Introduction

Medical use of ionizing radiation has grown rapidly and is now a significant source of radiation exposure for the USA and the European Union populations [1].There has been continuous emphasis on radiation dose optimization and development of newer CT technologies [2].

The IAEA initiated a Smart Card project in 2009 with the purpose of providing a methodology for tracking radiation exposure of patients [3, 4]. Currently, tracking has become a routine in many countries [4] and systems are available which could also provide radiation exposure history of patients. This requires using the same patient ID, converting the reference dose quantity of the imaging modality into effective dose, and summing to get cumulative effective dose (CED) [5].

A review of the literature has shown that patients undergoing diagnostic and interventional radiological procedures can receive cumulative effective doses (CED) in the range of 50–500 mSv in just a few years [6,7,8,9] but it was not known if this was a widespread phenomenon or restricted to isolated clinical situations and a small number of institutions. Following a large study through Massachusetts General Hospital (MGH study) in Boston, wherein several thousands of patients were found to have CED ≥ 100 mSv [10, 11], a joint effort was launched by IAEA and MGH to assess the situation in other parts of the world. In preparation for the meeting, nominated participants were asked to collect data from their hospitals and countries.

The focus of the meeting was to (a) look at data from different countries collected specifically for this meeting through the IAEA-MGH survey on patients with CED ≥ 100 mSv, (b) discuss available literature on patients with CED ≥ 100 mSv and radiation effects at this level of radiation dose, (c) create awareness about the findings on the number of patients with CED ≥ 100 mSv, (d) discuss limitations, if any, of the current framework on radiation protection in dealing with new findings, and (e) develop plans for future work.

This paper reports the summary of the findings and conclusions of the Technical Meeting on Radiation Exposure of Patients from Recurrent Radiological Imaging Procedures, organized by the IAEA on March 4 to 6, 2019.

Methods

The governments of all IAEA Member States were invited to nominate their representatives for the meeting. More than 50 experts from 26 different countries and 9 international organizations and professional bodies participated. A list of participating countries that provided presentations at the meeting is shown in Table 1. A list of participating international organizations and professional bodies is shown in Table 2. One of the sessions focused on “Radiation risks at doses 100–300 mSv” and summarized on current knowledge on potential and documented radiation risks in the does range 100–300 mSv.

Table 1 Participating countries grouped as less-developed or developed. Only countries which had a presentation in the technical meeting are included
Full size table
Table 2 Participating international organizations and professional societies
Full size table

The data was collected from several countries as per further details provided in “Basis and results of the IAEA-MGH survey on high CED patients.”

Results and discussion

Radiation risks at doses 100–300 mSv

The reviews undertaken by the United Nations Scientific Committee on Effects of Atomic Radiation (UNSCEAR), International Commission on Radiological Protection (ICRP), and National Council of Radiation Protection and Measurements (NCRP) have all been successful in evaluating radiation-related adverse effects (particularly cancer) at or above absorbed dose of around 100 mGy [12]. The recent review undertaken by NCRP of all studies available in the world at absorbed doses below 100 mGy concluded that the linear-no-threshold (LNT) model should be continually used for radiation protection purposes [13].

Altogether, the data available provide strong evidence of an increased cancer mortality risk at equivalent doses greater than 100 mSv, good evidence of an increased risk at doses between 50 and 100 mSv, and reasonable evidence for an increased risk at doses between 10 and 50 mSv. At the level of effective dose of 100 mSv, many organs have the potential to obtain doses of 100 mGy or more.

Individual patient radiation exposure tracking and cumulative dose estimation

The benefits of tracking have been documented in strengthening the process of justification and optimization of radiation protection [14, 15]. The use of tracking to assess cumulative radiation dose of individual patients has increased in recent years [16, 17], as this feature is being implemented in most commercial dose management systems.

Notwithstanding the well-known limitations of effective dose, there are numerous advantages in the use of effective dose to estimate the cumulative radiation dose: it summarizes the detriment using a single number, it can be used to enable comparison of relative detriment between procedures that utilize ionizing radiation, there is a large database and experience, and dose management systems provide it routinely. Finally, despite large uncertainties associated with the estimation of effective dose, at the level of 100 mSv of effective dose and higher, many organs receive high doses at which statistical increase in radiation risk has been demonstrated.

One approach to address the issue of identifying high CED patients is to study groups of patients who can be considered at risk for receiving high cumulative doses. These groups will likely include patients with chronic diseases who require ongoing care that results in recurrent imaging and exposure to radiation for both diagnosis and interventions. A second approach is to consider the cumulative radiation dose which can be accrued in a single episode of care, known to involve an extensive use of imaging with ionizing radiation. A third approach, which involves the use of the dose management systems, is to retrospectively identify patients who had accrued more than a pre-set alert value of effective dose. All three approaches are discussed below.

Assessing magnitude of cumulative doses

In a study in 2009, Fazel et al [18] reported a population-based survey of radiation doses from medical imaging procedures received by about 1 million people over a 3-year period. They obtained data from five healthcare markets across the USA on imaging procedures associated with ionizing radiation exposure. They used typical dose values from the literature, multiplying with the number of exams. The study reported a mean CED of 2.4 mSv with a standard deviation of 6.0 mSv per enrolee per year. CT and nuclear imaging accounted for 75.4% of the CED. The population-based rates of “moderate” (> 3 to 20 mSv per year), “high” (> 20 to 50 mSv per year), and “very high” (> 50 mSv per year) doses were estimated as 19.4%, 1.9%, and 0.2%, respectively.

Subsequently, several published studies have attempted to quantify the proportion of patients exposed to relatively higher level of cumulative doses. They are summarized in Table 3 and described below.

Table 3 Cumulative radiation exposure and patients with CED > 100 mSv
Full size table

Cardiac patients

Chen et al [19] assessed the CED from diagnostic and therapeutic cardiac imaging procedures in a general population of non-elderly adults. Among 90,121 patients who underwent one or more procedures, there were 3173 of 90,121 (3.5%) individuals who received > 60 mSv and 75 of 90,121 (0.08%) individuals who received > 150 mSv. Einstein et al [20] investigated the radiation exposure of 1097 patients who underwent myocardial perfusion imaging (MPI) study and their 20-year follow-up radiation exposures. They found that 344 of the 1097 (34%) patients received CED > 100 mSv, including 120 of the 1097 (11%) who received CED > 200 mSv. Stein et al [21] assessed 11,072 patients diagnosed with cardiac disease during a 5-year period and evaluated CED over a 3- to 8-year period, following initial diagnosis. After 3 years, 533 of 8656 (6.2%) of patients had a CED > 50 mSv. This percentage increased to 14.2% and 20% after 6 and 8 years, respectively. Kaul et al [22] did a retrospective assessment of CED in 64,071 patients admitted to hospitals in the USA, from 2006 to 2009, with acute myocardial infarction (AMI). In this case 1060 of 64,071 (1.7%) of patients received a CED > 50 mSv during their admission. Eisenberg et al [23] created a retrospective cohort of 82,861 patients admitted to a hospital in Quebec with an AMI between 1996 and 2006, and reported 15,090 of 82,861 (18%) patients who had a CED ≥ 30 mSv in the first year after admission. Lawler et al [24] in a population-based survey assessed the CED that 106,803 patients hospitalized with AMI from 1996 to 2004 accrued over a 3-year period after admission, finding that in a given year (2004), 825 of 11,427 (7%) patients had a CED ≥ 30 mSv in the first year after admission. Noor et al [40] in a single-center study assessed radiation exposure during a 10-year follow-up of 202 heart-transplanted patients and reported a mean CED of 84 mSv. Johnson et al [41] assessed cumulative radiation exposure in 10 children who underwent heart transplant, reporting a median post-transplantation CED of 45.8 mSv. Similar findings were reported by Seal et al [42] who reported a median CED of 81 mSv in 15 patients.

Some common aspects can be summarized despite the heterogeneity of the population under study: although most cardiac patients received low or moderate radiation exposure from medical procedures, there exist sizeable groups of patients aged 35–54 years, who received high CED in a short time period, with potential for long-term radiation effects. Patients admitted to hospital for AMI and patients undergoing heart transplant are two such groups.

End-stage kidney disease patients

In a study by Kinsella et al [25], 100 hemodialysis patients were followed for a median period of 3.4 years and it was reported that 13 of 100 (13%) patients had a CED > 75 mSv. In a study by De Mauri et al [26] of 106 hemodialysis patients during a 3-year follow-up, 17 of 106 (16%) patients had CED > 100 mSv. Coyle et al [27] performed a retrospective study including 244 hemodialysis and 150 kidney-transplanted patients with a median follow-up of 4 years, reporting 56 of 244 (23%) patients with CED > 100 mSv, for hemodialysis patients, and 12/150 (8%) kidney-transplanted patients with CED > 100 mSv.

The results of another study by De Mauri et al [28] of 92 kidney-transplanted patients with a 4.1-year median follow-up showed that 26 of 92 (28%) patients had a total CED > 50 mSv, and 11 of 92 (12%) patients had a total CED > 100 mSv. Seal et al [30] reported a median CED of 48 mSv in 15 pediatric kidney-transplanted patients.

In summary, the cumulative exposure is high in hemodialysis patients: exposures > 50–100 mSv of CED accrued in 3–4 years are common, occurring in almost one-third of subjects. The cumulative exposure in kidney-transplanted is lower than in hemodialysis patients. On the other hand, kidney-transplanted patients are younger than hemodialysis patients with more time left to develop radiation-induced malignancies. In addition, many transplant patients will have been on hemodialysis for year prior to transplant and the hemodialysis patients on the waiting list for kidney transplant have been demonstrated to be the most exposed subgroup of hemodialysis patients [43]

Crohn’s patients

Desmond et al [29] retrospectively calculated CED from an imaging performed in 354 patients with Crohn’s disease followed over a 15–year period. CED exceeded 75 mSv in 55 of 354 (16%) patients. With only one-third of follow-up duration, Levi et al [30] showed that 29 of 324 (8%) inflammatory bowel disease patients exceeded 75 mSv of CED. Similar results were also reported by other authors [31, 32, 44].

In summary, cumulative exposure in patients with Crohn’s disease is of concern. Exposures exceeding 50 mSv of total CED are not uncommon in this study cohort, occurring in almost 10–30% of subjects who underwent imaging [33,32,35, 45, 46]. Moreover, it is likely that many subjects diagnosed with Crohn’s disease in pediatric age [36,35,38] will eventually accrue more than 100 mSv at the age of 30 years and more than 200 mSv at the age of 50 years.

Endovascular aortic repair

Brambilla et al [39] reported in 71 patients with a mean follow-up of 1.84 years that 5 patients had a total CED < 100 mSv; 24 were in the 100 to < 200 mSv group, 24 were in the 200 to < 300 mSv group, and the remaining 18 had a total CED > 300 mSv. Kalender et al [47] reported in 59 patients with a median follow-up of 2.3 years a mean CED for the first-year post-EVAR of 109 mSv, and the ED delivered during each subsequent year was 16 mSv/year.

Estimated CEDs for endovascular aortic repair (EVAR) patients are very high with a predicted > 100 mSv in the first year following EVAR and 8–16 mSv yearly thereafter. Due to the elevated mean age of EVAR patients, the associated risk might not be considered of clinical relevance [48]. Also, the risk associated with not doing the EVAR is significant and alternative approaches carry different but very significant risks. However, despite all of these considerations, the pre- and post-procedure imaging should be properly justified and optimized to keep the overall dose as low as reasonably achievable [49].

Basis and results of the IAEA-MGH survey on high-CED patients

Publications in the scientific literature as listed above have shown that some patients undergoing recurrent radiological procedures can receive CED in the range of 50–1000 mSv or in some cases perhaps more. However, it was not known if this is a phenomenon restricted only to these disease categories or it is widespread among patients overall. This necessitated study on all patients undergoing CT exams at several hospitals in different countries. Following assessments carried out at the MGH in Boston, the study was extended to other hospitals in the USA and at national level in a European country. In the data from 324 hospitals, 254 of which were from the USA, it was found that the patients with CED ≥ 100 mSv vary from 0.64 to 3.4% with an average of 1.33% among more than 2.5 million patients undergoing CT scans in group of hospitals included in the study [10]). It must be noted that these assessments include only the CT exams underwent by patients and exclude other equally high-dose exams like PET/CT or interventional procedures these patients may have undergone. Further, unlike the study by Fazel et al [17] in which dose estimations were based on multiplying the number of exams with typical effective dose per exams with the possibility of inaccuracies, the MGH study used dose information provided by established dose management systems based on the dose-length-product (DLP) values for individual patients.

This created the need to widen this assessment and thus, a joint survey was launched by IAEA and MGH to assess the situation in other parts of the world. In the preparation for the meeting, nominated participants were asked to bring data from their hospitals and countries. The results of this IAEA-MGH survey are summarized in Table 4.

Table 4 Number of patients with CED ≥ 100 mSv in the IAEA-MGH survey
Full size table

The data available was from 20 hospitals: 18 of them in Europe, one in Africa, and one in Asia. The patients with CED ≥ 100 mSv vary from 0% for the hospitals from Africa and Asia to 5% for a hospital in a European country, with an average of 0.65% of 702,205 patients undergoing CT scans in these 20 hospitals.

Extrapolations

The data in this study from 20 countries covering 0.7 million patients indicated that 0.65% patients received CED ≥ 100 mSv. This is in line with the factor of 0.64% used by MGH study [10]. Thus, the estimate of 0.9 million patients with CED ≥ 100 mSv as reported by the MGH study appears reasonable, despite the fact that the data in Table 4 is not yet published anywhere else.

Methods to estimate the patient dose in automatic dose management systems

Numerous models to estimate dose from machine parameters are in use; the multiplicity of approaches means that estimates of effective or organ dose can be affected by discrepancies between systems [50]. However, at the level of CED > 100 mSv, the differences become small.

Clinical justification of exams leading to high cumulative doses

The patients who received such high cumulative radiation doses typically need radiological imaging exams for multiple clinical indications, for follow-up of malignant disease or some chronic conditions. There is a need for assessment of justification of recurrent CT examinations as an ongoing practice [11]. The selection of the appropriate investigation and the justification of radiological examinations should be facilitated by referral guidelines for medical imaging, considering the radiation doses [51]. Consideration of the cumulative radiation dose may provide impetus to the development of new or modification of existing strategies for imaging evaluation that could include substitution by other modalities, such as ultrasound or MR imaging. A European survey on imaging referral guidelines lead among representatives of national radiological societies, nuclear medicine societies, and competent authorities in 30 countries showed that only in 70% of the countries the representatives were aware of the legal requirement for guidelines and in only 60% of the countries the representatives were aware of the presence of referral guidelines [52].

Existing guidelines do not include specific criteria for referral of patients who require multiple and/or long-term imaging studies [11, 53].

Recommendations

The meeting discussed results available to date and deliberated on interpretation and recommended actions, which are summarized below:

  1. 1.

    Although the data available covers many hospitals and countries than previously reported, there is a need to do further work to fully understand the extent of recurrent exposures and validate doses involved and whether any of these exposures could be avoided or further optimized.

  2. 2.

    Urgent actions are required to heighten awareness among referring physicians of the current situation. The tools available include referral guidelines with information on radiation dose, clinical decision support systems, and information about previous imaging exams.

  3. 3.

    There is an urgent need to determine imaging strategies for patients with long-term illnesses.

  4. 4.

    There should be models for predicting patients with different clinical conditions who are likely to reach high cumulative dose range. Professional medical societies should develop or adopt appropriateness criteria/referral guidelines for patients who require multiple and/or long-term imaging studies. This requires a multi-disciplinary consensus to be achieved between the radiological medical practitioners and referring physicians of different specialities.

  5. 5.

    When a series of procedures can be reasonably foreseen, the risks and benefits of the entire series should be considered in the justification process. In these cases, the consideration of alternative non-ionizing techniques is especially important.

  6. 6.

    Special consideration is needed for optimization of individual procedures for patients with multiple imaging by utilizing the concept of diagnostic reference levels tied to defined clinical imaging task and other approaches for optimization.

  7. 7.

    Further strengthening and harmonizing education and training of radiological medical practitioners, radiographers, and medical physicists is emphasized.

  8. 8.

    Industry can play an important role by producing CT scanners capable of achieving adequate image quality at sub-mSv radiation dose.

  9. 9.

    There is an urgent need for inclusion of the concept of patient cumulative radiation exposure in radiation protection framework and standards.

  10. 10.

    There is a need to develop metrics that retains the useful aspects of the effective dose but also addresses its limitations for use in patients. Besides other aspects, it should be related to body habitus, age, and gender.

  11. 11.

    There should be standardization and validation of methods to estimate the patient dose in automatic dose management systems. Conversion factors to estimate organ dose should be standardized.

  12. 12.

    Alert values for cumulative radiation exposures of patients should be set up and introduced in dose management systems with suitable cautions provided to avoid misuse. Color scales could be adopted for alert values.

  13. 13.

    There is a need for health authorities to implement policies for applying technological solutions for patient exposure monitoring and integrating them into the electronic health records.