One major source of epidemiological data to evaluate health risks from radiation exposure is the Life Span Study (LSS) of atomic bomb survivors, which found a proportional relationship between cancer risk and exposure dose. Although numerous findings from the study provide scientific evidence for increased cancer risk from radiation exposure, radiation-associated cancer risk remains unclear at low-dose ranges under 100 mSv [11]. Studies related to the Chernobyl accident also demonstrated cancer risks from radiation exposure, especially an increase in thyroid cancer among children with high thyroid-absorbed doses. Except for this result, however, no definitive conclusions have been made regarding other cancers among Chernobyl residents who were exposed to low doses of radiation [12–15]. Some studies that have investigated the health of Chernobyl workers exposed to prolonged low to medium doses of radiation (average effective dose of 100 mSv) have reported increased risks of cancer as well as non-cancer diseases, such as cataracts and cardiovascular diseases [16–21]. However, due to screening effects (e.g., medical examinations) and limited sample sizes, it is difficult to draw definitive conclusions from these studies. Thus, it remains necessary to continue follow-ups of these cohorts with accurate assessments of exposure dose, health outcomes, and confounding factors [14, 22].

A major distinction between occupational exposure and accidental exposure is the period and dose levels of exposure. Whereas accidental exposure usually involves a single large exposure (acute), occupational exposure involves protracted exposures to low-dose radiation (chronic). A number of epidemiological investigations have been conducted among radiation workers in individual countries as well as in large-scale international cohort studies, and the cancer risk from occupational exposure to radiation continues to be updated. A few studies have reported elevated risks of cancer with statistical significance. One of the largest occupational studies in radiation workers is the 15-country collaborative study, which included 407,391 nuclear industry workers over 5.2 million person-years of follow-up [23]. In this study, an elevated risk of all-cancer mortality was observed, with an excess relative risk (ERR/Sv) of 0.97 (95% CI: 0.27, 1.8). However, this risk diminished after excluding data from workers in Canada, in whom the dose measurement was uncertain, and the observed risk was no longer significant. As a follow-up to the 15-country collaborative study, risks of leukemia and lymphoma were investigated among 308,297 radiation workers in France, the U.K., and the U.S. [24]. The association between exposure dose and cancer mortality was statistically significant with an ERR of 2.96 per Gy (90% CI: 1.17, 5.21) for leukemia, excluding chronic lymphocytic leukemia (CLL). The highest ERR/Gy of 10.45 (90% CI: 4.48, 19.65) was found for chronic myeloid leukemia, indicating a strong association between leukemia mortality and protracted low-dose radiation exposure [24]. Although the ERR of leukemia, excluding CLL, was not attenuated for doses less than 100 mGy, the 90% CIs were too wide to make a definitive conclusion about the low-dose ranges.

Cohort studies of the Mayak nuclear complex workers also reveal an elevated cancer risk [25–27]. Because this cohort had a broad range of cumulative doses due to high exposure levels during the early stages of the facility operation, the dose-response relationship had a degree of precision that is rarely observed in other studies of radiation workers, who are usually exposed to low-dose levels [26]. In addition to the Mayak cohort studies, other studies of radiation workers have reported increased risks of certain types of cancer, such as leukemia (excluding CLL), esophageal cancer, and lung cancer [28–31]. However, risks for individual cancer sites are inconsistent across most radiation epidemiological studies, and many studies do not find statistically significant results. Cancer risks from major health studies in nuclear workers are summarized in Tables 3 and 4.

Aircrews, such as pilots and flight attendants, are exposed to cosmic radiation. Although aircrews are not included in the national registry for radiation workers in Korea, they should be considered radiation workers and monitored for radiation exposure and health risks, because they are exposed to similar or even higher levels of radiation compared to common radiation-related occupations, such as nuclear workers and radiologists. An average effective dose in an aircrew flying over the poles at high latitudes is estimated to be 2–5 mSv/year, which may reach a cumulative dose of about 75 mSv over the course of a worker's career [32]. Many interesting health studies have been conducted in aircrews based in Nordic countries, the U.S., and Canada. These studies have reported higher risks of breast cancer, prostate cancer, brain cancer, skin cancer, non-Hodgkin's lymphoma, and acute myeloid leukemia among aircrews, compared with the general population [33–37]. However, given that no demonstrated dose-response relationship was found, these elevated cancer risks do not imply a causal relationship with radiation exposure.

In summary, despite the existence of several epidemiological studies in radiation workers, cancer risks from occupational exposure, especially for doses less than 100 mSv, remain poorly understood due to uncertainty about exposure dose and confounding factors, possible misclassification of health outcomes, and limited statistical power [24, 38].

Several criteria should be met cumulatively to obtain the recognition of radiation-related occupational cancer. These criteria are well described in Notification No. 2014-78 of the NSSC regarding regulations on occupational disease among radiation workers. The major criteria are summarized here. First, cancer must be eligible for radiation-induced cancer: liver cancer, except those cancers that involve cirrhosis or the hepatitis virus (e.g., types B or C); thyroid cancer; ovarian cancer; brain cancer; multiple myeloma; colon cancer; bladder cancer; Non-Hodgkin lymphoma; esophagus cancer; kidney cancer; female breast cancer; stomach cancer; pancreatic cancer; salivary gland cancer; lung cancer; skin cancer; and leukemia, excluding CLL. Several cancers are not recognized as radiation-related occupational cancer, namely Hodgkin's lymphoma, melanoma, malignant mesothelioma, and CLL. These classifications are based mainly on findings from epidemiological studies. For example, mesothelioma is a well-known asbestos-related cancer, and approximately 80–90% of mesotheliomas are caused by long-term inhalation of asbestos [39]. As another example, whereas leukemia is a radiation-sensitive cancer, CLL has not been associated with radiation exposure in most epidemiological studies (Table 6). Second, radiation exposure must be identified by dose assessment or circumstantial evidence. For the assessment of exposure levels, dose records from the NDR are considered a priority. Additional assessments, such as dose reconstruction, are necessary for unclear or omitted cases. Third, a latent period (i.e., time between the first exposure and the appearance of a tumor) must be considered as sufficient or relevant to cancer incidence. For example, solid cancer can be recognized as occupational cancer only if the cancer occurs at least 5 years after the first exposure, whereas leukemia (excluding CLL) can be recognized as occupational cancer only if the cancer occurs at least 2 years after the first exposure and within 20 years after the last exposure. Lastly, the probability of causation (PC), which is defined as the probability that a cancer was caused by occupational radiation exposure during employment, determines whether an individual's cancer is “at least as likely as not” (i.e., a PC of 50% or greater) related to occupational exposure [40]. The PC is calculated as cancer risk attributable to radiation exposure divided by the sum of baseline cancer risk to the general population plus the risk attributable to radiation exposure, considering personal information (e.g., birth year, gender), medical information (e.g., type of cancer, year of diagnosis), and exposure information (e.g., age at exposure, radiation dose). Given that a threshold dose for cancer has not been identified yet, risks of cancer are stochastic effects, and therefore the PC is an important objective measure to assess a causal relationship with radiation exposure. Based on the current guidelines from the NSSC, PCs for solid cancer and leukemia should exceed 50% and 33%, respectively. However, PC includes an estimation error due to uncertainties about dose and the dose rate effectiveness factor (DDREF), as well as a risk transfer error between different populations; therefore, there exist cases with a PC less than 50% that are fully or partially recognized as occupational cancer in civil litigation.

The recognition criteria for radiation-related occupational cancer are based on scientific evidence. However, ultimately, their acceptable range and levels are often affected by several factors unrelated to science, such as social, cultural, and economic factors. In particular, complex elements, such as the social status of the radiation-related occupation, number of workers, cancer incidence rate in the general population, specific risk perceptions of certain cancers, and economic wealth, factor into the recognition of occupational cancer. For these reasons, recognition criteria differ across countries or even across occupations within the same country. For example, CLL is generally excluded as an occupational cancer due to lack of scientific evidence regarding radiation-induced CLL. However, CLL is considered as being potentially caused by radiation, and hence, as potentially compensable under the Energy Employees Occupational Illness Compensation Program Act of 2000 (EEOICPA), effective March 7, 2012 in the U.S. In addition, eligible cancer sites differ according to occupation (e.g., special exposure cohort, uranium workers, energy employees, soldiers). Regarding the PC, the EEOICPA applies the upper 99% credibility (i.e., confidence) limit of the PC instead of the point estimate (i.e., 50th percentile) to the determination of causation between exposure and cancer, which provides each worker with the benefit of the doubt before a final compensation decision is made. In France, the criteria for recognition or compensation for cases not relevant to the regulatory guidelines are more relaxed through individual case assessments, meaning that cases with non-radiogenic disease or an inadequate latent period can be possibly compensated when the disease is obviously related to occupational exposure and the disability from the disease is over 25% [41]. Major recognition criteria of Korea and other countries are compared in Table 7.