, consistency, or statistical power to permit a conclusion regarding the presence or absence of a causal association between exposure and cancer, or no data on cancer in humans are available.

• Evidence suggesting lack of carcinogenicity. There are several adequate studies covering the full range of levels of exposure that humans are known to encounter, which are mutually consistent in not showing a positive association between exposure to the agent and any studied cancer at any observed level of exposure. The results from these studies alone or combined should have narrow confidence intervals with an upper limit close to the null value (eg, a relative risk of 1.0). Bias and confounding should be ruled out with reasonable confidence, and the studies should have an adequate length of follow-up. A conclusion of evidence suggesting lack of carcinogenicity is inevitably limited to the cancer sites, conditions, and levels of exposure, and length of observation covered by the available studies. In addition, the possibility of a very small risk at the levels of exposure studied can never be excluded.

Limitations of Epidemiologic Studies

Historical outbreaks of occupational cancer have often occurred with exposures at particularly high levels, where toxic effects or cellular proliferation might occur. To the extent that toxicity or cellular proliferation occur at high but not at the low doses to which humans generally are exposed at present, predictions of cancer risk at present day doses based upon results of historical epidemiology studies may not be accurate, because of the lack of clarity as to the shape of the dose-response curve at low doses.

Failure to demonstrate a positive association in an epidemiologic study does not always indicate that there is no association between the agent and the effect studied. In some cases, a false-negative epidemiologic study may result because of a variety of shortcomings. Some of these limitations include difficulties in identifying exposures and effects, difficulties in choosing appropriate study (exposed) and control populations, inadequate duration of follow-up given long induction-latency periods, and the relative lack of sensitivity of epidemiologic methods.

The existence of all these limitations accounts for the consensus among scientists that negative epidemiologic studies do not absolute provide proof of noncarcinogenicity of an agent. Such negative data generally are outweighed if there is a finding of convincingly positive results in animal experimental studies. Greater credence may be given to a negative study if the subjects studied had a sufficiently long period of exposure (an average of 15 or more years), if they were followed long enough to observe an effect (25 or more years), and if the number of exposed subjects was large enough so that a small excess risk for a particular cancer could be detected.

Animal Bioassays

Design

Experimental studies in animals involve the administration of a test chemical to a group of animals followed by observation for the development of tumors. Procedures for such studies are now to some extent standardized and accepted by most of the sponsoring or evaluating institutions, for example, the National Toxicology Program (NTP) and IARC. In brief, protocols include at least 100 or more animals of both sexes with two species in each of two to four dosage groups and incorporate thorough pathologic examination and proper statistical analysis of results. Animals most often are dosed for 2 years, a period that approaches their expected lifetimes, at the maximal tolerated dose (MTD), one-half of the MTD, and more recently at one lower dose (usually one-quarter of the MTD).

Interpretation

Results from well-conducted animal bioassays can yield clear evidence to support the carcinogenicity of a compound in a particular animal species and can be, with caution, extrapolated to humans. IARC, in the most recent Preamble to its Monograph series, has defined categories of evidence regarding carcinogenicity in experimental animals similar to those for humans, as noted above.

Correlation With Human Effects

Results from animal bioassays have served as good predictors of human carcinogenicity. A large number of agents now known to be human carcinogens were first discovered to be carcinogenic in animals. Available evidence suggests that there is good correlation between animal and human results. As IARC stated in its preamble:

There has been some concern that animal studies might not have 100% sensitivity to predict human carcinogenicity, that is, that animal studies might not demonstrate carcinogenicity, when ultimately epidemiologic studies indicate that the agent is carcinogenic to humans. In this regard, it had not been possible to demonstrate carcinogenicity of arsenic in animals until recently, even though it is a known skin, lung, and bladder carcinogen and a probable liver and kidney carcinogen in humans. However, subsequent animal bioassays have provided sufficient evidence in experimental animals for the carcinogenicity of inorganic arsenic compounds. Thus, all known human carcinogens for which adequate animal studies have been conducted have shown sufficient evidence of carcinogenicity in tested animals.

The issue of the specificity of animal testing as a predictor of human carcinogenicity is more difficult to resolve. Because of limitations inherent in epidemiologic studies (discussed earlier), it is unlikely that clear-cut evidence for human carcinogenicity can be derived for many chemicals proven to be carcinogenic in animals. Nevertheless, for the limited number of compounds for which there are adequate data in both humans and animals, there are no substances proven to be carcinogenic in animals that have been proven to be noncarcinogenic in humans.

Although there appears to be a good qualitative correlation between animal and human carcinogenicity, the target site at which cancers develop may be quite different for rodents and humans. For example, benzidine produces liver tumors in rats, hamsters, and mice but produces bladder tumors in humans and dogs. Nevertheless, for all known human carcinogens, at least one site of cancer in humans matched a site in at least one animal species tested.

Agents often are administered to animals in bioassays by a route, such as ingestion or intratracheal instillation, which is different from the one typically experienced by humans. As a result, it is possible that the outcome in the animal will not correctly predict the effect in humans. However, there is little evidence to suggest that a significant discrepancy in outcome occurs solely based on differences in the route of administration, except perhaps with nonphysiologic routes, such as intraperitoneal injection or subcutaneous implantation.

Another issue in the use of animal bioassays is the degree to which the susceptibility of humans to carcinogenic effects parallels the dose-response patterns in animals. There is no database that allows comparison of sensitivity between species. For the limited numbers of substances for which there are quantitative data in both humans and animals, it appears that the sensitivity of humans, on a total dose per body weight basis, is roughly similar to that of animals.

Finally, there is some controversy about the proper classification of an agent as a potential carcinogen in the rare circumstance where experimental results indicate an increased frequency of benign tumors only. However, both NTP and IARC consider that benign tumors induced in animals should raise the suspicion that the agent is potentially carcinogenic.

Limitations of Animal Bioassays

There are a number of difficult issues in the analysis of animal experimental studies that may limit their utility in human cancer risk assessment. Because of the relatively small number of animals studied, the bioassays are relatively insensitive. The agent under study must cause at least a 15% increase in the incidence of tumors in order for a statistically significant excess of tumors to be detected in a bioassay of standard size. A lower excess risk will not be demonstrable, particularly if there is any background rate of tumor development in untreated animals.

To increase the sensitivity of animal experiments, high-dose levels approaching the MTD are chosen. Depending on the agent studied, these dose levels typically are much higher than the human exposure levels in occupational or environmental settings. In some cases, an increased incidence of types of tumors not commonly observed in humans is found. Risk quantification based on animal studies conducted at high doses is difficult because predictions are based on extrapolation to the lower-dose exposures experienced by humans. The assumptions required and the associated mathematical models necessary for extrapolation make it essentially impossible to confirm the accuracy of quantitative risk assessments derived from them, and they are not universally accepted by scientists. Typically, health conservative assumptions are made in this process. These problems add uncertainty regarding the accuracy of risk estimates at low doses.

False positives may occur because certain toxic effects as a mechanism for cancer induction are largely observed in rodents but not in humans, such as peroxisome proliferation leading to liver tumor formation. If a carcinogenic metabolite is produced only in the animal or only in humans, either false-negative or false-positive results could be obtained. Genetic differences metabolic processes in humans may result in differences, in individuals, in susceptibility to some carcinogens. These differences may not be apparent in animal bioassays using genetically similar strains and small numbers of rodents, potentially leading to false negatives. For many chemicals, such differences in pharmacodynamics and pharmacokinetics between humans and animals are not always understood, which may impact the relevance of animal results to humans.

Mechanistic Studies

Short-Term Test-Types and Uses

A number of assays have been designed that provide evidence of mutagenicity or the ability to induce chromosomal damage by chemicals, without the long period of observation or follow-up required for epidemiologic studies or animal bioassays. These short-term tests therefore are much quicker and less expensive to perform. Assessed endpoints include gene mutation, induction of DNA damage and repair, DNA binding, chromosomal aberrations, sister chromatid exchange, and neoplastic transformation of mammalian cells and other endpoints. In general, these tests rely on the fact that most carcinogens, that is, initiators, covalently bind to DNA and thereby induce DNA damage.

One of the best-studied and most commonly performed short-term tests is the Ames test, which uses a mutant strain of Salmonella typhimurium that is deficient in the enzymes required to synthesize histidine and which will not grow unless histidine is added to the growth medium. The chemical to be tested, along with a liver microsomal enzyme fraction from rodents or humans that can metabolically activate “procarcinogens,” is added to the bacterial culture. Bacterial colonies that subsequently grow and can be counted indicate the occurrence of a reversion mutation to the wild strain, reflecting the mutagenic activity of the agent studied. Similar mutagenic testing is possible in cultured mammalian cells in vitro.

Tests for DNA repair can demonstrate that DNA damage has occurred following exposure to a chemical. Chromosomal aberrations in mammalian cells are detected by cytogenetic tests that assess changes in the morphologic structure of chromosomes. Such tests can be performed on animal or human cells, including human lymphocytes. Morphologic changes that may occur include chromosomal translocations and the formation of micronuclei.

Testing for sister chromatid exchange (SCE) is a more sophisticated form of cytogenetic investigation based on differential staining of sister chromatids, allowing for detection of the interchange of genetic material between chromatids. SCEs are more subtle than gross structural chromosomal aberrations. Tests for SCEs may be performed on animal or human cells.

Tests for neoplastic transformation of mammalian cells in culture assess the ability of chemicals to induce neoplastic growth. The treated cells are sometimes injected into animals to assess the ability of the cells to form tumors. A number of other short-term tests are being or have been developed.

Some of these tests (eg, tests for SCEs and chromosomal aberrations) can be performed on cells (typically lymphocytes) taken from humans exposed occupationally or environmentally to suspected carcinogens. Testing for SCEs has been performed on workers exposed to ethylene oxide and other chemicals. An increased frequency of SCEs has been found in some of these workers. These tests can be considered to be biomarkers of the effective dose of a potential carcinogen. Although this merits further study, the clinical significance of these findings for the workers is unknown. SCEs are also increased by other factors, including cigarette smoking. At this time, it is not possible to assess cancer risk based on results of this type of testing in workers. Consequently, use of these tests should be limited to well-designed prospective research studies.

Interpretation & Limitations

The predictive value of short-term tests to assess the potential animal or human carcinogenicity of a chemical is limited. The correlation between results in short-term assays and human or animal studies is imperfect. The sensitivity and specificity of short-term tests for predicting carcinogenicity is limited reflecting the occurrence of both false-negative and false-positive results, respectively. Moreover, while most genotoxic carcinogens are positive in some short-term tests, these in vitro tests generally do not detect chemicals that induce cancers by nongenotoxic or epigenetic mechanisms (ie, they are not sensitive to the effects of promoting agents). There are examples too of substances, such as asbestos, that are negative on an Ames test but are clearly carcinongenic.

Given the current state of knowledge, most authorities state that positive results in short-term tests on previously untested materials warrant further study in animal bioassays and further scrutiny in human-exposure situations. Similarly, positive results on these tests provide corroboration for positive findings in animal bioassays or epidemiologic studies, particularly when the animal or human results provide only limited or suggestive evidence of carcinogenicity. On the other hand, positive short-term assay results alone cannot demonstrate that an agent is a carcinogen.

The Role of Molecular Biology & Biomarkers in the Study of Occupational Cancer

Advances in biochemistry and molecular biology have opened the door to new types of studies that may advance an understanding of occupational cancers. These approaches include pharmacogenomic studies, evaluation of mutations in oncogenes or tumor suppressor genes, and measurement of DNA or protein adducts. The term molecular epidemiology has been applied to this use of molecular biology methodology in conjunction with epidemiologic studies. Application of some of these approaches may allow assessment of exposure to carcinogens and identification of possible early health effects. Methods of assessing internal dose through measurement of urinary metabolites of potential carcinogens can provide complementary information regarding absorbed dose. Measurement of adducts and urinary metabolites, along with the evaluation of chromosomal changes discussed above, may be useful in assessing the internal or biologically effective dose of a carcinogen. However, this would apply only to those individuals studied, and not necessarily a whole population.

Pharmacogenomics is the study of the impact of genetic variability on the response to or the metabolism of different drugs or toxins. Understanding these variations may lead to the ability to detect differences in individual susceptibility to chemically induced cancers and help to understand the mechanisms and etiology of occupational cancer. Most carcinogenic chemicals are metabolized to the active carcinogenic metabolite by phase I enzymes, primarily multiple enzymes in the cytochrome P450 monooxygenase class. Variations in the activity of these enzymes (polymorphisms), which may be genetically or environmentally determined, can result in differences in susceptibility to chemical carcinogenesis. For example, an isoform of one enzyme, aryl hydrocarbon hydroxylase, encoded by the gene CYP1A1, has been associated with an increased risk for lung cancer in smokers. This enzyme catalyzes the oxidation of PAHs to reactive metabolites such as epoxides. Genetically determined variants in phase II enzymes, which are involved in conjugation reactions such as glutathione S-transferase (GST) and N-acetyltransferase (NAT) appear to play a role in determining risk for lung cancer associated with PAHs and bladder cancer associated with exposure to cigarette smoke and aromatic amines, respectively. Polymorphisms in NAT result in delayed metabolism (termed slow acetylation) of aromatic amines (contained in cigarette smoke and some organic pigments), presumably increasing the risk for bladder cancer by increasing duration and intensity of exposure. These examples illustrate the concept of gene-environment interactions in which there are differences in the response to an occupational or environmental exposure on the risk for diseases, such as cancer, in persons with different genotypes.

Mutations in oncogenes or tumor suppressor genes or in the abnormal protein products encoded by them potentially may be detectable in tumor tissue or in the serum and urine. For example, studies of lung cancers in smokers and in never smokers have identified different levels of mutations of p53 tumor suppressor genes or K-ras oncogenes. To the extent that exposure-specific mutations can be identified, these mutations and the resulting protein products could serve as markers for exposure and potentially for risk for specific occupationally induced cancers. It is also possible that further analysis of altered genes or proteins may be useful as preclinical response indicators in premalignant and early malignant lesions in occupationally exposed cohorts, with a role in early detection. Considerable additional information from molecular epidemiologic studies will be necessary before these measurements can be readily applied in clinical situations.

Another potentially valuable tool is the measurement of levels of specific carcinogens covalently bound to DNA or proteins, referred to as DNA or protein adducts. Binding to DNA, detectable in white blood cells or tissues, may lead to DNA damage. Protein adducts (to albumin or hemoglobin) serve as potential surrogates for DNA binding and have the advantage of representing an internalized dose that is more readily measurable. Measurement of adducts presents a promising and perhaps better method of quantifying ­internal dose compared to older available methods, such as air monitoring or measuring blood or urine levels of an agent. Application of these methods has included measurement of PAH–DNA adducts in smokers and lung cancer patients and of hemoglobin adducts of aromatic amines in smokers and occupationally exposed groups.

These approaches have begun to advance the understanding of the causes and mechanisms of chemically induced cancers. However, they do not yet have clinical application in the determination of the risk for cancer in exposed individuals or confirmation that a specific cancer is related to an exposure. These limitations occur for a number of reasons, for example, there are imperfect and inconsistent correlations between these measures; the association between them and individual cancer risk is not clear; and they do not yet permit differentiation of the cause of any specific cancer. In addition, many other factors contribute to the induction of cancers, which may preclude effective clinical application of these tests. In addition, information about polymorphisms could be improperly employed to attempt to predict individual risk and to discriminate against individuals, raising important ethical issues, although recently passed legislation, the Genetic Information Nondiscrimination Act (GINA), legally precludes this use of such information.

Integration of Study Results in the Assessment of Cancer Risk

Based on analysis of epidemiologic and animal studies through its Monograph Series, IARC regularly updates lists that categorize chemicals and other agents by the level of evidence for human carcinogenicity. The Monographs, currently through Volume 101 from 2012, discuss the potential for carcinogenicity of these agents. On its Web site, IARC stated: “Since 1971, more than 900 agents have been evaluated, of which more than 400 have been identified as carcinogenic, probably carcinogenic, or possibly carcinogenic to humans.” (http://monographs.iarc.fr/) These agents include specific chemicals, groups of chemicals, complex mixtures, occupational exposures, industrial processes, lifestyle factors, biological agents, and physical agents. The groups (categories) are defined as:

• Group 1: The agent is carcinogenic to humans.

• Group 2A: The agent is probably carcinogenic to humans.

• Group 2B: The agent is possibly carcinogenic to humans.

• Group 3: The agent is not classifiable as to its carcinogenicity to humans.

• Group 4: The agent is probably not carcinogenic to humans.

The current IARC Preamble provides detailed criteria for placing an agent into one of these groups. For example, IARC provides the criteria for Group 1:

Current lists of the agents in each of these groups are available at http://monographs.iarc.fr/ENG/Classification/index.php. Other classifications of specific agents and lists of agents (categorized somewhat differently) can also be found in the current report on carcinogens (12th Report on Carcinogens) prepared by the U.S. National Toxicology Program, available at http://ntp.niehs.nih.gov/. Of note, while these agencies provide levels of evidence to assess the potential for cancer causation by these agents, they do not provide information about quantitative dose-response relationships, including the dose of the agent required to induce cancers in humans.

The results of all of these studies that attempt to assess carcinogenicity of and cancer risk from chemicals should serve as the basis for risk management, which involves public policy as well as scientific considerations. When a sufficient body of evidence supporting carcinogenicity in humans exists, corrective action to protect public and worker health should proceed, even if there is some remaining uncertainty in the conclusions. Sufficient evidence of carcinogenicity in humans should prompt immediate protective interventions. Convincingly positive results from a well-conducted epidemiologic study or sufficient evidence of carcinogenicity in animals, as defined by IARC or NTP, should prompt attempts to reduce worker exposure as much as possible. The finding of limited evidence in animal bioassays or positive results in short-term tests should, at the very least, serve as a stimulus for further study of the suspect chemical. When the results in different studies are contradictory, results suggesting carcinogenicity generally outweigh the negative evidence. Given the limited sensitivity of epidemiologic methods, this principle for managing exposures should be considered when animal studies are clearly positive while epidemiologic studies do not appear to show increased risks. These thoughts incorporate the European view of the precautionary principle.

While it is difficult scientifically to establish definitively that there is no increment in risk associated with very low doses of known human carcinogens, quantitative risk assessment can provide approximations of the lifetime doses that will likely produce only very small increases in risk. Mathematical models have been designed that allow for the extrapolation from high-dose studies to lower-dose exposures, providing an estimate of the approximate upper limit of excess risk (or the resulting excess number of cases) of a specific cancer that might be seen in a given population as a result of a particular exposure. However, there are substantial uncertainties in applying this approach, because of several factors such as limitations in the evidence from available studies and the requirement for assumptions and models to assess the shape of the dose-response curve at low doses. The principle of lowest possible exposure helps in reducing risk.

The nature of the appropriate response by government to evidence for carcinogenicity of a chemical is controversial, in part because of these uncertainties. In an ideal world, human exposure to known carcinogens would be nil, but, in practice, political, economic, social, and technical factors constrain the power of regulators to adopt such a stringent standard. Quantitative risk-assessment methodology can provide some assistance in making risk management decisions. From this information, regulatory agencies can set dose levels that do not exceed “acceptable” risks, defined by some agencies as 1 in 10,000 or 1 in a million. The agency may then mandate control of exposures to ensure that the cumulative lifetime or occupational dose is sufficiently low such that the incremental cancer risk is not likely to exceed acceptable levels. Such risk assessments can also help to place hazards from various chemicals in any given setting into perspective by allowing comparison of the risk estimate against that from other known hazards. The use of risk comparisons may allow regulators or decision makers to prioritize exposure problems and to allocate scarce resources for remediation activities.

Of course, industrial organizations, relying upon recommendations from occupational health and toxicology professionals, have the opportunity to go further than the regulations require in using information regarding carcinogenicity to make decisions about chemical use and control. Unfortunately they rarely do so, and industries have been known to purposefully create uncertainty. For example, decisions to avoid the use of chemicals in IARC groups 1 and 2A (or those with evidence that would likely place them in these categories) or, alternatively, to tightly control exposures to these agents (to completely prevent exposure) may lead to reductions in the risk for occupational cancers. Pressure applied by chemical users on chemical manufacturers to study suspect chemicals and to develop safe alternatives to potential carcinogens could be another effective approach. Additionally, proper hazard communication should inform workers of potential carcinogens in their work area and provide them with the training and tools to prevent or minimize exposure.

Exposure to known or suspected carcinogens clearly has declined in the United States and other economically developed countries as a result of both regulations that control exposure and changes in the chemicals produced and used and in the methods of production. Unfortunately, exposures to carcinogens in some developing countries are increasing in frequency and intensity. When measurements exist and are reported for workers in developing countries, exposure levels in given industries to carcinogens tend to be considerably higher than in developed countries, sometimes rising above regulatory standards in developed countries, and even the higher standards adopted in some developing countries The transfer of hazardous industries to developing countries likely will further increase carcinogen exposure for workers in these countries. Recognizing the reduced ability of many developing countries to regulate these hazards effectively, it is incumbent on industrial concerns from developed nations to attempt to control these hazards for their workers (or contract workers) in developing countries.

Medical Surveillance

The role of medical surveillance in workers currently or previously exposed to known or suspected carcinogens should follow well established principles. Surveillance of populations at high risk of cancer is only effective if the screening test is sensitive and easy to perform, if it detects premalignant abnormalities or tumors at an early stage in their development, and if there is an effective intervention that reduces morbidity and mortality when applied to individuals at risk. For certain tumors not known to be associated with chemical exposures (eg, cervical cancer), screening techniques and effective therapy for early lesions have had a significant impact on the disease. There is some evidence that a small group of workers at high risk of bladder tumors, as a result of prior exposure to aromatic amines used in dyestuff manufacturing, can benefit from early detection by the use of urine cytology and cystoscopy as screening tools. The finding in The National Lung Screening Trial that screening of current and former high-risk cigarette smokers (≥30 pack year history) with low-dose CT scans reduced mortality from lung cancer by 20% raises the possibility that use of these tests might ultimately play a role in the screening of high-risk groups exposed to other carcinogens.

Despite these limitations, properly collected medical surveillance data combined with industrial hygiene data collection may prove useful in future epidemiologic studies and in the refinement of our knowledge regarding human dose-response relationships. If medical surveillance is to be performed, the protocol should be designed for each agent of concern based on the presumed target site from prior human and animal studies and the availability of screening tools. In practice, some form of medical surveillance is required by Occupational Safety and Health Administration (OSHA) standards for asbestos, arsenic, benzene, and a variety of other carcinogens, as listed in Table 19–2.

Evaluation of Individual Cancer Cases or Concerns

The practice of occupational medicine sometimes requires an assessment as to whether a cancer in an exposed worker is causally related to an exposure at work or to exposure. Such an assessment may occur informally in discussion with a concerned affected employee or more formally in the setting of a workers' compensation claim or toxic tort case. Unfortunately, neither the principles of carcinogenesis nor the investigative methods for assessing the carcinogenicity of a particular chemical were designed to be used in and cannot readily be applied to the assessment of an individual case.

In a medicolegal context, there are two elements involved in the assessment of whether a cancer might be exposure related. The first is an assessment of whether the agent is capable of causing the relevant type of cancer. This assessment involves the evaluation of the level of scientific ­evidence for human carcinogenicity, as discussed in the section Integration of Study Results in the Assessment of Cancer Risk, and is ideally supported by an IARC or similar carcinogenicity classification. If an agent has not been shown to be causally associated with a cancer in humans, it is not possible to causally associate exposure to that agent with a cancer in a specific individual. This assessment is often, in a legal context, termed general causation. Establishing general causation is a necessary, but not sufficient, condition for a determination that a specific cancer in an individual is causally related to his or her particular exposures, referred to as specific causation. The latter assessment also involves consideration of potential alternative causes, latency issues, and synergy.

Some of the same factors used in the assessment of the work-relatedness of any illness, largely derived from the medical and occupational history and medical, employment, and exposure records, are important in the assessment of a possible occupational cancer. Obtaining such information may be complicated by the long time elapsed since exposures began and the absence of industrial hygiene or exposure records. Nevertheless, it is critical to assess, to the extent possible, the nature of the agents involved and the intensity, setting, control, timing, and duration of the exposures, and other sources of exposure. Potential sources for this information include the individual, coworkers and managers, material safety data sheets or other sources of chemical use information, and if available, industrial hygiene data. Knowledge of the presence or absence of other symptoms or conditions that may be due to exposures may be helpful in establishing that substantial exposure had occurred. For some industries, there may be published exposure-assessment and dose-effect information. From these sources it may be possible to derive a qualitative or semi-quantitative sense of the intensity, timing, and duration, as well as the potential health significance, of the exposures.

The medical history and medical records provide information about the cancer site and cell type and the presence of any other known risk factors for the cancer. Physical examination may be helpful, by demonstrating physical findings suggestive of other conditions associated with the exposure or other risk factors.

With the rare exception of certain tumors, such as mesotheliomas that almost universally stem from asbestos exposure, there is generally nothing about the appearance or behavior of a particular cancer that allows differentiation between work-related and spontaneous etiologies. The occurrence of some pathological types of cancer may suggest a particular etiology.

Literature review can provide the descriptive epidemiology of the tumor type, including age, sex, and racial patterns of incidence, as well as information regarding known nonoccupational risk factors for the tumor type. Literature searches can identify any relevant epidemiologic studies, which may be chemical specific or job-or process specific. In addition, there are a number of published occupational mortality studies that can provide information about mortality rates from cancers at certain sites and from other causes in specific occupational groups. Literature searches also can identify animal experimental studies of specific chemicals that provide information about tumor sites, type, frequency, and dose response. Synthesis of this information first involves assessment of the quality of the epidemiologic and animal experimental evidence using the criteria discussed earlier. The literature review may identify published data regarding exposures in different settings, which may be helpful in estimating exposures in an individual, although such estimated dose reconstructions may be fraught with difficulty. Assessing the time between initial exposure to the chemical agent and the development of the cancer may help to determine if there has been an adequate latency period. For solid tumors in humans thought to be caused by exposures, the minimum latency from first exposure to clinical evidence of cancer appears to be about 10–12 years or more. Therefore, it is unlikely that tumors that develop within a few years of initial exposure to the suspect agent are causally related to that exposure. Cohort studies of workers exposed to known carcinogens sometimes demonstrate the occurrence of certain cancers following a shorter interval since first exposure. In contrast to the prolonged latency required for solid tumors, exposure-related cancers of the blood (eg, leukemia) and lymphatic system (eg, lymphoma) generally are seen within 3–7 years following exposure. Slower-growing hematologic cancers (eg, myelodysplasia or low-grade lymphomas) may be more delayed in appearance.

All of the above information may be useful in developing a “differential assessment” of potential causative factors in an individual, akin to a differential diagnosis. In addition to an analysis of exposures, demographic features, latency period, and nonoccupational risk factors in the affected individual need to be factored into the assessment of causation.

In firefighters and police officers, there may be legislated presumptions in workers' compensation cases that cancers occurring during or after employment are considered to be work related. When such individuals, exposed to an agent known to be capable of causing cancer in humans, develop a cancer, it is presumed that the exposure caused the cancer unless there is evidence indicating that there is no “reasonable link” between the exposure and the cancer (thus the term, rebuttable presumption). These presumptions do not necessarily reflect that there is convincing epidemiologic or other evidence to support causation in the case. Nevertheless, the individual may be eligible for an award of compensation.

In the end, there will often be limitations to the ability to evaluate specific causation in a case. These limitations include: (1) there may be limited or no epidemiologic (or animal experimental) data to allow an assessment of the carcinogenicity of an agent (leading to problems in the assessment of general causation); (2) often there is incomplete exposure/dose information in the published literature or for the individual to permit estimation of the individual's risk from exposure; and (3) there can be multiple potential causes in an individual that may contribute to risk or interact with one another (such as cigarette smoking and asbestos exposure), making it difficult to assess the relative contribution of different risk factors in an individual case. In workers' compensation and toxic tort cases, the standard of proof generally is reasonable medical probability (ie, that the exposure more likely than not caused the medical condition). Possible connections are not sufficient to establish causation. In an individual case, assessment of specific causation at best is necessarily probabilistic (not definitive).

Outside the medical/legal context, groups of employees may express concerns about whether they are at risk of developing cancer because of specific workplace exposures. Evaluation of these concerns initially involves similar assessments of the potential for general and specific causation related to the exposures. Risk communication efforts are then warranted to share meaningful information and provide perspective about possible risks. It is important to recognize that lay people often have a poor understanding of and misperceptions about risk. For example, surveys have shown that individuals often perceive that pollution and occupational exposures are the main causes of cancer, when, in fact, lifestyle factors, such as smoking, diet, and obesity are overall much more important risk factors for cancer. Notions that “There must be a cause for the cancer” are common, without recognition of the existence of background risks. Application of the principles of risk communication, informed by scientific information and an understanding of the nature and magnitude of exposures, may help to reduce misperceptions of risk and to provide reassurance, when appropriate.

Possible clusters of cancer in a working population pose somewhat different challenges to the occupational physician in terms of investigating causation and communicating risk. Clusters are defined as groups of like or similar illnesses, both pathologically and etiologically, aggregated in space and time within a group of individuals with the same occupation. A commonly encountered scenario involves the recognition by a group of workers that two or more individuals in their group have had cancer. The first challenge is to confirm that there is indeed a cluster, perhaps by interviewing members of the group or by reviewing medical records. Individuals may have distinctly different types of tumors that are typically etiologically unrelated (eg, breast cancer, Hodgkin lymphoma, and lung cancer). The individuals may not have shared the same space for very long, or one may have had cancer prior to joining the group. It is also important to assess exposure in the work area and to look for potential sources of exposure to carcinogens or other hazardous chemical or physical agents. If initial investigation does reveal that a true cluster might exist, the next step is to confirm that the observed incidence exceeds what would have been expected in a population of comparable size and demographics (ie, that the apparent clustering did not occur by chance). This assessment requires appropriate statistical methods, typically using a Poisson distribution for low-frequency data, with comparison to cancer incidence data for a comparable population. There are several published investigations of “true” clusters where no plausible responsible environmental factor could be identified, even when the clusters did not appear to be due to chance. Thus, failure to determine an environmental cause after a thorough investigation should not be surprising. In the absence of known causes for observed clusters of cancer, careful and accurate presentation of the investigation results may help to ease concerns in the work force.

LUNG CANCER

Occupations at Risk

• Asbestos-exposed workers, including miners, insulators, and shipyard workers

• Workers exposed to radon, for example, uranium miners

• Chemical production workers exposed to chloromethyl ethers

  • Workers exposed to diesel exhaust/diesel particulate matter

• Workers exposed to polycyclic aromatic hydrocarbons, for example, aluminum reduction workers, coke oven workers, roofers, and rubber production workers

• Workers exposed to hexavalent chromium compounds, for example, in chromate production

• Workers exposed to nickel compounds, for example, in nickel mining and refining

• Workers exposed to inorganic arsenic compounds, for example, in arsenical pesticide production and use; and in copper, lead, and zinc smelting

General Considerations

Lung cancer is the leading cause of cancer-related death in North America and Europe. Lung cancers account for 33% of new cases and 25–30% of deaths in the United States. When the number of new cases equals or comes close to the number of deaths, it is an indication that the success of treatment is not good. There has been a decline in lung cancer seen most clearly in men; only recently has the decline become apparent among women in the United States. Unfortunately, in many parts of the world, especially in countries with developing economies, cigarette use continues to increase, and along with it, the incidence of lung cancers is also rising. While tobacco smoking remains the primary cause of lung cancer worldwide, more than 60% of new lung cancers occur in never smokers or former smokers, many of whom quit decades ago. Moreover, 1 in 5 women and 1 in 12 men diagnosed with lung cancer have never smoked.

Etiology

Cigarette smoking is the most important and most preventable risk factor for cancer of the lung. More than 80% of lung cancer deaths are attributable to cigarette smoking. Although its relative importance may decline if recent trends toward reduced cigarette consumption and the use of cigarettes with decreased tar and nicotine continue, the increasing incidence of lung cancer in women correlates with an increase in the smoking habit. Occupations with a high smoking prevalence have an increased risk of cancer. This includes restaurant wait staff, cashiers, orderlies, drivers, construction workers, watchmen, and others where smoking prevalence may be higher than 40%. The high levels of environmental tobacco smoke found in some workplaces may increase the risk of lung cancer as well, based upon IARC's conclusion in 2002 that involuntary smoking (exposure to secondhand or “environmental” tobacco smoke) is carcinogenic to humans.

Exposures at work have been estimated to contribute to 10% of all lung cancer cases In addition to asbestos, other agents either proven or suspected to be respiratory carcinogens include acrylonitrile, arsenic compounds, beryllium, bis(chloromethyl) ether, chromium (hexavalent), formaldehyde, mustard gas, nickel carbonyl (nickel smelting), polyaromatic hydrocarbons (coke oven emissions and diesel exhaust), secondhand tobacco smoke, silica (both mining and processing), talc (possible asbestos contamination in both mining and milling), vinyl chloride (sarcomas), and uranium. Workers at risk of radiation-related lung cancer include not only those involved in mining or processing uranium but also those exposed in underground mining operations of other ores where radon daughters may be emitted from rock formations. The association of lung cancer with exposure to most of these agents appears to be independent of cigarette smoking. However, the effects of some known occupational carcinogens are greatly enhanced by smoking (eg, asbestos, radon).

A. Asbestos

Asbestos is the substance generally considered to pose the greatest carcinogenic threat in the workplace. About 125 million people around the world are exposed to asbestos in their work environments, and many millions more workers have been exposed to asbestos in years past. NIOSH has estimated that current occupational exposures to asbestos will cause five deaths from lung cancer in every 1000 workers exposed for a working lifetime. About 20–40% of adult men report past occupations that may have entailed asbestos exposures. In the most highly affected age groups, mesothelioma may account for over 1% of all deaths. In addition to mesothelioma, 5–7% of all lung cancers are potentially attributable to occupational exposures to asbestos.

Asbestos refers to a group of fibrous silicates of several types. The minerals are divided into two classes: serpentine (chrysotile) and amphiboles (amosite, crocidolite, actinolite, anthophyllite, and tremolite). The three most common commercial forms are chrysotile, amosite, and crocidolite. Chrysotile represents 95 percent of all the asbestos ever used worldwide. All three commonly used forms of asbestos are known to cause an increased risk of cancer. Exposures are frequently to mixed fiber types. Despite all that is known about the health effects of asbestos, annual world production remains at over 2 million tons. Global asbestos trade increased by more than 20% in 2012.

Lung cancer is a major asbestos-related disease, accounting for 20% of all deaths in asbestos-exposed cohorts. A latency period of approximately 20 years has been noted before the majority of lung cancer cases are seen. Asbestos exposure increases the risk of lung cancer fivefold in nonsmokers. Several studies show evidence that cigarette smokers who were also exposed to asbestos have a much greater risk of developing cancer of the lung, indicating a synergistic effect between these carcinogens.

B. Radon

Radon exposure is known to increase the risk of lung cancer. This carcinogenic effect was discovered when increased mortality rates from lung cancer were identified in uranium miners. Large-scale mining of uranium began in the United States in 1948 because of the need for uranium to make nuclear weapons. By the 1960s, 20% of deaths in uranium miners in the United States were a result of lung disease. Excessive lung cancer in uranium miners is independent of cigarette smoking, although exposure to both is synergistic.

Ores containing uranium include all its decay products, which form a series of radionuclides, of which one is the inert gas radon. Radon diffuses out of the rock into the mine atmosphere, where it decays into radioisotopes of polonium, bismuth, and lead—termed radon daughters. These radionuclides are found in the air and then are inhaled as free ions or as attachments to dust particles. Epidemiologic studies of workers in US uranium mines demonstrate that the risk of lung cancer is proportionate to the cumulative radon daughter exposure. Increased risk of lung cancer also has been found in fluorspar miners, iron ore (hematite) miners, and hard-rock miners. Data from animal models support the carcinogenic effect of radon; respiratory tumors can be induced by inhaled radon daughter products.

Domestic radon exposure has been an issue of concern since 1984, when high radon levels were discovered in homes built on the Reading Prong geologic formation in Pennsylvania. The risk of lung cancer from low-level radon exposure has been extrapolated from studies of mine workers to the general population but appears to be very low.

C. Chloromethyl Ethers

Exposure to multiple chemical substances can cause an increase in lung cancers in exposed workers. Among the most historically important of these are the chloromethyl ethers, which include chloromethylmethyl ether (CMME) and bischloromethyl ether (BCME). Chloromethyl ethers are produced in order to chloromethylate other organic chemicals in the manufacture of ion-exchange resins, bactericides, pesticides, dispersing agents, water repellents, solvents for industrial polymerization reactions, and flameproofing agents. The potential for chloromethyl ethers to cause cancer was first suspected in humans in 1962. In Philadelphia, cases of small-cell lung cancer occurred among approximately 45 men working in a single building of a large chemical plant. A large proportion of tumors occurred in young men and nonsmokers. Numerous other studies confirm these findings, with increased risk seen in workers with prolonged or intense exposure. Unlike other chemical carcinogens, which can cause a variety of cancers, the chloromethyl ethers are associated primarily with the induction of small-cell lung cancer. Inhalation studies in animals show that the chloromethyl ethers produce bronchial epithelial metaplasia and atypia, and both carcinogens are active alkylating agents. BCME is a more potent carcinogen than CMME.

D. Polycyclic Aromatic Hydrocarbons

PAHs, formed from the incomplete combustion of coal tar, pitch, oil, and coke, have long been recognized as carcinogens. In 1775, Sir Percival Pott reported an increased risk of scrotal cancer in chimney sweeps as a consequence of dermal exposure to soot. Epidemiologic evidence linking PAHs to lung cancer was provided in 1936, when a study of exposed workers in a coal carbonization plant in Japan revealed a marked increase in the rate of lung cancer.

Exposures to PAHs linked to an increased risk of lung cancer have been found in coke oven workers, roofers, printers, and truckers. Rubber plant workers and those employed in asphalt production, coal gasification, and aluminum reduction facilities are also at risk. The best-described occupational group is coke oven workers, where direct exposure to the coke oven emissions results in increased rates of lung cancer. A clear dose-response relationship has been described based on proximity of work to the ovens and the potential for exposure to PAHs.

E. Diesel-Engine Exhaust

Studies of miners, railroad workers, and truckers have demonstrated significant increases in the risk of lung cancer associated with exposures to diesel-engine exhaust. Further, this association has been observed in multiple case-control studies, including a large pooled analysis of 11 population-based case-control studies from Europe and Canada, which was adjusted for cigarette smoking. In 2012, IARC concluded, based on these studies, that there was sufficient evidence in humans for the carcinogenicity of diesel-engine exhaust. Similarly, IARC stated that animal bioassays demonstrated sufficient evidence of carcinogenicity. Diesel-engine exhaust contains a number of nitroarenes, which are nitro-substituted derivatives of polycyclic aromatic hydrocarbons (arenes). Many of these agents are animal carcinogens and are genotoxic.

F. Other Chemicals

1. Arsenic

Exposure to inorganic arsenic increases the risk of lung cancer; the first cases of arsenic-induced lung cancer were reported in 1930. Arsenic exposure in copper smelting, fur handling, sheep-dip compound manufacturing, and arsenical pesticide production and use has resulted in increased rates of lung cancer. Long latency periods of approximately 25 years are seen after exposure before the development of cancer. Arsenic is thought to act as a late-stage promoter of cancer and may interfere with DNA repair mechanisms. A dose-response relationship in exposed workers has been described There is some evidence for a synergistic effect of smoking and arsenic exposure in increasing the risk of lung cancer.

2. Beryllium­

Increased risks of lung cancer have been observed in studies of beryllium-processing workers. IARC concluded in 2012 that beryllium and beryllium compounds cause cancer of the lung in humans.

3. Cadmium

Increased risks of lung cancer have been reported in some studies of cadmium-processing workers, nickel-cadmium (Ni-Cd) battery workers, and workers in a cadmium recovery plant. Despite the possibility that coexposure to other lung carcinogens, such as cigarette smoke, arsenic, and nickel, could have contributed to excess risks, IARC concluded in 2012 that cadmium and cadmium compounds cause cancer of the lung in humans.

4. Chromium

Increased rates of lung cancer have been reported in industries, such as chromate production, chrome plating, and chrome-alloy production, which use chromium (VI) compounds, also known as hexavalent chromium compounds. Other lung carcinogens used in the electroplating industry, such as nickel and PAHs, may confound this relationship. IARC has concluded that chromium (VI) compounds cause cancer of the lung.

5. Nickel

Exposure to nickel in mining, refining, and subsulfide roasting facilities is associated with increased rates of lung and nasal cancer. While exposure to both soluble and insoluble nickel compounds has been associated with lung cancer risk, the evidence is strongest for water-soluble nickel compounds. IARC has concluded that there is sufficient evidence in humans for the carcinogenicity of mixtures that include nickel compounds and nickel metal.

6. Mustard gas

Studies of Japanese and German workers in factories that manufactured mustard gas during World War II show an excess of respiratory cancers. This is consistent with the finding that mustard gas can produce lung tumors in laboratory animals. There may be a higher rate of squamous cell cancer of the lung in humans.

7. Silica

In a number of occupational settings, workers exposed to crystalline silica had increased risks for lung cancer, including in quarries and granite works and in refractory brick and diatomaceous earth industries. Studies of individuals with documented silicosis have also demonstrated increased lung cancer risk. Based upon this information, IARC concluded in 1997 that there is sufficient evidence in humans for the carcinogenicity of inhaled crystalline silica in the form of quartz or cristobalite from occupational sources.

8. Other agents

IARC has concluded that painting, ionizing radiation, and rubber and aluminum production are known to cause lung cancer in humans.

9. Probable or possible human lung carcinogens

Some studies have suggested but not demonstrated an increased risk of lung cancer in humans associated with exposure to certain other agents. Agents for which there are varying degrees of evidence suggesting human lung carcinogenicity include acrylonitrile, formaldehyde, mists from strong inorganic acids (including sulfuric acid), and vinyl chloride.

Pathology

The four major types of lung cancer are squamous cell (epidermoid) carcinoma, adenocarcinoma, large-cell carcinoma, and small-cell (oat-cell) carcinoma. All histologic types of lung cancer are linked to cigarette smoking. There is no one cell type that is pathognomonic of an occupationally related lung cancer. Even in studies of workers exposed to CMME or BCME, who are much more likely to develop the relatively uncommon small-cell histology, other types of lung cancer have been observed. Although early work suggested that the peripheral distribution of asbestos fibers was associated with a higher incidence of adenocarcinomas in this region, this has not been found in recent, more thorough studies. It appears that lung cancers in asbestos-exposed persons occur equally throughout the lung, and all pathologic types are seen.

Clinical Findings

Symptoms and Signs

• 75–90% are symptomatic at diagnosis.

• Presentation depends on

  • Type and location of tumor

  • Extent of spread

  • Presence of distant metastases and any paraneoplastic syndromes

• Anorexia, weight loss, and asthenia in 55–90%.

• New or changed cough in up to 60%.

• Hemoptysis in 5–30%.

• Pain, often from bony metastases, in 25–40%

• Local spread may result in endobronchial obstruction and postobstructive pneumonia, effusions, or a change in voice due to recurrent laryngeal nerve involvement

• Superior vena cava (SVC) syndrome

• Horner syndrome

• Liver metastases are associated with asthenia and weight loss

• Possible presentation of brain metastases

The symptoms and signs and laboratory and imaging procedure findings in occupational lung cancer generally do not differ from lung cancers of nonoccupational etiology. In some cases, an imaging or other finding may suggest a particular etiology, for example, the presence of pleural plaques, in conjunction with a lung tumor, would suggest heavy asbestos exposure as the cause.

Prevention

Avoidance of exposure to lung carcinogens is the most important preventive measure, but complete avoidance is typically not possible, especially for those agents that occur naturally in the environment such as asbestos, arsenic, and silica. The most effective method of reducing the mortality rate for lung cancer is primary prevention. This includes identification of etiologic agents in the workplace, adherence to strict workplace standards, and worker education. Because tobacco use is known to increase the incidence of lung cancer in occupationally exposed groups, aggressive anti-smoking campaigns in the workplace are critically important.

Medical monitoring in the workplace has been attempted as a method of secondary prevention to aid in early detection. Serial chest radiographs and sputum cytologic examinations are recommended by the National Institute for Occupational Safety and Health (NIOSH) and mandated by OSHA in some high-risk occupational groups. The main problem with this approach is that there is no evidence that early detection improves the prognosis for persons with occupationally-induced lung cancer. Thus far serial chest radiographs have been more useful than sputum cytologic examinations in detecting lung cancer. However, sputum cytology may reveal signs of mucosal damage, such as atypia, that could identify individuals at increased risk and lead to decreased exposure.

There is currently a lack of evidence supporting the use of chemoprevention for lung cancer in high-risk populations. A study of primary chemoprevention of lung cancer, using retinol and beta carotene, in current and former smokers and asbestos workers was discontinued after increases in risk were observed.

Treatment & Prognosis

Therapy of occupationally-induced lung cancers is no different from treatment for each of the specific cell types of lung cancer that may be seen in other settings. In general, even in patients with localized disease, long-term survival is the exception rather than the rule.

MESOTHELIOMA

Occupations at Risk

• Asbestos miners

• Construction workers

• Workers exposed to insulation materials in production, installation, and removal

• Shipyard workers

• Asbestos textile manufacturing

• Welders, plumbers, electricians

General Considerations

Mesothelioma is uncommon, accounting for only a small fraction of deaths caused by cancer, but it and other asbestos-related diseases have been of great interest to occupational health physicians and to public health professionals. This is because both community-based and industrial exposures to asbestos and asbestiform fibers increase risks for mesothelioma. Exposure to asbestos from the use of construction materials that contain asbestos is a serious and often neglected problem throughout the world.

Worldwide, the yearly number of asbestos-related cancer deaths in workers is estimated to be 100,000–140,000. In Western Europe, North America, Japan, and Australia, 20,000 new cases of lung cancer and 10,000 cases of ­mesothelioma result every year from exposures to asbestos. The incidence of mesothelioma has been increasing despite international efforts to ban the mining and manufacture of asbestos. The age-adjusted annual incidence for adults in North America is approximately 19 cases per million for men and 4 cases per million for women. The national incidence rates for mesothelioma in Australia are the highest in the world. In the United Kingdom, at least 3500 people die from asbestos-related illnesses each year. The British mesothelioma death rate is now the highest in the world, accounting for 1 in 40 of all male cancer deaths. About 1 in 170 of all British men born in the 1940s will die of mesothelioma. The incidence rates of peritoneal mesothelioma are about an order of magnitude less than those for pleural tumors.

Etiology

Diffuse malignant mesotheliomas of the peritoneum and pleura are considered “sentinel tumors” or pathognomonic of exposure to asbestos. The large majority of mesothelioma cases report past asbestos exposure. The latency period from asbestos exposure to the diagnosis of mesothelioma is often 30 years or more. Higher quantitative asbestos fiber content of dried lung has been found in some patients with mesothelioma. Further evidence of the etiologic role of asbestos has been shown in experimental animals in which intrapleural injection or administration by inhalation of asbestos fibers causes mesothelioma that is histologically identical to human tumors.

Epidemiologic data show that variable levels of exposure to asbestos can result in mesothelioma, despite the known dose-response relationship. While most cases occur in individuals with a history of heavy asbestos exposure, some cases occur in individuals with relatively trivial contact at work or in the home environment (eg, exposure of wives washing their husbands' contaminated work clothes).

The major value of studies done to date has been to identify segments of the population at risk, but reports of patients with mesothelioma who do not have a history of occupational or paraoccupational exposure to asbestos raise other questions. The proportion of patients with no exposure history ranges from 0 to 87% in various studies. The long latency period from exposure to disease results in problems with forgotten or unknown exposures. In addition, the variety of occupations associated with asbestos exposure leads to problems with overlooked exposures. Exposure occurs in the milling, mining, and transportation of raw asbestos and in the manufacture of asbestos cement pipe, friction materials, textiles, and roofing materials. Construction workers, plumbers, welders, and electricians are all exposed, and shipyard tradesmen can be “innocent bystanders” when they are exposed to airborne asbestos fibers. There is also some evidence that nonasbestos agents may induce malignant mesotheliomas, including erionite, a nonasbestos fiber, and ionizing radiation. Cigarette smoking does not increase the risk of malignant mesothelioma. Unlike lung cancer, there is no evidence for synergy between cigarette smoking and asbestos exposure in the development of this tumor.

Pathogenesis

All types of asbestos are capable of causing mesothelioma, although there is evidence that amphiboles, particularly crocidolite and amosite, are the most potent carcinogens. The mechanisms of induction are unknown. Cancer development is apparently related not to chemical composition but to physical properties (ie, fiber size and dimension). In work done in rats, long, thin fibers of a variety of types have proved carcinogenic, whereas short fibers and those with a relatively broad diameter have failed to produce mesothelioma. Inhaled fibers are expectorated or swallowed. Short fibers are cleared more readily than long fibers and are more likely to end up in the pleura. Fibers that remain accumulated in the lower lung, adjacent to the pleura. The pathogenesis of peritoneal mesothelioma is thought to be similar to that of pleural tumors. Fibers of asbestos are transported in lymphatics to the abdomen, and asbestos is also transported across the mucosa of the gut after ingestion. This location of mesothelioma is related to the type of asbestos fiber as well, in that peritoneal mesothelioma occurs much more frequently in individuals exposed to amphibole asbestos rather than chrysotile asbestos.

Pathology

A major area of difficulty in the study of mesothelioma has been distinguishing its pathologic features. Many tumors metastasize and spread to the mesothelial lining of the chest and abdomen. This has led to misdiagnosis of mesothelioma when it was, in fact, a metastatic tumor, such as an adenocarcinoma, and the reverse is true as well. Confusion also exists because of the tumor's diverse microscopic appearance.

Two types of mesothelioma have been described: benign solitary and diffuse malignant. The benign solitary type remains localized, although it may become large and compress neighboring thoracic structures. This tumor has not been associated with asbestos exposure; it is a benign tumor arising from fibroblasts and other connective-tissue elements in the areolar submesothelial cell layers of the pleura and is not occupational in origin. By contrast, diffuse malignant mesothelioma arises from either the pluripotential mesenchymal cell or the primitive submesothelial mesenchymal cell, which retains the ability to form epithelial or connective-tissue elements.

Malignant mesothelioma is a diffuse lesion that spreads widely in the pleural space and usually is associated with extensive pleural effusion and direct invasion of thoracic structures. On gross examination, numerous tumor nodules may be noted, and in advanced cases, the tumor has a hard, woody consistency. Microscopically, malignant mesotheliomas consist of three histologic types: an epithelial (or epithelioid) type that may resemble metastatic adenocarcinoma, a mesenchymal type, and a mixed type. Histochemical and immunohistochemical techniques that use Alcian blue stains and a panel of antibodies to specific cellular antigens, respectively, can be employed to help distinguish mesothelioma from metastatic adenocarcinoma. Studies with the electron microscope have defined certain characteristic features that are also helpful in differentiating the tumor from metastatic disease.

Clinical Findings

A. Symptoms

Symptoms in diffuse pleural mesothelioma may be entirely absent or minimal at the time of onset of the disease. Disease progression results in the most common symptom of a persistent gnawing chest pain on the involved side, which may radiate to the shoulder and arm. In most patients, pain becomes the most incapacitating symptom. Dyspnea on exertion, dry cough (occasionally hemoptysis), and increasing weight loss are frequent accompanying symptoms. Some patients have low-grade fever, which can result in an incorrect diagnosis of chronic infection. The symptoms of peritoneal mesothelioma are nonspecific but may include increased abdominal girth, pain, and weight loss.

B. Signs

Physical findings vary with the stage of disease. Most patients present with pleural effusion. Local tumor growth may depress the diaphragm and displace the liver or spleen, giving the impression of hepatomegaly or splenomegaly. In advanced disease, there may be obvious enlargement of the affected hemithorax, with bulging of the intercostal spaces and displacement of the trachea and mediastinum to the unaffected side. After removal of pleural fluid, a pericardial or pleuropericardial rub may be heard. Advanced signs also may include mediastinal lymph node enlargement, subcutaneous nodules in the chest wall, and clubbing. Encroachment on the mediastinal structures may lead to neuropathic signs such as vocal cord paralysis or Horner syndrome. Congestion and edema may develop in the upper trunk or lower limbs secondary to compression of the superior or inferior vena cava.

C. Laboratory Findings

Laboratory findings are nonspecific but may include anemia and thrombocytosis.

D. Imaging

Radiographic studies of the chest most commonly show unilateral pleural effusion. After thoracentesis, the pleura may show thickening or nodularity, seen usually at the bases. CT scanning, which is the most sensitive test for evaluating the pleural surface, may show thickened tumor along the chest wall, and late in the disease, tomograms or an overpenetrated film will show compressed lung surrounded on all sides by a tumor 2–3 cm thick. Extrapleural extension can result in soft-tissue masses or radiologic evidence of rib destruction. Signs of asbestosis such as interstitial pulmonary fibrosis, pleural plaques, and calcification are valuable findings when present.

E. Special Examination

1. Sputum Cytology

Microscopic examination of sputum rarely shows malignant cells unless the tumor has invaded lung parenchyma. Asbestos bodies may be seen.

2. Thoracentesis

The considerable force necessary to enter the pleural space with a thoracentesis needle may be a clue to the presence of pleural mesothelioma. Pleural fluid is serosanguineous or hemorrhagic in 30–50% of cases but is commonly straw-colored. Cytologic examination of pleural fluid is not typically helpful diagnostically. Mesothelial hyperplasia is not uncommon in benign pleural effusions and easily can be mistaken for malignant cells.

3. Pleural Biopsy

Because of the limitations of pleural fluid cytologic examination, biopsy confirmation is required. A CT-guided pleural biopsy may permit diagnosis in some cases. Thoracoscopy (pleuroscopy) with biopsy of pleural masses can be an effective technique and is less invasive than an open biopsy. Pleurodesis (obliteration of the pleural space) with insufflation of talc to reduce recurrence of pleural effusions can be performed as part of this procedure. An open thoracotomy with multiple biopsies from different pleural areas is sometimes required for diagnosis.

Differential Diagnosis

The major disorders that must be differentiated from mesothelioma are inflammatory pleurisy, primary lung cancer, and metastatic adenocarcinoma or sarcoma. Inflammatory pleurisy is suggested by the associated clinical picture and by typical findings in the analysis of sputum and pleural fluid. In primary lung cancer, the more prominent symptom of cough, the less common presence of severe chest pain, the presence of parenchymal tumors, and the absence of pleural abnormalities after thoracentesis help to differentiate between these two types of cancer. Primary tumors of the pancreas, gastrointestinal tract, or ovary should be excluded because these tumors can metastasize to the pleural or peritoneal space and mimic mesothelioma.

Prevention

Avoidance of exposure to asbestos is the most effective means to prevent mesothelioma. The exposures to asbestos that lead to mesothelioma may be less intense and of shorter duration than the exposures that lead to asbestosis or lung cancer. Setting permissible limits requires establishment of dose-response relationships, with subsequent determination of an acceptable level of risk. The difficulty is that all industrial processes, fiber types, and asbestos-related diseases have dissimilar dose-response relationships. Control of asbestos dust in industry has become progressively more rigorous over the last 40 years. Recommendations for levels of asbestos in the air of occupational settings were first established in the 1940s, but it was not until 1970 that federal regulations began as a result of the passage of the Occupational Safety and Health Act and the Clean Air Act. The current OSHA standard is 0.1 fibers/cc of air on an 8-hour time-weighted average (TWA) basis, although adherence to this standard may not be fully protective against the development of mesotheliomas.

Treatment

A. Surgical Measures

Surgery has been used with some success as the primary method of treatment in pleural mesotheliomas, both for tumor debulking and for palliation of symptoms. Even with tumors with extensive infiltration of adjoining viscera, partial surgical resection has led to an apparent increase in longevity, although it is not curative. Subtotal pleurectomy with decortication is the accepted procedure. More radical surgeries such as pleuropneumonectomy (extrapleural pneumonectomy) may be appropriate for selected patients. Postoperative adjuvant chemotherapy and radiation therapy sometimes are used, but there are no studies to support their use. Surgical resection of all visible disease is believed to be the treatment of choice. Surgical excision has no role in the management of peritoneal mesothelioma unless the tumor is localized.

B. External Radiotherapy

Radiation therapy clearly has been shown to be of benefit in controlling pain and pleural effusion in mesothelioma. Although antitumor efficacy has been noted using high-dose radiation, this modality is relatively ineffective in altering the dismal survival statistics for this disease.

C. Chemotherapy

There has been no systematic study of the role of cytotoxic drugs in mesothelioma. While there are well-documented reports of definite antitumor effects in some patients, chemotherapy is not curative. Pemetrexed (a folate antimetabolite), cisplatin, gemcitabine (a nucleoside analog), methotrexate, and other drugs, sometimes in combination, have been used. U.S. Food and Drug Administration has approved combination treatment with pemetrexed and cisplatin for malignant pleural mesothelioma that is not surgically resectable.

Course & Prognosis

Approximately 75% of patients die within 1 year after diagnosis, with an average survival after diagnosis of 8–10 months. Several factors correlate with improved survival in mesothelioma. Patients whose tumors are in the pleura survive twice as long as those with peritoneal tumors; survival is longer for patients with epithelial types than for those with mixed or fibrosarcomatous types; and survival is longer for patients younger than age 65 years, those who respond well to chemotherapy, and those able to undergo surgical resection.

CANCER OF THE NASAL CAVITY & SINUSES

Occupations at Risk

• Wood and other dusts

  • Boot and shoe manufacturing

  • Furniture workers

  • Textile manufacturing

• Nickel

  • Nickel refinery workers

• Chromium

  • Chromate pigment manufacturing

  • Metal plating workers

Cancers of the nasal cavity and sinuses are rare and account for fewer than 10 cases per million in the United States per year. This disease is uncommon in younger than 40–50 years of age, and rates increase with age. Evidence suggests a fairly steady incidence over the years. Over 50% of all sinonasal tumors are squamous cell, while about 10% are adenocarcinomas. Both these histologies are linked to occupational exposures. Other histologic types include other carcinoma, sarcoma, and melanoma.

Etiology

IARC has concluded that cigarette smoking causes cancer of the nose and paranasal sinuses. Many different occupational exposures are linked to cancer of the nasal cavity and paranasal sinuses. These include wood and leather dust, nickel, radium, and isopropyl alcohol production (by the strong acid process), for which IARC has concluded that there is sufficient evidence in humans. Agents or industries, for which IARC has concluded that there is limited evidence in humans for causation of these tumors, include hexavalent chromium compounds, formaldehyde, carpentry and joinery, and textile manufacturing (possibly due to textile dusts, dyes, and/or formaldehyde). Employment in several other industries, including furniture and shoe manufacturing, with corresponding exposures to wood and leather dusts, respectively, also has frequently been associated with these cancers.

A. Wood and Other Organic Dusts

Many studies have shown an increased incidence of carcinoma of the sinonasal area in persons exposed to wood dust. Adenocarcinoma of the ethmoids and middle turbinates is the most frequent cell type encountered in these workers. The exact substance in wood dust responsible for carcinogenesis has not been identified.

An excess of both adenocarcinomas and squamous cell carcinomas of the nasal sinuses also has been observed among workers in the boot and shoe industry, exposed to leather dust. As in the case of woodworkers, the specific etiologic agent in leather dust is unknown. Dusts involved in the textile industry and flour dusts in bakeries and flour mills also have been associated with the development of sinonasal cancers.

B. Nickel

Both nasal cancer and lung cancer are linked to occupational nickel exposure. Most studies have been done on nickel refinery workers exposed to complex particulates (insoluble nickel sulfide dust, nickel oxides, and soluble nickel sulfate, nitrate, or chloride) and gaseous nickel carbonyl. Nickel and nickel carbonyl are carcinogenic under experimental conditions, yet epidemiologic evidence points away from the nickel carbonyl process and incriminates exposure to dust from the preliminary processes. The mean latency period between exposure and diagnosis of cancer in refinery workers is 20–30 years.

C. Other Occupational Exposures

Tumors of the nasal epithelium and mastoid air cells have been noted in women exposed to radium used for painting dials of watches and in radon chemists. Chromium is known to cause ulceration and perforation of the nasal septum, and there is an excess risk of sinonasal cancer in workers involved in manufacturing chromate pigments. Mustard gas, cutting oils (mineral oils), and formaldehyde are also linked to excess cancers of the nasal cavity and paranasal sinuses.

Clinical Findings

The earliest symptoms of nasal cavity neoplasms are a low-grade chronic infection associated with discharge, obstruction, and minor intermittent bleeding. The patient often complains of “sinus trouble” and may have been treated inappropriately with antibiotics for prolonged periods before the true diagnosis was known. Subsequent symptoms depend on the pattern of local growth. Maxillary sinus tumors develop silently when they are confined to the sinus, producing symptoms only with extension outside the walls. With extension into the oral cavity, pain may be referred to the upper teeth. Nasal obstruction and bleeding are common complaints, along with “sinus pain” or “fullness” of the involved antrum.

Diagnosis & Treatment

In all cases, the patient should receive careful inspection and palpation of the facial structures, with attention to the eye and especially the extraocular movements. The nasal and orbital cavities should be examined closely. Helpful radiologic studies include facial bone or sinus radiograph series and CT scan of the involved areas. Biopsies are required for diagnosis.

Therapy is the same for occupational cancers as with other nasal and sinus cancers, including surgical therapy and radiation therapy, with chemotherapy reserved for advanced disease. The prognosis is better for nasal cavity cancers because they tend to be diagnosed at an early stage.

CANCER OF THE LARYNX

Occupations at Risk

• Asbestos-exposed workers, including miners, insulators, and shipyard workers

• Workers exposed to strong inorganic acid mists

Cancer of the larynx is much more common than sinonasal cancer, representing about 2% of the total cancer risk in the United States. In the United States, there is evidence that the incidence of cancer of the larynx is decreasing.

Etiology

Cancer of the larynx appears to be related primarily to cigarette smoking. Alcohol is less important in the causation of laryngeal cancer than in other tumors of the head and neck. IARC has indicated that there is sufficient evidence in humans that both smoking and alcohol cause laryngeal cancer. Similarly, IARC concluded that there is sufficient evidence in humans for occupational exposure to asbestos. Asbestos exposure in a variety of occupations, including miners, asbestos product manufacturers, and insulators, is associated with high rates of laryngeal cancers. Similarly, IARC has concluded that there is sufficient evidence in humans that strong inorganic acid mists cause laryngeal cancer. Agents or industries for which IARC has concluded that there is limited evidence in humans are mustard gas (sulfur mustard), human papilloma virus type 16, and rubber production.

Laryngeal cancer is primarily a disease of older individuals, including workers, with incidence rates rising sharply after age 50. At the time of diagnosis, approximately 60% are localized, 30% show regional spread, and 10% have distant metastases. Laryngeal tumors in the United States are classified into three groups according to anatomic site of origin, with about 30–40% supraglottic, 60% glottic, and 1% subglottic cancers. Nearly all are squamous cell carcinomas.

Clinical Findings

Symptoms of laryngeal carcinoma vary depending on the site of involvement. Any patient who complains of persistent hoarseness, difficulty in swallowing, pain on swallowing, a “lump in the throat,” or a change in voice quality should be examined promptly by indirect laryngoscopy. Any limitation of motion or rigidity should be noted, and direct laryngoscopy with biopsy of suspicious lesions is necessary. Lateral soft-tissue radiographs of the neck and CT scanning are also useful, especially to delineate extent of disease.

Treatment

The treatment plan must include preservation of the patient's life and voice. Therapy is no different for work-related laryngeal cancers than for other laryngeal cancers.

BLADDER CANCER

Occupations at Risk

• Work with aromatic amines, such as 2-naphthylamine, 4-aminobiphenyl, benzidine, and ortho-toluidine

  • Dye/pigment manufacturing and use

• Workers in rubber manufacturing industries

  • Painters

  • Workers in rubber production

  • Truck drivers

• Methylene dianiline

  • Drill press operators

• Benzidine-derived azo dyes

General Considerations

Bladder cancer accounts for approximately 5% of all malignant tumors. In the United States, nearly 60,000 cases are diagnosed each year. The male-to-female ratio is about 4:1. The highest incidence of bladder cancer occurs in industrialized countries such as the United States, Canada, and France with lower incidence in less developed countries Smokers have two to three times the risk of nonsmokers. The increased frequency in men may reflect higher smoking rates and the fact that more men work in potentially hazardous occupations than do women. As with most cancers, the incidence of bladder cancer increases with age, with most cases occurring in individuals 65 years and older.

Etiology

Cigarette smoking is the most important known preventable cause of bladder cancer, with as many as 60% of cases attributed to this common habit. Occupational exposure is also a major cause, particularly in nonsmokers.

Large-scale production of aromatic amines as dye intermediates was started in the United States during World War I, and by 1934, the first occupational bladder cancers were described. Twenty-five cases of bladder cancer were reported in workers exposed to 2-naphthylamine (β-naphthylamine) or benzidine and two cases in workers exposed to α-naphthylamine. Several years later, 58 additional cases were reported from the same plant. In addition, β-naphthylamine was reported to induce urinary bladder tumors in dogs, and subcutaneous injections of benzidine were shown to induce carcinomas in rats. During the next three decades, several studies in the United States and Great Britain showed an increase in urinary bladder tumors in workers exposed to these chemicals. The latency period between exposure and cancer was quite variable, with a mean of 20 years. NIOSH has concluded that all benzidine-derived dyes should be recognized as potential human carcinogens, and since then, virtually all companies in the United States have stopped or reduced their manufacture.

Agents that IARC has classified as causing bladder cancer in humans include tobacco, arsenic (in drinking water), several aromatic amines (as noted above), painting, work in aluminum and rubber production, Schistosoma haematobium, and ionizing radiation (from atomic bomb explosions and therapeutic radiation). Occupational exposures with limited evidence for causing bladder cancer in humans include coal-tar pitch, soot, dry cleaning and tetrachloroethylene, diesel engine exhaust, hairdressing (perhaps due to hair dyes), printing and textile manufacturing. For environmental exposures, coffee also falls in this category. Artificial sweeteners are likely not associated with risk for bladder cancer, based upon results of epidemiologic studies.

Pathogenesis

Cigarette smoking leads to exposure to a number of carcinogens linked with bladder cancer, including a number of aromatic amines, PAHs, and nitrosamines. Most occupation-­related urinary tract tumors are thought to be caused by contact of the bladder epithelium with carcinogens in the urine. Because of the concentrating ability of the kidney, the bladder is exposed to higher concentrations of these materials than other body tissues. In addition, this exposure occurs over prolonged periods of time in certain areas of the urinary tract, most notably the bladder trigone area. Most of the proved urinary carcinogens are aromatic amines, which may be inhaled, ingested, or absorbed through the skin. Recent data indicate that hereditary polymorphisms of the arylamine N-acetyltransferase gene may play a role in the etiology of bladder cancer by modulating the effect of as well as the interaction between carcinogens, including cigarette smoke and aromatic amines. The risk of cancer appears to be the highest in slow acetylators, suggesting that individual mechanisms of detoxification play an important role in the risk of toxin induced bladder cancer.

Pathology

In addition to bladder cancer, other less common urologic neoplasms, which are sometimes work related, include tumors of the renal pelvis, ureter, and urethra—largely with the same histologic and etiologic features as bladder tumors. All four types usually are considered together as “lower urinary tract cancers” for epidemiologic purposes. More than 90% of bladder tumors are of the transitional-cell type, approximately 6–8% are squamous cell, and 2% are adenocarcinoma. The tumors may be papillary or flat, in situ or invasive, and are graded according to degree of cellular atypia, nuclear abnormalities, and number of mitotic figures.

Multiple genetic changes have been associated with bladder cancer, such as expression of the ras and myc protooncogenes. Mutation of the tumor-suppressor gene p53 is correlated with an increased risk of disease progression.

Clinical Findings

The most common presenting symptom of bladder cancer is hematuria, which occurs in 80% of patients and usually is painless, gross, and intermittent. More than 20% of patients have vesical irritability alone, with increased frequency, dysuria, urgency, and nocturia. The diagnosis of bladder cancer may be made on the basis of urinary cytologic examination, which has been proposed as a screening tool. Up to 75% of patients with bladder cancer have abnormal urine cytology. Most patients undergo excretory urography, which is useful in ruling out upper tract disease and may show a filling defect in the bladder. Definitive diagnosis relies on cystoscopy and transurethral biopsy of the suspicious areas.

Bladder carcinoma that has invaded the muscular wall is potentially lethal and may metastasize even before urinary symptoms bring the patient to a physician. Bladder cancer generally spreads by local extension, through lymphatics, or by hematogenous dissemination. Clinical sites of metastatic disease include the pelvic lymph nodes, lungs, bones, and liver (in decreasing order of occurrence). Once the diagnosis has been confirmed by biopsy, a chest radiograph, radionuclide bone scan, and liver and renal function studies should be done. CT scans are extremely useful in staging. Current staging depends on depth of involvement, nodal involvement, and the presence or absence of distant metastases.

Prevention

Avoidance of cigarette smoking is the most important preventive measure. Prevention of exposure to known carcinogens is the most effective means of preventing occupational bladder cancer. One appealing means of control is screening, and the use of urinary cytological examinations, followed by cystoscopy when indicates, has been suggested for this purpose, in addition to urinalysis to look for microscopic hematuria. Screening of high-risk patients such as certain occupational groups may result in a reduction in the stage of disease at diagnosis, although studies have demonstrated a favorable impact on outcomes or survival.

Treatment

Treatment and prognosis are not different for occupationally-­induced cancer to that of other bladder cancers. Therapy varies with the stage of cancer, although initial treatment for nonmetastatic disease is surgical. Carcinoma in situ and superficial lesions are treated with transurethral resection of the malignant areas, occasionally followed by intravesical immunotherapy or chemotherapy. More advanced disease requires more aggressive surgery, including potentially radical cystectomy. Systemic chemotherapy is reserved for metastatic disease.

Prognosis varies with the stage of the disease.

LIVER CANCER: HEPATIC ANGIOSARCOMA

General Considerations

Angiosarcoma of the liver is a rare tumor with strong epidemiologic links to vinyl chloride and arsenic exposures. Thorotrast (thorium dioxide) exposure was the main nonoccupational risk factor when this agent was used as a radiographic contrast agent from about 1930 to 1955. This cancer occurs most commonly in middle-aged men, with a male-to-female ratio of 4:1. The mean age at presentation is 53 years. Characteristic features of the disease include a long period of asymptomatic laboratory abnormalities, difficulty in diagnosis, and poor response to treatment.

Etiology

Vinyl chloride is the raw material with which the common plastic polyvinyl chloride is made. In 1974, a cluster of cases of angiosarcoma of the liver in men was reported by an alert physician in Louisville, Kentucky. The men were all workers at a local industrial plant that polymerized vinyl chloride. By 1981, 10 cases of hepatic angiosarcoma were identified among 1855 employees older than 35 years of age, with no other cases of angiosarcoma identified in the Louisville area. In one review of 20 patients with angiosarcoma of the liver after vinyl chloride exposure, the mean time from first exposure to development of tumor was 19 years, with a range of 11–37 years. In addition to the Louisville experience, cancer in other patients from plants elsewhere producing vinyl chloride has been noted. Similar hepatic lesions in experimental animals exposed to high concentrations of vinyl chloride also have been observed. Although the evidence is not as striking, angiosarcoma of the liver is also associated with arsenical pesticides, arsenic-contaminated wine, and Fowler solution used medicinally.

Pathology & Pathophysiology

The two distinctive hepatic lesions seen after exposure to vinyl chloride are a peculiar hepatic fibrosis and angiosarcoma. The hepatic fibrosis is characterized by three features: a nonspecific portal fibrosis, capsular and subcapsular fibrosis in a nodular form (the most characteristic lesion), and focal intralobular accumulation of connective tissue fibers. The neoplasm is hemorrhagic and cystic and replaces most of the normal tissue. The carcinogenicity of the vinyl chloride monomer is related to the metabolic formation of reactive metabolites.

Hepatic angiosarcomas caused by Thorotrast and inorganic arsenicals show many of the histologic features observed in the evolution of the hepatic angiosarcoma in the vinyl chloride workers.

Clinical Findings

A. Symptoms and Signs

The symptoms of hepatic angiosarcoma are nonspecific, and some patients may be asymptomatic. Abdominal pain is the most common symptom, usually in the right upper quadrant. Fatigue, weakness, and weight loss are seen in 25–50% of patients. Physical examination reveals hepatomegaly with ascites, jaundice, and, less often, splenomegaly

B. Laboratory Findings

Almost all patients have some abnormality of liver function testing. Most common is elevation of serum alkaline phosphatase. Tests for α-fetoprotein, carcinoembryonic antigen, and hepatitis B antigen are negative.

C. Imaging and Diagnosis

Routine abdominal radiographs and gastrointestinal contrast studies usually are normal. Radionuclide liver scans are abnormal in most patients, but the findings can range from distinct filling defects to nonspecific nonhomogeneous uptake (which can be confused with cirrhosis and splenomegaly). Hepatic arteriograms are the most helpful diagnostic tool, usually demonstrating normal-sized hepatic arteries that may be displaced by tumor, peripheral tumor stain, puddling during the middle of the arterial phase, and a central area of hypovascularity. Hepatic ultrasonography also may demonstrate a hepatic mass. Definitive diagnosis of angiosarcoma is best made by thoracoscopic liver biopsy. Because of the difficulty in making the diagnosis and rapid clinical deterioration, more than 50% of hepatic angiosarcomas are diagnosed only after death.

D. Screening Tests

Medical Surveillance & Prevention

As part of the U.S. OSHA standard for vinyl chloride, exposed employees are to receive periodic testing, including history and physical examination and liver function tests. Further testing such as imaging tests (ultrasound, liver scan), angiography, and biopsy should be performed as indicated for significant abnormalities. Though OSHA mandates medical surveillance for arsenic-exposed workers, it is not directed toward detection of angiosarcoma. Preventive measures for angiosarcoma include stringent limitations on employee exposure to vinyl chloride and arsenic compounds.

Treatment & Prognosis

Partial hepatectomy with intent to cure is possible in only a very limited number of patients because of extensive fibrosis in the uninvolved liver. No forms of treatment, including radiation, chemotherapy, or liver transplantation have been shown to improve survival.

Overall survival usually is measured in months, with the median survival approximately 6 months and only a small percentage of patients surviving 2 years. The major cause of death is irreversible, rapidly progressive hepatic failure.

SKIN CANCER (NONMELANOMATOUS)

Occupations at Risk

• Solar (ultraviolet) radiation

  • Outdoor workers

• PAHs

  • Workers exposed to coal-tar, coal-tar pitch and soot, such as roofers

  • Workers exposed to untreated or mildly treated mineral oils, for example, metal workers

• Workers exposed to shale oil

• Arsenic

  • Arsenical pesticide production and use

  • Copper, lead, zinc smelting

  • Sheep-dip manufacturers (contamination of drinking water)

• Ionizing radiation

  • Uranium miners

  • Health workers

General Considerations

Neoplastic diseases of the skin are commonly divided into melanoma and nonmelanomatous skin cancer, which consists mainly of basal cell and squamous cell carcinoma. Nonmelanomatous skin cancer (NMSC) is currently the most common form of cancer in the white population of the United States, accounting for one-third of all diagnosed cases of cancer. Although the dominant risk factor for NMSC (ultraviolet light) has been established, epidemiologic study of skin cancer has been limited. Nonmelanomatous skin cancer has an excellent prognosis, with 96–99% cure rates, making death certificate reviews useless.

There is an incorrect perception that skin cancer other than melanoma is a trivial disease. In addition, patients are rarely hospitalized, with the result that they are commonly not included in cancer registries. Because of failure to register or record skin cancers, much of the data on incidence are from surveys conducted many years ago. It is projected that more than 80,000 Americans will develop NMSC each year. Basal cell cancer is more than three times as common as squamous cell cancer.

Globally, NMSC is the most common form of cancer, more common than all other cancers combined. Most of these cancers are due to solar (ultraviolet) radiation and occur at high rates in people who work or play in the sunlight or use tanning booths and tanning lights.

Etiology

The primary causes of skin cancer in industry include ultraviolet radiation (UV), PAHs, arsenic, and ionizing radiation. The information presented below refers primarily to NMSC. An increased risk of melanoma is associated with UV light exposure.

A. Uv Radiation

Clearly, the major risk factor for skin cancer in lightly pigmented persons is radiation from the sun. The experiment of nature in which different intensities of UV radiation occur at different global latitudes has provided the opportunity for many epidemiologic studies to show an increased incidence of NMSC in whites at latitudes closer to the equator. The earliest realization that excess sun exposure leads to skin cancer was made on the basis of occupation in 1890, when Unna described changes of the skin of sailors, including skin cancer that resulted from prolonged exposure to the weather.

There are approximately 4.8 million outdoor workers in the United States, with certain occupations at greater risk, such as those in agriculture and professional sports. In experimental animals, the most carcinogenic wavelength is in the range of 290–300 nm (sunlight does not include wavelengths lower than 290 nm). The actual carcinogenic spectrum for humans is unknown. It is also notable that in experimental animals, a variety of foreign substances, including phototoxic chemicals (eg, coal tar, methoxsalen), chemical carcinogens (eg, benzo[a]pyrene), and nonspecific irritants (eg, xylene), under suitable conditions augment UV carcinogenesis.

B. Polycyclic Aromatic Hydrocarbons

Although chemical carcinogenesis of the skin does not seem to be nearly as frequent a cause of NMSC as UV radiation, it was described more than a century earlier. Percival Pott described the increased incidence of scrotal cancer in chimney sweeps in 1775, but it was not until the 1940s that a polycyclic aromatic hydrocarbon (PAH), benzo[a]pyrene, was shown to be a constituent of soot. These hydrocarbons have the ability to induce skin cancers in laboratory animals, and mixtures of them are found in coal tar, pitch, asphalt, soot, creosotes, anthracenes, paraffin waxes, and lubricating and cutting oils. Exposures to untreated or mildly treated mineral oil containing PAHs have been linked to skin and scrotal cancers among mule spinners, wax pressmen, metal workers exposed to poorly refined cutting oils, and machine operators using lubricating oils. Latent periods between exposure to polycyclic aromatic hydrocarbons and skin cancer vary from about 20 (coal tar) to 50 years or more (mineral oil).

C. Arsenic

Arsenic causes cancer in experimental animals and is a well-recognized human carcinogen. Skin tumors associated with arsenic occur following ingestion, injection, or inhalation, as well as from skin contact. Medicinal inorganic arsenicals and arsenic in drinking water are the sources most commonly implicated. Recent detailed studies in Taiwan established that use of well water with high arsenic concentrations resulted in skin cancer, with a dose-response relationship. An estimated 1.5 million workers in the United States are exposed to inorganic arsenic in such diverse trades as copper and lead smelting, the metallurgical industry, and the production and use of pesticides; however, skin tumors attributed to occupational arsenic exposure are very uncommon. It is thought that some of the cases cited in the literature of agricultural workers with arsenic-induced skin cancers may be the result of other carcinogenic influences, such as sunlight and tars. The simultaneous presence of arsenical hyperkeratoses or hyperpigmentation, which occurs at lower exposure levels, strongly implicates arsenic as the etiologic agent in an individual with NMSC. In addition, cancers tend to be multiple and occur in younger patients than those attributable to UV light.

D. Ionizing Radiation

Ionizing radiation is as carcinogenic for skin as it is for many other tissues. Roentgen radiation–induced skin carcinoma was first reported in 1902, shortly after the discovery of x-rays, in those who worked the machines. There was a definite excess in skin cancer deaths among radiologists in the period 1920–1939, and an excess risk also has been found for uranium miners. Patients receiving radiation for acne, tinea capitis, and facial hair in the past had an increased risk of invasive skin cancers. The latent period for radiation-induced skin cancers varies inversely with the dose, with the overall range from 7 weeks to 56 years (average 25–30 years), and the skin cancers often occur in areas with chronic radiation dermatitis. Although epidemiologic studies do not give reliable data on dose-response relationships, the risk from exposures under 1000 cGy appears to be small, and skin cancer may be induced by dose equivalents of 3000 cGy. There are now strict controls on industrial and occupational exposure to ionizing radiation, and currently, it appears that ionizing radiation is not responsible for much cutaneous carcinogenesis.

E. Other Factors

Other risk factors for the development of NMSC include chewing tobacco or betel nuts, where squamous cell cancer of the lip and oral cavity have been described. Chronic irritation or inflammation is thought to induce these cancers. Patients with either primary or secondary (long-term immunosuppressive therapy) immunodeficiencies are at increased risk for skin cancer. Several genetically inherited syndromes such as xeroderma pigmentosum and albinism are associated with increased susceptibility to skin cancers.

Pathophysiology

Early studies elucidated the two-stage theory of carcinogenesis. They found that a single application of a potent carcinogen such as benzo[a]pyrene applied in a quantity insufficient to cause tumors allowed tumor development after subsequent application of croton oil, which by itself produced no tumors at all. The authors theorized that the production of a tumor was initiated by the carcinogen but that its subsequent development could be promoted nonspecifically. It appears that initiation is permanent and irreversible, but promotion, up to a point, is reversible.

UV light fits into this theory of chemical carcinogenesis in that it appears to be both an initiator and a promoter for carcinoma of the skin. Two major effects of UV radiation on the skin that seem likely to be responsible for the carcinogenic effects are photochemical alteration of the DNA and alterations in immunity. Certain immunologic defects, both in skin and in lymphocytes, can be induced by UV radiation. Exposure to UV light depletes the dermis of Langerhans cells and renders it unable to be sensitized to potent allergens. Alterations at the level of DNA are also thought to be responsible for ionizing radiation–induced skin cancers.

Pathology

The histologic types of skin lesions associated with sun exposure include solar keratoses, basal cell epitheliomas, squamous cell carcinomas, keratoacanthomas, as well as malignant melanomas. Solar keratoses contain morphologically cancerous cells, but they are considered premalignant because invasion is limited to the most superficial part of the dermis. Approximately 13% of all solar keratoses develop into squamous cell carcinomas, but these are rarely aggressive. The estimated incidence of metastases from all sun-induced squamous cell carcinomas is 0.5% or less. Almost all squamous cell carcinomas in whites occur in highly sun-exposed areas, but 40% of basal cell epitheliomas occur on shaded areas of the head and neck.

Regardless of the source of exposure, certain features are common in all cases of arsenic-induced skin cancers. Punctate keratoses of the palms and soles and hyperpigmentation are seen frequently. The skin tumors are of several types. Squamous cell carcinomas arise either from normal skin or from keratoses. Basal cell epitheliomas, including multiple superficial squamous cell and basal cell epitheliomas, as well as areas of intraepidermal (in situ) squamous cell carcinoma (Bowen disease), have been described. Multiple tumors, most of which are found on unexposed areas, are the rule. Cancer of the scrotum, which is seen following topical exposure to PAHs, is rare.

Early radiation workers with heavy exposure from uncalibrated machines developed predominantly squamous cell carcinomas, found mainly on the hands and feet and occasionally on the face. More recently, basal cell cancers have been described following repeated occupational exposures. Radiation-related tumors usually arise in areas of chronic radiation dermatitis, and whether they can occur on clinically normal skin is a matter of dispute.

Clinical Findings

Basal cell epithelioma frequently presents as a nodular or nodular-ulcerative lesion on the skin of the head and neck and only 10% of the time on the skin of the trunk. It is much less common on the upper extremities and very uncommon on the lower extremities. The lesion generally is smooth, shiny, and translucent, with telangiectatic vessels just beneath the surface. It is usually not painful or tender, even with ulceration, except when crusting or bleeding is seen with minor trauma. Basal cell carcinomas rarely metastasize, but they can invade widely and deeply, extending through the subcutaneous tissue to involve neurovascular structures and occasionally erode into bone.

Squamous cell carcinoma presents first in a premalignant stage characterized by actinic keratosis, a rough, reddened plaque on sun-exposed skin. There is then an in situ stage, which appears as a well-demarcated, slightly raised erythematous plaque with more substance and scaling than actinic keratosis. Squamous cell cancers arising in sun-exposed areas of the body tend to be on the most highly irradiated areas, such as the tip of the nose, the forehead, the tips of the helices of the ears, the lower lip, or the backs of the hands. Metastases are more common than from basal cell cancer, and squamous cell cancers on mucosal membranes metastasize more frequently than do those found on the skin surface.

Prevention

The most important step in prevention of occupation-related skin cancers is avoidance of UV light. This is especially true for workers who are more susceptible to UV light, such as those with fair complexions or with certain hereditary diseases (eg, albinism and xeroderma pigmentosum).

Protective clothing, such as wide-brimmed hats and long sleeves, is the most effective barrier to UV radiation exposure in outdoor workers. Sunscreens that provide protection in the UVA and UVB spectrum should be used daily. The effectiveness of sunscreens in preventing nonmelanomatous skin cancer and melanoma is unknown, though their effectiveness for avoidance of erythema has been proved. Periodic examinations are recommended to detect the presence of malignant and premalignant skin lesions among those at risk.

The incidence of scrotal cancer is now rare because of preventive measures. If possible, a noncarcinogenic material should be substituted for a carcinogenic one. Good personal hygiene should include compulsory showering and changing of clothes when entering and leaving the plant, as well as washing of exposed skin after leaving contaminated areas. Isolated or closed-system operations, protective clothing, and employee education are also critical in avoidance of skin cancer induced by PAHs.

Currently, the maximum allowable dose equivalent of ionizing radiation for occupational exposure to the skin is 30 rems in any year, except that forearms and hands are allowed 75 rems in any year (because there is little red marrow in the forearms and hands). These recommendations are based mainly on avoidance of hematologic disease and may need to be revised in order to prevent skin cancer. Exposure can be limited further by the use of shielding devices such as lead gloves and aprons.

Diagnosis & Treatment

Biopsy is necessary in all cases of suspected skin carcinoma. Treatment for occupationally-induced skin cancers is not different from other skin cancers.

LEUKEMIA

Occupations at Risk

• Worker exposed to ionizing radiation, including some exposed radiologists, nuclear industry workers, and ­military personnel

• Workers exposed to benzene

• Workers in the rubber manufacturing industry (leukemia and lymphoma)

General Considerations

The two major forms of leukemia that have been linked to occupation are acute nonlymphocytic leukemia (ANLL), including myelodysplasia or preleukemia, and chronic myelogenous leukemia (CML). The acute leukemias are malignant diseases of the blood-forming organs characterized by a proliferation of immature blood cell progenitors in the bone marrow and other tissues. Together with replacement of the normal marrow with leukemic cells, there is a diminished production of normal erythrocytes, granulocytes, and platelets. Acute leukemias are classified morphologically by reference to the predominant cell line involved as lymphocytic and nonlymphocytic forms. Nonlymphocytic leukemias are further classified as de novo (no underlying cause known and without preexisting myelodysplasia) or secondary (known cause such as chemical exposure or preexisting myelodysplasia or chronic leukemia). The incidence rates for all types of leukemia vary widely by geographic areas and ethnic groups, but, in North America and Europe, they vary from about 8–12 cases per 100,000 person-years in men to about 5–8 cases per 100,000 person-years in women, with ANLL accounting for about 3–4 cases per 100,000 person-years in men and 1–2 cases per 100,000 person-years in women. Thus, both total leukemias and ANLL are more common in men. The incidence of ANLL increases with age with the highest rates above age 50.

Chronic leukemias are classified as lymphocytic and myelogenous; only CML has been reported as an industrial disease. CML is a neoplastic disease resulting from the development of an abnormal hematopoietic stem cell. There is excessive growth of the blood cell progenitors in the marrow, which initially function as normal hematopoietic cells. The leukemic cells gradually undergo further malignant transformation, with loss of the ability to differentiate in the later stages of the disease, with the resulting development of acute leukemia and death. In the early stages of the disease, large numbers of mature and immature granulocytic cells accumulate in the blood, and extramedullary hematopoiesis produces gross enlargement of the liver and spleen. CML accounts for approximately 20% of all deaths from leukemia in the Western world, with an incidence that, unlike other forms of leukemia, has not been increasing recently. Although rare cases are reported in infants, most patients with CML are age 25–60 years, with a median age of about 45 years.

Etiology

The cause of most cases of human leukemia is unknown. As in the case of most other cancers, it is probable that no single factor is responsible. Most cases are thought to result from the interaction of host susceptibility factors, chemical or physical injury to chromosomes, and in animals and presumably in humans, incorporation of genetic information of viral origin into susceptible stem cells. While certain occupational exposures, such as benzene and ionizing radiation, cause leukemia, most cases of leukemia are idiopathic and cannot be attributed to recognized causes. A recent analysis, using information about the extent of exposure to three known and suspected human leukemogens—benzene, ionizing radiation, and ethylene oxide—and relative risk estimates from epidemiologic studies, suggested that these exposures might account for only 1–3% of all cases of leukemia.

A. Radiation

Radiation remains the most conclusively identified leukemogenic factor in human beings. The earliest evidence began to accumulate soon after the discovery of x-rays, which were used mainly in the medical workplace; thus radiologists, radiation therapists, and radiation technicians were all at risk. Several studies showed an excess risk of leukemia among radiologists (approximately nine times that of other physicians) during the years 1930–1950, with a latency period of about 18 years. With the institution of dose limits, careful monitoring, and adequate shielding since that time, this excess risk has decreased significantly and should be eliminated.

The data from Hiroshima and Nagasaki atomic bomb survivors leave little doubt that the incidence of leukemia is increased following exposure to mixed gamma and neutron radiation and that the response is dose dependent. The risk of leukemia is increased in populations exposed to ionizing radiation at doses as low as 50–100 cGy. Between 100 and 500 cGy, there is a linear correlation between dose and leukemia incidence. The data suggest that the risk of leukemia is increased at a rate of 1–2 cases per million population per year per centigray. Maximal risk occurs approximately 4–7 years after exposure, and an increased risk has been seen in Japanese people as long as 14 years after exposure.

Whole-body exposure to radiation in single doses results in suppression of marrow growth, and a single whole-body dose of more than 400 cGy usually is fatal in humans. In sublethal exposure, cytopenias may occur, which gradually recover but indicate significant damage to the marrow precursor elements. Patients are then at risk to develop leukemia with a delay between exposure and disease of 8–18 years. Following radiation exposure, both acute and chronic myelogenous leukemia may occur. The specific rates per 100,000 for people within 1500 m (4921 ft) of the hypocenter are 8.1 for ANLL, 25.6 for chronic myelogenous leukemia, and 21.7 for acute lymphocytic leukemia. Chronic lymphocytic leukemia has not been associated with ionizing radiation exposure.

Workers at risk secondary to exposure to ionizing radiation include military personnel in the vicinity of nuclear tests, uranium miners, and workers in nuclear power plants. Approximately 250,000 troops are estimated to have been present at multiple detonations of nuclear devices carried out by the United States from 1945 to 1976. In 1976, more than 3000 men exposed at the 1957 nuclear test explosion “Smoky” were studied, and a significant excess of leukemia was discovered. A review of death certificates of former workers at the Portsmouth Naval Shipyard in Portsmouth, New Hampshire (where nuclear submarines are repaired and refueled), revealed an observed-to-expected ratio of leukemia deaths of 5.62 among former nuclear workers.

B. Benzene

Certain chemicals (eg, chemotherapeutic agents) are known to be toxic to marrow cells, and many of these also possess leukemogenic potential. Occupational evidence of leukemogenicity is strongest for benzene, where epidemiologic studies have shown significant increases in leukemia in workers with past exposure to benzene. Benzene has been known for more than a century to be a powerful bone marrow poison, leading to aplastic or hypoplastic anemia. In addition, benzene causes leukemia, with fatal cases of leukemia outnumbering those of aplastic anemia. In 1928, the first case of acute leukemia was reported in a worker from a plant in which there was very heavy exposure to benzene. Benzene is a cyclic hydrocarbon obtained in the distillation of petroleum and coal tar. It is used widely in chemical synthesis in many industries, in the manufacture of explosives, and in the production of cosmetics, soaps, perfumes, drugs, and dyes. Benzene once was used in the dry-cleaning industry, but that is no longer the case. In addition, gasoline contains benzene, typically at about 1% in American gasoline but sometimes up to 5%.

An estimated 2 million workers in the United States have exposure to benzene. In a cohort of rubber hydrochloride workers with significant exposures to benzene, there was an overall 2.5 to 3-fold increase in risk of leukemia, with progressively higher risks with increasing cumulative exposure (in ppm-years). Many other studies, including several undertaken in the shoe manufacturing industry, have shown an increase in the risk of leukemia in workers with exposure to benzene. Studies also suggest a link between benzene exposure and an increased risk of chronic lymphocytic leukemia, multiple myeloma and non-Hodgkins lymphoma. In 2012, IARC concluded that benzene causes ANLL (as well as myelodysplastic syndromes), with limited evidence in humans for a causal association with acute lymphocytic leukemia and chronic lymphocytic leukemia.

C. Other Agents

Chemicals other than benzene are known or suspected to cause leukemia In 2012, IARC concluded that formaldehyde causes leukemia, supported particularly by studies demonstrating increased risk for professional workers, for example, embalmers and pathologists. The evidence is stronger for myeloid leukemias. Similarly, IARC concluded in 2012 that 1,3-butadiene, a monomer used in the production of synthetic rubber and polymers, causes hematolymphatic malignancies, with the evidence stronger for leukemia than for lymphoma, but the epidemiologic data in this area are incomplete. There is limited evidence according to IARC that ethylene oxide causes leukemia in humans. Exposure to ethylene oxide, used as a sterilant and in chemical processing, has been associated with an increased risk of lymphatic and hematopoietic cancers (specifically lymphoid tumors, ie, non-Hodgkin lymphoma, multiple myeloma, and chronic lymphocytic leukemia). Among agents to which exposure is largely nonoccupational, tobacco smoking causes myeloid leukemia in humans, perhaps related to the significant presence of benzene, formaldehyde, and 1,3-butadiene in cigarette smoke. Treatment with a variety of chemotherapeutic agents has been causally associated with an increased risk of leukemia, typically within 2–5 years of initiation of chemotherapy. There is sufficient evidence in humans that some alkylating agents, such as busulfan, chlorambucil, cyclophosphamide, melphalan, and thiotepa, cause leukemia. There is a known synergistic interaction of radiation therapy and treatment with alkylating chemotherapeutic drugs, used to treat Hodgkin disease and other malignancies, resulting in a significant increase in the risk of subsequent leukemia. Some immunosuppressive agents, such as cyclosporine and azathioprine also cause leukemia in humans. IARC has concluded that there is limited evidence that exposure to extremely low frequency (ELF) magnetic fields, for example, from power lines, and to parental smoking causes leukemia in children.

Pathophysiology

A. Ionizing Radiation

The effects of radiation on human tissue depend on multiple factors, such as type of radiation, dose of radiation, length of exposure, body part exposed, and oxygen content of the exposed tissue. Damage secondary to radiation is greatest in rapidly dividing cells such as bone marrow stem cells, epithelial cells, and gamete-forming cells. The mechanism of radiation-induced injury at the cellular level involves direct and indirect damage to nucleic acids and proteins. DNA is a radiosensitive target, with even minor molecular damage resulting in profound effects on the cell and the organism. Radiation-induced molecular damage may be so severe that the cell no longer functions, and cell death results. Cells exposed to radiation may survive with no effects (if only a small number of nonessential molecules are affected) or may survive with altered structure and function. If the alteration is within the DNA, clinical disease may not appear until after a latency period. Cancer induction appears to depend on an interaction of defective cellular repair and damage to the cell's regulator genes.

B. Benzene

Benzene toxicity may present as an acute illness or as a chronic disease developing up to 30 years after exposure. Chronic or recurrent exposure to concentrations of benzene exceeding 100 ppm (320 mg/m³) leads to a very high incidence of cytopenias. When the exposure ends, there is usually spontaneous remission. Among workers who have been exposed to atmospheric concentrations of benzene in excess of 300 ppm for at least 1 year, as many as 20% will acquire pancytopenia or aplastic anemia. Aplastic anemia generally occurs in subjects while they are still exposed to high concentrations of benzene; leukemia may occur at the same time or after cessation of exposure.

There are likely multiple mechanisms by which benzene induces ANLL and myelodysplastic syndromes (MDS). In experimental studies, including in human cells, benzene induces chromosomal aberrations and mutations. Workers exposed occupationally to benzene have exhibited chromosomal aberrations in peripheral lymphocytes. Chromosomal aberrations occur in individuals with leukemia, in therapy-related and benzene-related leukemias, and in spontaneous (de novo) leukemias. The leukemogenicity of benzene requires its metabolism to other compounds in the liver and bone marrow, such as hydroquinone and 1,4-benzoquinone, which may be the active carcinogenic metabolites. Like certain chemotherapeutic drugs, known as topoisomerase-II inhibitors, benzene metabolites may act by inhibition of this enzyme that is responsible for the maintenance of proper chromosome structure. Hematotoxic effects from high-dose exposure to benzene, as discussed above, secondary immune system dysfunction, and epigenetic changes induced by benzene may also contribute to its leukemogenicity.

Clinical Findings

The clinical findings, including symptoms, signs, and laboratory findings, in occupationally-or environmentally-induced leukemias are not different from those observed in de novo leukemias. However, certain agents produce preceding toxic effects, which may be present prior to the development of leukemia.

1. Radiation. As noted earlier, 300–400 cGy of whole-body radiation is lethal in humans. Sublethal exposures will cause symptoms of nausea and vomiting, after which bone marrow suppression occurs. Thrombocytopenia, anemia, and neutropenia will develop, with their attendant symptoms. The development of leukemia occurs after a variable but relatively short latency period. When the disease progresses, symptoms are identical to acute leukemia.

2. Benzene. Benzene-induced hematotoxicity may lead to anemia, thrombocytopenia, and leukopenia, with attendant symptoms and clinical findings; however, such manifestations may not precede the development of benzene-induced leukemia. The presentation of leukemia due to benzene exposure is no different than that of de novo leukemia.

Prevention

Avoidance of exposure to potential leukemogens, including ionizing radiation, benzene, and cigarette smoking, should reduce the occurrence of leukemia secondary to these agents.

A. Radiation

X-rays were discovered by Roentgen in 1895, and by 1902, the basic principles of radiation protection already had been elaborated: to minimize dose by reducing the time of exposure and by using shielding and distance. Since 1928, the International Council on Radiation Protection (ICRP) and the National Council on Radiation Protection have defined acceptable levels of radiation exposure for workers. The concept of dose equivalent or rem (Roentgen-equivalent man) is used because the same amounts of absorbed radiation energy can produce different levels of damage depending on the type of radiation present. Acceptable exposures for different organs vary, with a maximum permissible dose ranging from 5 rems of whole-body exposure to 30 rems of skin or bone exposure.

B. Benzene

Regulation of benzene exposure began in 1926. In 1974, NIOSH published a recommended standard based on the evidence for hematotoxic effects: 10 ppm as an 8-hour time-weighted average (TWA), with a ceiling limit of 25 ppm. The current OSHA 8 hour time-weighted average workplace exposure limit is 1 ppm. Even this “acceptable” level remains an area of controversy, in that quantitative risk assessment analysis suggests that the risk of leukemia mortality from a working lifetime of exposure at this level would be about 1.7-fold compared to the background risk. OSHA requires periodic medical surveillance annually, including a complete blood count, for workers exposed to benzene above the 8-hour TWA action level of 0.5 ppm.

Treatment & Prognosis

There have been recent major advances in the treatment of acute leukemias with the use of combination chemotherapy and bone marrow transplantation. The treatment of occupationally-­induced leukemia is essentially the same as that for spontaneous leukemias.

OTHER CANCERS

A number of agents are known to cause certain other cancers in humans. Formaldehyde and wood dust are known to cause nasopharyngeal carcinoma in humans. Work in the rubber production industry is causally associated with stomach cancer, as well as leukemia, lymphoma, and lung and urinary bladder cancers. Ionizing radiation, specifically x-and γ-radiation, is known to cause a number of cancers in humans—salivary gland, esophagus, stomach, colon, lung, bone, skin (basal cell), female breast, kidney, urinary bladder, brain and central nervous system, thyroid, and leukemia (excluding chronic lymphocytic leukemia), although not all of these causal associations were observed in occupational epidemiology studies. Many other cancers are reported to be associated with specific occupational or environmental exposures in humans, most with limited evidence based upon epidemiologic studies. Semiconductor workers demonstrate excess risks for non-Hodgkin lymphoma, leukemia, brain tumor, and breast cancer. An increased incidence of renal cell cancer has been reported in some workers, with limited evidence of renal carcinogenicity for exposure to arsenic and cadmium compounds. There is limited evidence that exposure to asbestos causes cancers of the pharynx, stomach, and colorectum.