Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Department of Surgery, Section of Hepatobiliary Surgery and Liver Transplantation, New York Presbyterian Hospital – Weill-Cornell Medical Center, New York, NY
Address reprint requests to: Michael Kluger, M.D., M.P.H., Assistant Professor of Surgery, Division of Hepatobiliary Surgery and Liver Transplantation, Weill Cornell Medical College, New York-Presbyterian Hospital, 525 East 68th St., Baker 1913, New York, NY 10065. E-mail: firstname.lastname@example.org; Tel.: 212-305-6514; fax: 212-305-5992.
Potential conflict of interest: Nothing to report.
Substantial evidence has linked ionizing radiation exposure (RE) to oncogenesis. Patients evaluated for transplantation undergo extensive diagnostic imaging and have increased baseline cancer risk factors. The objective was to examine exposure in a cohort of patients undergoing evaluation and liver transplantation. Radiation exposure from all diagnostic examinations and procedures were retrospectively recorded. Radiation exposure is reported in mSv, a standardized measure of the detrimental biologic effect of radiation which allows for population-level comparisons. Seventy-four patients (69% male, mean 57 years) were evaluated, of which 13 of 35 subsequently listed patients were transplanted; an additional 18 previously evaluated patients were also transplanted during 2010. The most common indications were hepatitis C (55%) and hepatocellular carcinoma (HCC) (30%). The median observation period was 14 months. In all, 1,826 imaging examinations were performed, of which 408 (22%) involved considerable ionizing radiation and were the focus of investigation. Median annualized effective RE was 51 mSv (interquartile range [IQR]: 19,126), with 10% exposed to almost twice the amount of radiation recommended for a 5-year period. Patients with HCC received significantly (P < 0.00001) higher median annualized effective RE than patients without HCC, 137 mSv (IQR: 87,259) versus 32 mSv (IQR: 13,57), respectively. Computed tomography (CT) abdomen (23%) and chest (16%) accounted for the most common exposures, with CT abdomen accounting for 46% of overall cohort RE. Conclusion: Patients undergoing evaluation and liver transplantation at our center are exposed to very high levels of ionizing radiation. Although long-term effects in these patients are yet to be defined, the theoretical increased risk of malignancy must be given its due consideration. Routine use of nonradiation imaging and reconsideration of indications may be preferred and justified in this population. (Hepatology 2014;59:496–504)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
American Association for the Study of Liver Diseases
computed tomography angiography
dual-energy X-ray absorptiometry
European Association for the Study of the Liver
endoscopic retrograde cholangiopancreatography
magnetic resonance imaging
New York Presbyterian Hospital
The contribution of radiation-based imaging technologies such as computed tomography (CT) and fluoroscopy to the diagnosis and management of liver disease is undeniable. Over the past decade, concerns have been raised by physicians and patients about the increased cancer risks of medical radiation exposure (RE).[2, 3] Exposure to ionizing radiation has been linked to the development of both solid cancers and leukemias.[2-4] The small but real risks must be balanced against the benefits of improved diagnostic accuracy and guidance in management decisions.[4-6] Patients being evaluated for liver transplantation undergo an extensive battery of radiographic examinations and often undergo imaging for treatment of complications and long-term surveillance posttransplantation.[7, 8]
The potential oncologic impact of radiation and immunosuppression on this patient population cannot be ignored. However, there are no data in the current literature documenting the amount of RE these patients receive.[4, 9-13] The objective of this study was to quantify RE at a single transplant center during liver transplantation evaluation, wait listing, and posttransplantation.
Materials and Methods
Study approval was obtained from the Weill Cornell Medical College Institutional Review Board. This was a retrospective review of all patients who underwent evaluation for and/or liver transplantation between 1 January and 31 December 2010, with follow-up through 31 December 2011. Electronic medical records were reviewed to identify all imaging performed at our center or independent facilities. The evaluation period was defined as 60 days before signing informed consent for transplantation evaluation through either listing or denial. The waitlist period was the time the patient was listed to transplantation, or to the end of follow-up. The posttransplantation period was defined as the time from transplantation to postoperative day 60.
Diagnostic examinations and interventions are directed by written program policies based on American Association for the Study of Liver Diseases (AASLD) practice guidelines for liver transplantation evaluation and hepatocellular carcinoma (HCC).[7, 8] Comprehensive age/history/symptom-based evaluations assess malignancies and comorbidities that affect candidacy.[7, 14-18] Cardiovascular evaluation is based on AASLD guidelines and American College of Cardiology / American Heart Association practice guidelines for major vascular/high-risk procedures. Bridging strategies with the aim of HCC control are practiced as the mean time from diagnosis to transplantation is 12 ± 9 months at our center. Endoscopic retrograde cholangiopancreatography (ERCP) is performed based on symptoms or magnetic resonance imaging (MRI)-based malignancy concerns. Although individual program guidelines are not detailed in the literature, our practices appear similar to major international centers and standards.[7, 20-25]
All modalities emitting radiation were included, as well as MRI and ultrasonography (US). Radiation exposure is reported in millisieverts (mSv). This is a measure designed to represent the overall detrimental biologic effect of RE and the potential for radiation-related mutagenic changes in each organ in a reference subject.[4, 6, 26] It allows for meaningful population-level comparisons across different types of RE.[2, 6, 10]
Radiation exposure from plain radiography was estimated based on published effective doses. For CT and nuclear imaging studies, doses were abstracted from the screen capture information or from diagnostic reports. Ionizing RE from CT is typically recorded as dose-length product, a measure related to patient exposure; for the purposes of the current investigation, this was converted to mSv using factors derived by Huda et al. For interventional radiology, gastroenterology, and cardiology procedures involving fluoroscopy, departmental logs were reviewed. When RE was not available for an examination, the median for all patients undergoing the same procedure was substituted to reflect RE based on institutional practices. The effective RE for 94% of examinations, excluding cardiac catheterization and ERCP, were obtained. The effective RE for cardiac catheterization and ERCP could not be reliably obtained from hospital records, so standardized values were used; these procedures accounted for only 4% (17/408) of all exposures.
The frequency of each modality and effective RE was calculated for the study population, and categorized by evaluation, waitlist, and posttransplantation period and the presence of HCC. Importantly, there exist no limits or categories for patient RE. Radiation exposure was differentiated among the following categories of occupational RE for at-risk workers: low (≤3 mSv/year), moderate (>3-20 mSv/year, the upper annual limit averaged over 5 years), high (>20-50 mSv/year, the upper annual limit for in any given year), and very high (>50 mSv/year).[2, 4, 29]
Continuous variables were compared with a Student t-test, and categorical variables with the Pearson's x2 or Fisher's exact tests, where appropriate. All data were analyzed using Stata 11.2 (StataCorp, College Station, TX).
Seventy-four patients were included in the study: 56 patients who consented for liver transplantation evaluation in 2010 and 18 patients evaluated prior to 2010 who were transplanted during 2010 (Table 1). Twenty-two patients (30%) had HCC. The median overall observation period, was 14 months (interquartile range [IQR]: 5,21).
Table 1. Demographic and Clinical Characteristics at Time of Consent for Liver Transplantation Evaluation (n = 74)
Patients n (%)
Age, median (range) years
Patients listed in 2010
Patients denied listing in 2010
Listed in 2010, transplanted 2010-12/2011
Listed prior to 2010, transplanted 2010
Primary Liver Disease
Hepatitis B virus
Hepatitis C virus
Primary sclerosing cholangitis
Primary biliary cirrhosis
Cryptogenic, NASFLD, other
Intervals, median months (IQR)
Overall study period
Consent to listing
Consent to denial
Listing to transplantation
A total of 1,826 diagnostic imaging examinations and procedures were performed during the study period: 27% were MRI or US examinations, 51% were simple chest or abdominal radiographs, or dual-energy X-ray absorptiometry (DEXA) scans. Patients underwent a median of six (IQR: 2,10) chest radiographs with a median overall effective RE of 0.11 mSv (IQR: 0.04, 0.2), and a median of 0 (IQR: 0, 4) abdominal radiographs with a median overall effective RE of 0 mSv (IQR: 0, 2.8). Fifty-seven DEXA scans were performed in 57 individuals, with an effective RE of 0.001 mSv each.
The utilization and proportion of different imaging modalities in the study, and the proportion of total RE each modality contributed, are summarized in Figs. 1 and 2, respectively. CT of the abdomen/pelvis (23%) and chest (16%) accounted for the two most commonly performed diagnostics with radiation. With regard to effective RE, CT abdomen/pelvis for diagnostic and ablative purposes accounted for 46% of RE, followed by embolization procedures (22%). This demonstrates that frequency of an examination did not correlate with overall RE because of differences in emission among modalities. Twenty-two percent (n = 408) of studies and procedures involved substantial radiation and were the focus of this investigation. Table 2 demonstrates individual patient utilization trends and effective RE of studies performed; these did not notably differ from standardized values. Radiation exposure ranged from 0.4 mSv for mammography to 48.48 mSv for a CT-guided ablation of an HCC.
Table 2. Utilization of Diagnostic Examinations and Procedures, and Effective Radiation Exposure (mSv) During the Overall Study Period for 74 Patients
There were clear trends in utilization over the study period (Fig. 3). For studies with considerable ionization, 215 studies were performed during the evaluation period, versus 130 and 63 studies for the waitlist and the posttransplantation periods, respectively. The most studies occurred during the evaluation period, with a greater utilization of nuclear studies and fluoroscopy or angiography. As candidacy is being assessed, this is a logical finding; based on these findings, this period should be targeted to reduce exposure. There was a proportional increase in the utilization of CT in the postoperative period as a consequence of its accessibility and speed when imaging and intervening upon in-patients with complications. Given the benefits of CT under these circumstances, this is a difficult period to reduce exposures (Fig. 3). For the nonionizing studies, there was a relatively constant number of studies performed throughout the evaluation, waitlist, and posttransplantation period. US became a dominant examination in the postoperative period, surpassing ionizing studies secondary to its portability. MRI abdomen was the dominant abdominal imaging modality during the evaluation and waiting periods.
Patients had a median RE of 53 mSv (range: 0-569) during the study period; patients in the 90th percentile were exposed to greater than 189 mSv (Table 3). The annualized median RE for the overall study period was 51 mSv (range: 0-427). Based on annual occupational limits for at-risk workers, 51% of patients had very high RE (>50 mSv), with 10% being exposed to almost twice the 5-year occupational limit for at-risk workers.[2, 4, 29]
Table 3. Rates of Radiation Exposure and Categorization According to Occupational Limits for At-Risk Workers by Study Period, and Diagnosis of Hepatocellular Carcinoma
Radiation Exposure Categories (Occupational Limits for At-Risk Workers)
During the evaluation period, patients had a median effective RE of 20 mSv (range: 0-332); patients in the 90th percentile were exposed to greater than 112 mSv. The annualized median RE during the evaluation period was 36 mSv (range: 0-621); the 90th percentile received over 217 mSv. Almost half the patients (46%) had very high RE based on annual occupational limits for at-risk workers during the evaluation period, signifying the evaluation period as the interval with the greatest RE; 10% were exposed to over twice the 5-year occupational limits for at-risk workers. There was no significant difference in effective RE whether a patient was denied or listed (P > 0.31).
For the waitlist period, the median effective RE was 6 mSv per patient (range: 0-353); the 90th percentile were exposed to over 107 mSv. The median annualized RE during this period was 11 mSv (range: 0-424); patients in the 90th percentile were exposed to greater than 127 mSv. During the waitlist period, conversely, almost half of the cohort (47%) had low exposure based on occupational limits for at-risk workers. This decreased to 36% when considering the annualized rate, illustrating that cumulative exposure increases with time on the waitlist.
The median effective RE during the 60-day posttransplantation period was 16 mSv (range: 0-202); patients in the 90% percentile were exposed to over 75 mSv; 23% had very high RE based on occupational limits for at-risk workers, a designation used to categorize exposure over a 12-month period. Over the 2-month period, 10% of patients were exposed to three-quarters of the 5-year occupational limits for at-risk workers.
Patients with HCC had significantly higher cumulative RE than patients without (P < 0.00001). For patients with HCC, the median annualized effective RE was 137 mSv (range: 32-427) as compared to 32 mSv (range: 0-331) in patients without HCC. HCC patients in the 90th percentile were exposed to greater than 314 mSv. Based on annualized rates, 96% of patients with HCC had very high RE, with 10% being exposed to over three times the 5-year occupational limits for at-risk workers.
Medical imaging accounts for the largest source of ionizing RE in the United States. Ionizing radiation can cause deoxyribonucleic acid mutations that may eventually result in cell death or the development of cancer in a dose-dependent manner.[6, 30-32] Epidemiological and experimental evidence has shown that exposure to medical range radiation is associated with radiation-induced cancer.[2-6] It is estimated that at current rates of RE from medical imaging, up to 2% of future cancers will occur as a consequence of this cumulative exposure.
Liver transplantation recipients are already at increased risk of cancer because of a higher prevalence of established risk factors: advanced age, environmental exposures to alcohol and tobacco, infectious agents such as hepatitis C, hepatitis B, human herpesvirus 8, human papillomavirus, Epstein-Barr, and human immunodeficiency virus. They are all important risk factors with deleterious effects on the immune system and direct oncogenic capabilities.[9, 33-37] Posttransplantation, risk is largely attributed to the effects of immunosuppression.[33, 38, 39] Malignancy is now recognized as a leading cause of mortality in transplantation patients; population-based studies have revealed an overall doubling or tripling of solid organ malignancies and at least a 30-fold increase in the rate of lymphoproliferative malignancies.[9, 33]
Liver transplantation candidates undergo extensive diagnostic evaluations to rule out cardiovascular disease and primary and secondary malignancies. If diagnosed, patients may undergo treatment or continued screening. This comprises a battery of investigations, of which the bulk involve ionizing medical radiation. In the current study, 73% of all investigations involved ionizing radiation, of which almost a third involved considerable exposures. Patients were exposed to a median of 53 mSv during the study period. Based on annualized exposure rates, over half the patients were exposed to >50 mSv and 10% were exposed to almost twice the occupational RE limits for a 5-year period. Patients with HCC had significantly higher RE than patients without HCC due to more prevalent utilization of CT for diagnosis and screening, and the need for CT- and fluoroscopy-guided bridging therapies. Specifically, patients with HCC were exposed to a median of 137 mSv versus 32 mSv in patients without HCC.
To put this in perspective, exposure from background radiation is on average 2-3 mSv per year and the typical patient undergoing at least one imaging procedure in the U.S. receives 2.4 ± 6.0 mSv per year.[3, 27, 29] Nuclear power plant workers are legally protected from excessive exposure and limited to annual exposures of 20 mSv in Europe and 50 mSv in the U.S. Such protections do not exist for patients undergoing diagnostic imaging. Even in dire emergency situations, such as the recent 2011 Fukushima incident, the Japanese authorities raised the allowable annual exposure for emergency workers to 250 mSv. It is estimated that the radiation-induced risk of cancer in those emergency workers receiving a total of 250 mSv will be 1 in 40. However, as outlined earlier, liver transplantation candidates are already at a much higher baseline risk of developing de novo cancer. Coupled with the multihit hypothesis of accumulated mutations and the adverse impact of immunosuppression, the amount of medical radiation these patients receive may advance malignant transformation.[40, 41]
The total amount of ionizing radiation a candidate receives as part of their liver transplant course should not be ignored. Responsible providers need to be mindful of cumulative RE and to avoid unnecessary and excessive radiation when alternative modalities exist, or indications are unclear. Guidelines for the medical evaluation of liver transplantation candidates are well documented.[7, 8] This typically encompasses a range of studies that emit ionizing radiation to evaluate for cardiovascular disease, malignancies, and infections that would preclude transplantation.
Coronary artery disease (CAD) impacts morbidity and mortality in liver transplant candidates. According to the AASLD guidelines, dobutamine stress echocardiography is an effective screening test for occult CAD in candidates.[7, 43] Nuclear stress testing and angiography accounted for 12% of the total RE in the current study, yet did not result in a single denial for listing, or intervention (e.g., angioplasty or stent insertion). Therefore, it is important to critically evaluate a way to minimize unnecessary cardiac testing. Exposure from the nuclear pharmaceutical agents is not equal: effective RE for thallium testing is ∼40.7 mSv, whereas exposure from technetium sestamibi testing is 9.4 mSv. The latter agent should thus be considered. Cardiac catheterization should be reserved for patients with positive stress tests only if treatment is necessary to continue with evaluation.[19, 42] Computed tomography angiography (CTA) has gained attention, as it has high diagnostic accuracy to predict the absence of CAD. Recognizing that CT imaging of the chest and abdomen contributed more than half the total radiation patients were exposed to in the current study, we caution against the indiscriminate use of CTA as an additional diagnostic tool rather than an alternative to catheterization and stress echocardiography.
Chest CT accounted for 11% of total RE. These studies are superior to chest radiographs for diagnosis of chest disease but not necessarily a more effective surveillance strategy even in transplantation candidates. Indeterminate and solitary pulmonary nodules in transplant candidates mandate precise clarification because either malignancy or infection will be exacerbated by immunosuppression. However, the risk of having a pulmonary nodule does not necessitate CT screening in all candidates. In one study, 9 of 152 (5.9%) HCC candidates were diagnosed with a solitary nodule on chest radiography. On chest CT, three had no corresponding nodule, two had pulmonary tuberculosis, and CT was nondiagnostic in four. Three of these patients subsequently underwent thoracoscopic biopsy for which an infectious agent was identified, and one was treated empirically for tuberculosis with appropriate response. Primary lung cancer or HCC metastasis were not identified in any of the studied patients. The incidence of solitary pulmonary nodules was low in a patient population with HCC and at increased risk of tuberculosis, and CT identified no corresponding nodules or was nondiagnostic in 78%. Routine chest CT screening is not supported in the transplantation population.
Screening and treatment of HCC resulted in high levels of ionizing RE. Although safety and technical measures can potentially be taken to reduce exposure during CT- and fluoroscopic-guided treatment of HCC, these therapies are necessary to prevent progression when wait times are greater than 6 months, or to down-stage patients. Such interventional procedures during the evaluation and waitlist period were responsible for a quarter (26%) of the total RE. However, greater steps can clearly be taken to reduce exposure from screening for HCC and metastases. Both the AASLD and European Association for the Study of the Liver (EASL) guidelines provide size-based algorithms for the diagnosis of HCC.[8, 45] In screening for HCC, US should be the first-line strategy, even in transplantation candidates. For lesions less than 1 cm, cross-sectional imaging is not required or typically diagnostic, and US screening at 3- to 4-month intervals is recommended. For lesions >1 cm, either a CT scan or MRI is an essential diagnostic tool because of the highly specific characteristics of HCC on contrast-enhanced imaging.[47, 48] However, it is important to recognize that a 4-phase CT scan exposes patients to four times the radiation of a single phase scan. A multiphase MRI has no RE and is superior to CT in detection and characterization of HCC in cirrhosis patients. We recommend that CT be reserved for situations where it is absolutely necessary, as this is a clear opportunity to reduce exposure to ionizing radiation given the superiority of other diagnostic modalities.
Finally, the regular use of CT chest and bone scans to detect extrahepatic metastases need to be reconsidered given that 76% of HCC patients with local disease treated with locoregional therapies die without metastases. In a multicenter study of 117 patients with stage I or II HCC being evaluated for liver transplantation, routine use of chest CT and bone scans neither demonstrated metastatic disease nor impacted listing decisions. Twenty-nine patients had indeterminate scans requiring repeat imaging or, in six individuals, invasive procedures. None of the latter six patients were found to have metastases, and one died as a result of the biopsy procedure. In an investigation of bone scans in patients with hepatobiliary tumors, only 1 of the 47 patients with HCC had a positive bone scan. These studies should be reserved for patients with risk factors, thereby further minimizing unnecessary RE.
This study is limited by its retrospective design, but the comprehensive care that patients receive from a single team of collaborating providers make it unlikely that a significant number of studies were overlooked. We did not have access to accurate RE data for cardiac and ERCP procedures, both of which have a wide range of exposures. This is especially significant for patients with primary sclerosing cholangitis, who may have required lengthy therapeutic ERCP interventions. Lastly, these single-center data may not be representative of all liver transplantation programs; however, our practices adhere to standard guidelines for the evaluation and treatment of liver transplantation patients and cardiovascular disease, and is consistent with the practices of major transplant centers worldwide.[7, 8, 19-25] It will be important to validate these findings based on the experience of other institutions.
In conclusion, patients being evaluated at our center according to standard recommendations were exposed to worrisome levels of radiation, with exposure significantly greater for HCC patients. Providers must consider the benefits, the cumulative exposure, and risk. Although RE varied by period, it was clearly proportional to the length of time in evaluation or on the waitlist. It is difficult to reduce exposure from interventions necessary to bridge HCC patients to transplantation, especially given increasing wait times driven by a shortage of grafts. This is also true in the postoperative period, where patients may not be able to safely tolerate lengthy MRIs and when CT- or fluoroscopic-guided interventions may be necessary to prevent morbidity, mortality, or graft loss. However, there is ample opportunity to reduce RE during the screening and waitlist periods through the more thoughtful ordering of diagnostics. The long-term effects of ionizing radiation in these susceptible patients is yet to be defined, but the theoretical risk of increased malignancy warrants routine use of nonradiating imaging studies when possible and demands the need for better patient safety guidelines.