Radiation Safety and Protection in the Interventional Fluoroscopy Environment

Stochastic Effects

The Oxford English Dictionary defines stochastic as “randomly determined; having a random probability distribution or pattern that may be analyzed statistically but may not be predicted precisely.” The need for radiation protection arises from the detrimental effects, including those that are stochastic or random, that can occur due to exposure to ionizing radiation. Our main knowledge regarding radiation effects has been obtained from studying the survivors of the atomic bombs (known as the Life Span Study [LSS]). However, the doses from these acute exposures were delivered much faster (higher dose rate), were high linear energy transfer radiation, and in a higher dose range (much greater than 100 millisieverts [mSv]) than is typically delivered in a diagnostic or interventional medical imaging procedure. Radiation-induced cancer is one of the main outcomes studied in the LSS. These cancers can appear up to years after the exposure (latent period) and are stochastic in nature. When translating (by extrapolation and scaling to take account of the lower dose rate and dose) the results from the atomic bomb survivors, the dose response for solid cancer incidence follows a linear function with leukemia incidence better described by a linear-quadratic fit.

As interventionalists we are chronically exposed to low-dose radiation. This is a very different phenomenon from the atomic bomb survivor population. This leads to controversy. There are polarized schools of thought about the risk of chronic low-dose radiation. On one end of the spectrum are those who say no exposure is safe. On the extreme far end are people who believe that radiation exposure promotes cell mutation that may result in beneficial evolution of humanity. This phenomenon is called hormesis . Large national studies from developed countries like South Korea report the long-term risks of chronic low-dose radiation in radiographers, radiologists, industrial radiographers who use cobalt sources to x-ray steel in aircraft turbine blades and buildings, and nuclear power station workers. Other international studies pool data on survivors of Chernobyl, employees of the nuclear power industry, and those exposed during nuclear accidents such as Fukushima. These studies include hundreds of thousands of workers with occupational exposure and their long-term follow-up. From these studies the concept of the linear no threshold model has evolved. This essentially conveys that the risk of cancer increases approximately with dose, and risk exists even at the lowest doses. At higher doses where the likelihood of multiple interactions in a cell increases, the shape of the response curve is more likely to be linear-quadratic. For low linear energy transfer radiation and low dose rates there is a good chance of repair between radiation insults.

At the low end of the dose range encountered in medical imaging, the linear no threshold response is the premise used for radiation protection purposes within the clinical environment. However, controversy regarding the actual shape of the dose-response curve in this low-dose region remains. It is most likely that the linear no threshold model is a conservatively safe basis for radiation protection.

It is difficult to show a causal connection with radiation and cancer directly in either patient groups undergoing medical imaging or staff who have been occupationally exposed. Roguin et al. have facilitated a quasi registry of brain cancers in interventionalists (n=31) primarily through self-reporting. Their findings show a greater percentage (22 of 26 cases, 85%) of left-sided malignancy , which is speculated to correlate with the higher levels of radiation typically received on the left side of the head because of the positioning of the interventionalist during a case. Obviously whether these cancers are radiation induced and whether they demonstrate an excess over natural incidence cannot be established with only anecdotal reporting.

A large-scale epidemiology study of US radiologic technologists found no evidence of an excess risk of malignant intracranial tumor due to occupational radiation exposure, including in a subcohort with consistently higher brain doses who assisted with fluoroscopy-guided interventions. However, they did find an excess risk for unknown reasons among those technologists who had ever (compared with those who had never) worked in fluoroscopy-guided procedures. They observe that these results do not necessarily mean that there is no association between radiation and brain tumors, but instead that the small number of cases and the low brain doses (mean cumulative weighted brain dose of 12 milligray [mGy]; range, 0–290 mGy) may limit identification of any causal connection.

Similarly, recent large-scale studies of radiologists and physicians likely to perform image-guided interventions using a control group of psychiatrists found no association between occupational radiation exposure and cancer. Dosimetry was not available for these studies, so dose response was not assessed. However, excess mortality risks were seen in radiologists and physicians who graduated before 1940, seemingly correlating with higher dose exposure in the earlier years.

The other type of stochastic response to radiation exposure is heritable effects, where exposure of the reproductive cells may lead to disease in the children of the person exposed. These types of genetic effects are considered to follow the same linear no threshold dose response for the purpose of radiation protection. Interestingly, there is no direct evidence of heritable effects in human studies, although experiments with animals have shown otherwise. Overall, the International Commission on Radiological Protection (ICRP) recommends a risk coefficient of 5% per sievert of radiation for fatal stochastic effects. At the dose levels relevant to medical imaging this equates to 0.005% per millisievert (mSv) or a risk of fatal radiation-induced cancer of 1 in 20,000 per mSv. It is essential to recognize that this is a population risk rather than specific to an individual.

At a cellular level, radiation exposure can cause direct damage to DNA resulting in double- or single-stranded breaks (lesions) or indirect damage through the production of reactive oxygen species that also lead to DNA damage. Repair begins immediately, with opportunity for full repair or misrepair resulting in chromosomal aberrations. It has been shown that the cellular damage mechanism may be inhibited or modulated and/or repair mechanisms up-regulated by a variety of radioprotective agents . It is thought that reducing damage will ultimately reduce the risk of radiation-induced carcinogenesis. However, this is difficult to prove because it would require long-term follow-up through epidemiologic studies. To date, most research in this area has been achieved through in vitro experimentation, and although some agents are used to reduce radiotherapy side effects, none have been administered outside of research in diagnostic or interventional imaging. Velauthapillai et al. have shown that oral antioxidants taken 1 hour before technetium-99m methylene diphosphonate bone scans can block 90% of DNA breaks seen in the control group. This could lead to patients and healthcare workers being medicated before radiation exposure. This was an entirely in vivo study.

Assessments of cellular damage are usually achieved through quantifying radiation-induced double-stranded breaks through a biomarker, γ-H2AX. Radiation insult leads to activation of a response through protein kinase, predominantly by ataxia-telangiectasia-mutated (ATM), which causes phosphorylation of the histone H2AX producing γ-H2AX. The levels of γ-H2AX (as a damage/repair marker) and phosphorylated ATM (as a response marker) are time dependent after radiation exposure and have become more readily quantifiable through cytometry using fluorescence-activated cell sorting. Despite the difficulty in performing in vivo research in this area, a recent study using blood samples taken from interventionalists before and after performing image-guided endovascular aneurysm repair (EVAR) cases showed that levels of both γ-H2AX and pATM increased significantly. Interestingly, levels of both markers normalized after 24 hours.

In the study of interventionalists performing EVAR procedures, blood samples were also exposed to known doses of radiation in vitro. The results demonstrated a variable individual response to radiation, as measured by γ-H2AX prevalence. This study and others have suggested that individual susceptibility to radiation exists. It is recognized that people with genetic disorders involving an impaired genetic response to DNA repair—for example, ataxia telangiectasia—have higher susceptibility to radiation effects or are considered more radiosensitive. It is worth considering.

One feasible risk-reduction method would be the development of a prescreening procedure for current or future interventionalists for inherited mutations that would predispose individuals to radiation-induced cancer. The first step would be to look at family cancer history and to identify high-risk populations. Some of this is simple population genetics. The Ashkenazi Jewish population, for instance, has a 2% chance of carrying a BRCA germ-line mutation that predisposes to impaired p53-mediated DNA, the consequence of which manifests as increased risk of breast, ovarian, peritoneal, and prostate cancers. Interestingly, concern has been raised regarding the potentially carcinogenic effects of low-dose ionizing radiation from annual or even biannual mammograms in BRCA1 and/or BRCA2 mutation carriers. If we are concerned about biannual low-dose mammographic radiation breast cancer induction, should we not be concerned about the higher doses potentially received by interventionalists who carry BRCA1/2 mutations?

A follow-up step might be to offer potentially at-risk interventionalists and trainees access to genetic screening for germline mutations (likely including mutation types in BRCA1, BRCA2, and MLH1, among others) that increase susceptibility to radiation-induced DNA damage. Certainly, screening for specific mutations is a well-established and increasingly affordable practice in the context of patients with positive family histories of breast, ovarian, and colorectal cancers. The intent of such a screening process would be to provide an objective estimate of the risk an individual might incur by pursuing a career in interventional radiology, and to allow them to make informed decisions on the basis of that knowledge. The choice of what to do in the event that impaired DNA repair ability were identified would be completely up to the individual; perhaps an early trainee would choose to pursue a magnetic resonance imaging or ultrasound intervention-based career, whereas an established interventionalist might heighten their adherence to lead barrier-based risk-reduction methods or reduce the number of high-exposure procedures they perform each year. Of course, the potential ethical ramifications of such screening would have to be considered carefully. For one, the test results would need to be strictly confidential to safeguard against any possible discrimination, however well meaning, that might ensue from their dissemination. Screening would have to be performed strictly in the context of appropriate counselling and informed interpretation. Finally, participation in screening would have to be purely optional, with the choice to know more about their genetic risk profile left up to each individual.

Tissue Reactions

Radiation exposure can also induce a reaction in the tissues or organs exposed. Tissue reactions are similarly known as deterministic effects . These are typically evident on a much shorter time frame than stochastic effects and occur above a threshold and hence are not probabilistic in nature. As doses increase above the threshold the severity of the effect also increases. The most common tissue reactions for patients seen in image-guided interventional procedures are those occurring in the skin. Above a threshold, which is thought to be around 2 Gy, the initial effect will be transient erythema, but as doses increase this can progress to permanent epilation, radiation-induced telangiectasia, dermal atrophy or induration, and, at particularly high doses, necrosis with possible surgical intervention necessary.

Radiation-induced skin injury is well known in radiotherapy. However, it is relatively rare in image-guided procedures. In fact, it is likely that an interventionalist may not observe a skin effect in a patient during their working lifetime when using good practice in an environment with a radiation protection program. There are many reported cases of skin injury despite being infrequent. When an injury occurs, it can be severe and significantly affect quality of life. Some injuries result from patient factors (obesity, previous radiation exposure); others may be considered avoidable (machine malfunction, unintentional direct irradiation of body parts (arms), and lengthy procedures with no real-time dose feedback. In addition, the medications that the patient is taking should be considered. A patient undergoing chemotherapy will have diminished cellular repair ability and may be at higher risk of radiation dermatitis.

Again, the issue of individual sensitivity to radiation exposure is relevant. Knowledge from both radiotherapy and image-guided interventional procedures that produce skin injury shows that predisposing factors for radiation-induced skin damage include smoking, diet, and impaired skin integrity as well as an increase in severity when combined with certain chemotherapy agents even at a period of months postprocedure in so-called radiation recall. It has been suggested that follow-up by a dermatologist would identify more skin injury than follow-up by a radiologist.

Although not well recognized, radiation-induced circulatory effects (cardiovascular and cerebrovascular, such as heart attack and stroke) are noncancer outcomes that have also been observed in the LSS and radiotherapy patients. These are believed to be due to elevated circulating reactive oxygen species resulting in a chronic low-grade inflammatory state. The ICRP first included circulatory disease as a radiation outcome in 2012, with a threshold dose of 0.5 Gy for acute, fractionated, and chronic exposures. It is acknowledged that much uncertainty remains, hence the application of a single threshold regardless of dose rate. These effects, if occurring, are relevant to patient exposure of the heart and brain in complex interventional procedures where doses may exceed these values. The threshold dose is defined as the level at which 1% incidence of the detrimental effect in the exposed population is expected. For circulatory disease, this is a small increase over the natural incidence of 30% to 50% and the radiation-induced effect is expected to occur after a latency of at least 10 years.

For interventional staff, it is becoming increasingly understood that opacification of the lens of the eye, or cataracts, is a potential risk from radiation due to occupational exposure. As vision impairment worsens, cataract surgery can be required. In revising the scientific literature, the ICRP in recent years has lowered the theoretical threshold for radiation-induced cataract induction to 0.5 Gy for acute and protracted exposures. This is indicative of the higher radiosensitivity of the lens compared with other tissues. In fact, it is not clear whether cataract formation is potentially a stochastic effect because there is considerable uncertainty regarding the dose-response mechanism in the lens of the eye and if a threshold is indeed applicable.

Several factors appear to increase the risk of radiation-induced cataract, including increasing age, gender (female), marital status (single), socioeconomic status (low), smoking history, and body mass index, among others. However, these also tend to be predisposing factors for natural incidence, making it difficult to isolate the radiation causal relationship.