Journal of Gastrointestinal Surgery

, Volume 14, Issue 10, pp 1529–1535

Radiation Dose from Computed Tomography in Patients with Necrotizing Pancreatitis: How Much Is Too Much?


    • Departments of SurgeryIndiana University
  • Camilo Correa-Gallego
    • Departments of SurgeryIndiana University
  • Thomas J. Howard
    • Departments of SurgeryIndiana University
  • Nicholas J. Zyromski
    • Departments of SurgeryIndiana University
  • Michael G. House
    • Departments of SurgeryIndiana University
  • Henry A. Pitt
    • Departments of SurgeryIndiana University
  • Atilla Nakeeb
    • Departments of SurgeryIndiana University
  • Christian M. Schmidt
    • Departments of SurgeryIndiana University
  • Fatih Akisik
    • Departments of RadiologyIndiana University
  • Keith D. Lillemoe
    • Departments of SurgeryIndiana University
2010 SSAT Plenary Presentation

DOI: 10.1007/s11605-010-1314-8

Cite this article as:
Ball, C.G., Correa-Gallego, C., Howard, T.J. et al. J Gastrointest Surg (2010) 14: 1529. doi:10.1007/s11605-010-1314-8



Low-dose ionizing radiation from medical imaging has been indirectly linked with subsequent cancer. Computed tomography (CT) is the gold standard for defining pancreatic necrosis. The primary goal was to identify the frequency and effective radiation dose of CT imaging for patients with necrotizing pancreatitis.


All patients with necrotizing pancreatitis (2003–2007) were retrospectively analyzed for CT-related radiation exposure.


Necrosis was identified in 18% (238/1290) of patients with acute pancreatitis (mean age = 53 years; hospital/ICU length of stay = 23/7 days; mortality = 9%). A median of five CTs/patient [interquartile range (IQR) = 4] were performed during a median 2.6-month interval. The average effective dose was 40 mSv per patient (equivalent to 2,000 chest X-rays; 13.2 years of background radiation; one out of 250 increased risk of fatal cancer). The actual effective dose was 63 mSv considering various scanner technologies. CTs were infrequently (20%) followed by direct intervention (199 interventional radiology, 118 operative, 12 endoscopic) (median = 1; IQR = 2). Magnetic resonance imaging did not have a CT-sparing effect. Mean direct hospital costs increased linearly with CT number (R = 0.7).


The effective radiation dose received by patients with necrotizing pancreatitis is significant. Management changes infrequently follow CT imaging. The ubiquitous use of CT in necrotizing pancreatitis raises substantial public health concerns and mandates a careful reassessment of its utility.


Necrotizing pancreatitisRadiationComputed tomography


Acute pancreatitis represents a continuum of disease that challenges our clinical, social, and financial management.16 Necrotizing pancreatitis is particularly virulent because it involves degradation of the pancreatic gland and/or surrounding peripancreatic tissues.46 It also increases the risk of developing acute organ dysfunction associated with severe acute pancreatitis.7 More specifically, the extent of necrosis can predict both local complications as well as the degree of overall organ degredation.8,9

The current non-invasive, gold standard modality for identifying the initial extent as well as the evolution of pancreatic necrosis is computed tomography (CT) with intravenous contrast medium. As a result, multiple CT-based classification schemes have been developed in an attempt to better prognosticate the clinical course of this disease.10,11 In addition to the inherent risk of contrast-associated nephropathy, CT imaging also exposes patients to a measurable dose of ionizing radiation.1215 Considering the frequent need for multiple CT scans during the course of necrotizing pancreatitis, a patient’s potential risk from repeated radiation exposure is substantial. The increasing use of CT imaging1621 coupled to the growing incidence of acute pancreatitis,3,2225 makes this public health issue especially topical.

Given indirect evidence suggesting that low-dose ionizing radiation is associated with the subsequent development of both solid cancers and leukemia,26 the primary goal of this study was to identify the frequency and effective radiation dose of CT imaging for patients with necrotizing pancreatitis at a high-volume pancreas referral center. A secondary goal was to identify the proportion of CT examinations that resulted in a subsequent therapeutic intervention or change in management.

Materials and Methods

The study population consisted of all patients with necrotizing pancreatitis treated at Indiana University Hospital (IUH) between January 1, 2003 and December 31, 2007. IUH is a tertiary care hospital with a high-volume referral pattern for both pancreatitis and malignant pancreatic disease. The pancreatitis database, computer-based charts, and picture archiving and communication system supplied all data in this retrospective study. CT scans were reviewed to identify all patients with pancreatic and peripancreatic necrosis (both sterile and infected necrosis were included). Operative reports were also reviewed. Patients with equivocal radiologic findings of pancreatic necrosis on CT imaging who did not undergo surgical debridement were excluded. Therapeutic interventions following CT imaging were defined as: (1) operative pancreatic debridements/necrosectomies (with or without a cholecystectomy); (2) percutaneous drainage, feeding access placement, and/or angioembolization by interventional radiologists; and (3) ductal imaging, sphincterotomy, and/or pancreatic duct stent placement using endoscopic retrograde cholangiopancreatography (ERCP). Interventions occurring within 96 h of imaging were considered to have been influenced by that CT scan.

IUH utilized four different CT scanners during the study interval. From January 1, 2003 to June 17, 2003 a Phillips four channel (Mx8000 CT Twin 7180 Gantry) was employed. Between June 17, 2003 and March 5, 2004, a Phillips 16 channel (Mx8000 IDT Gantry) was utilized. Between March 5, 2004 and September 2, 2008 both a Phillips 40 channel (Brilliance 40 Gantry with DMS) and a Phillips 64 channel (Brilliance 64 Gantry with DMS) were employed. After September 2, 2008, all CT imaging was performed on a Phillips 64 channel (Brilliance 64 Gantry with DMS). Newer CT technologies (40- and 64-channel detectors) included automatic exposure control software. Pancreatic CT studies at IUH typically included a very generous torso profile that included cross-sectional images of the majority of the pelvis.

All radiation dosing is discussed using “effective doses.” This entity is reported as Sieverts (Sv) in standard SI units [1 Sievert (effective dose equivalent) = 1 Roentgen equivalent man]. The effective dose accounts for the absorption of radiation dose and estimates the whole-body dose that is actually delivered during a radiologic procedure. As a result, this measure allows comparisons to other types of non-medical radiation exposure. Natural background radiation dose is defined as 3 mSv per year. Chest radiographs (posteroanterior) deliver an individual effective dose of 0.02 mSv.18 An increased risk of fatal cancer is calculated by multiplying the effective dose (Sv) by the risk coefficient of fatal cancer in adults.

Data analysis was performed using Stata version 8.0 (Stata Corp, College Station, TX). Normally or near-normally distributed variables were reported as means and non-normally distributed variables as medians. Means were compared using the Student’s t test and medians using the Mann–Whitney U test. Differences in proportions among categorical data were assessed using Fischer’s exact test. A p value less than 0.05 was considered to represent statistical significance for all comparisons.


Over the 5-year study period, 1,290 patients were diagnosed with acute pancreatitis. Of these, 255 (20%) displayed a necrotizing variant (38, 47, 50, 60 and 60 patients for 2003 to 2007, respectively). Seventy percent (178) of these patients required operative pancreatic or peripancreatic debridement. Seventeen patients were excluded from the analysis because of missing patient data. The mean age of the remaining 238 patients was 53 years [standard deviation (SD) = 16.1] (Table 1). Sixty-two percent were male. The cohort displayed a mean hospital and intensive care unit (ICU) length of stay of 23 and 7 days, respectively (SD = 7.8 and 12.1, respectively). The overall mortality rate was 8.8%. Patients surviving to discharge were released to home with self care only (41%), home with home-care assistance (30%), long-term care hospitalization (10%), skilled nursing facility (9%), rehabilitation facility (7%), or other destinations (3%). Patients had a median imaging follow-up interval of 2.6 months [interquartile range (IQR) = 7.2) (mean = 8.3 months].
Table 1

Patient demographics

Total no. of patients


Mean age (years)


Male gender (%)


Mean hospital length of stay (days)


Mean intensive care length of stay (days)


Overall mortality (%)


A median of five abdominal CT examinations per patient (IQR = 4) were completed during their initial hospitalization and outpatient follow-up interval. The average effective dose of this imaging load is 40 mSv (Table 2).18 By isolating the effective doses delivered by each specific CT scanner over the course of the study period (four-, 16-, 40- and 64-channel machines), the average effective dose among the study cohorts at IUH was actually 47 mSv per patient. Thirty-five percent of all CT scans also used two phases (arterial and venous). This increased the delivered average effective dose to 63 mSv per patient (Table 2). The mean ICU length of stay increased concurrent to the number of CT examinations in a stepwise manner (R = 0.69).
Table 2

Comparison of radiation equivalents for a median of five abdominal CT scans

Scanner-specific effective dose (mSv)

Total effective dose (mSv)

Chest radiograph equivalents

Equivalent background radiation time (years)

Increased risk of fatal cancer





















aDoses for study patients when dual-phase and scanner technologies are included

bUnited States mean abdominal CT scan effective dose

cUnited Kingdom mean abdominal CT scan effective dose

dDoses using the lowest reported effective dose possible for an abdominal CT scan

The median number of post imaging interventions was one (IQR = 2) with a resultant CT/intervention rate of 20%. Of the 1,202 total CT examinations performed, 531 (44%) were completed in patients considered to be physiologically ill at the time of imaging (ICU admission with sepsis and/or organ failure). This patient subset displayed a higher (31%) CT/intervention rate compared to patients without acute physiologic illness (p < 0.001). Postimaging interventions included 189 (57%) interventional (non-angiography), 118 (36%) operative, 12 (4%) ERCP, and ten (3%) angioembolization procedures. MRI scans (78 patients) did not have a CT-sparing effect as these patients still underwent a median of five CT scans. The median number of CT scans for patients who underwent an initial operative intervention was similar to patients initially managed with non-operative [interventional radiology (IR) or gastroenterology] techniques (p = 0.19). The time interval between CT scans for patients who underwent a post-imaging intervention was similar to those who underwent no procedures (IR or ERCP) (p = 0.11). CT examinations performed after discharge from the hospital were indicated for evaluation of a known pancreatic fistula (72%), interval follow-up (19%), or for unclear reasons (9%).

The mean direct hospital cost increased in a stepwise manner with the number of CT examinations obtained (R = 0.72). These increased from a mean of US $14,831 with one CT scan to US $67,470 with ten scans. The cost of performing as well as interpreting a CT scan for pancreatitis at IUH is US $600–1,200 (charge to insurer). The mean hospital (variable direct) costs for patients with necrotizing pancreatitis are approximately threefold higher than for non-necrotizing acute pancreatitis. Radiology costs account for 5% of the total hospital cost in all cases of pancreatitis.


Although CT technology was invented in 1971,27 recent improvements in scanner speed, image resolution, and ease of use have created a veritable explosion in both applications and indications.28,29 In 1980, approximately 3 million CT scans were performed in the United States, compared to 62 million in 2006.17 This change has led to a nearly sixfold increase in the per capita radiation exposure from medical imaging.13 The revolution in spiral CT technology is also evident in terms of the absolute number of scanners. As of 1996, the United States and Japan had 26 and 64 machines per 1 million people, respectively.16 Based on its utility for a broad range of screening endeavors, from evaluating seasoned astronauts for cardiovascular disease30 to identifying occult injuries in severely injured blunt trauma patients,31 CT use is again primed to increase. More specifically, interest in CT colonography,32,33 CT lung screening for smokers,34,35 coronary artery CT screening,36 and whole-body health screening examinations37,38 is significant.

While the majority (80% to 85%) of human radiation exposure arises from equal amounts of solar and radon sources (background dose = 1 to 3 mSv per year), medical imaging creates most of the remaining 15% to 20%.12,28,39,40 Of all CT imaging, 75% is obtained in a hospital setting, with up to half being scans of the torso.41 Furthermore, abdominal CT imaging accounted for up to 31% of the annual cumulative effective dose from medical imaging procedures in a study of nearly 1 million non-elderly adults.13

The stochastic risk of DNA mutations and therefore carcinogenesis (solid organ, thyroid, leukemia) following exposure to medical imaging currently assumes a linear, no-threshold extrapolation model based on data from the Japanese atomic bomb survivors (organ doses compared to organ-specific cancer incidence).1215,42 This dose-biologic effect relationship is the subject of significant controversy given its reliance on risks extrapolated from high doses as well as the possibility that it overestimates risk by ignoring the human body’s natural defense mechanisms against radiocarcinogenesis at low doses.4346 Unfortunately, no large-scale epidemiologic data exist to confirm the potential cancer risks associated with CT imaging using this conservative approach.47

In addition to the unclear oncologic risks of medical imaging exposure, the delivered effective dose can vary significantly based on the individual CT scanner (i.e., number of “slices”). The reported effective dose for a standard single-phase abdominal CT scan ranges from 1.5 to 10 mSv depending on the number of channels.12,20 If the generally recognized average effective dose of 8 mSv is utilized, our patients would have been exposed to a mean of 40 mSv. This exposure is classified as a high annual dose, with less than 1% of the United States population being exposed to greater than 20 cumulative mSv per year.13 In comparison, exposure for both pilots/flight crews (1,000 flight hours per year) and occupational radiation workers approximate 5 mSv per year. It also far exceeds goal occupational radiation exposure levels defined by the International Commission on Radiological Protection guidelines.48,49 Recent estimates of the lifetime risk of radiation-induced cancer approximate one person in 100 for those exposed to 100 mSv (relative risk = 1.024)(Table 1).50 The lifetime risk of cancer from all other causes is 42 in 100, and the risk of dying from a motor vehicle crash in the United States is one in 77.26,5053

The need for repeat CT imaging in the same patient extends beyond pancreatitis. Mettler and colleagues reported that among all patients in the literature undergoing CT imaging, 30% underwent at least three scans (7% underwent more than five and 4% more than eight scans).20 Given evidence that radiation exposure is more harmful in younger patients,13 the best studied adult population is the trauma cohort. The number of CT examinations in a subset of severely injured patients who spent at least 30 days in the ICU (mean injury severity score (ISS) = 32) was 7.8.40 Similarly, a study of 172 trauma patients with a mean ISS of 23 used multi-site dosimeters to identify a mean effective dose of 22.7 mSv.53 This led to an extrapolation of 190 cancer deaths per 100,000 patients exposed to imaging studies following major trauma.53

Given the calculated effective dose of 40 mSv in our necrotizing pancreatitis patients (assuming a similar life expectancy), we predict significantly more deaths from radiation-induced cancer. Although this estimate accounts for the fact that medical radiation exposure tends to accumulate in ill patients with an inherently reduced life expectancy (less time to manifest radiation-induced cancers), the precise relationship between trauma patients and those with severe acute pancreatitis is unknown (mean age = 53 vs. 43% of all injured U.S. patients >45 years of age).3,54 Interestingly, the effective dose of 40 mSv is identical to that reported for patients with pancreatic cancer during their first year (40.1 mSv).15 Unfortunately, patients with pancreatic cancer have significantly shorter life expectancies than those with necrotizing pancreatitis as evidenced by a 5-year exposure of only 68.8 mSv per patient.15 It should also be noted that our estimated effective dose does not include adjunctive radiologic investigations. These procedures most commonly include fluoroscopy and angiography as well as other occasional CT studies (pulmonary emboli protocols = 15 mSv). Unfortunately, despite our center’s generally aggressive use of MRI for pancreas disease, this philosophy did not have a CT-sparing effect in this patient population. More specifically, although we have found great utility for MRI in evaluating the integrity of the pancreatic duct (i.e., diagnosing disconnected left pancreatic remnants), its utility was limited elsewhere.

The actual effective dose of our cohort was 47 mSv because of the progression from four- to 64-channel scanners over the study interval. With the application of recent dose optimization strategies such as automatic exposure control available for 64-channel scanners, the effective dose per scan has been reported to be as low as 1.52 mSv.12 Had this technology been available for each of our patients with necrotizing pancreatitis, the actual delivered effective dose could have been reduced by over 80%. Put into perspective, this would lower the lifetime risk of cancer to less than one in 15,000 per individual 64-channel CT examination. Although we used single-phase arterial-enhanced CT scans in the majority of patients (65%) to determine the extent of disease, some authors have proposed routine use of tri-phasic CT to improve the delineation between pancreatitis and cancer.55 The total effective dose would therefore need to be multiplied by the number of phases. More specifically, the effective dose in our patients was actually 63 mSv when dual-phase CTs were accounted for (Table 1). This compares to a mean of 31 mSv in a recent study of patients with acute pancreatitis.56 As a result of this variance, each center must assess its own CT technology and clinical practice in an effort to quantify the associated risk to patients with pancreatitis.

As noted by Fazel and colleagues,13 unlike the exposure of workers in health care and nuclear industries, the exposure of patients to radiation cannot be restricted. As a result, the potential stochastic risks of CT imaging must be carefully weighed against the clinical importance of each individual procedure.57,58 Although the precise real-time thought process of the ordering clinicians was unavailable for this retrospective study, the rate of post-imaging therapeutic interventions was employed as a surrogate for a change in clinical management. Only 20% of all CT scans in the cohort were followed by subsequent interventions (interventional radiology, operative, or ERCP procedures). This value echoes the 20% rate of subsequent alterations in management among trauma patients who undergo a CT scan of their chest.59 This rate increased to 31% in patients with necrotizing pancreatitis who displayed more severe physiologic illness (sepsis, organ failure with ICU admission). While the importance of an individual CT study cannot be understated, the likelihood of altering a patient’s clinical pathway based on the subsequent findings must be considered before exposing them to radiation. This includes not only adding an intervention, but also the ability to avoid a planned procedure. Although the contribution of a CT scan to the overall cost of a hospital stay for patients with necrotizing pancreatitis is relatively small, the observation of increasing direct costs concurrent to the number of CT scans in our patient cohort is notable. It not only reflects an increasing severity of illness, but also highlights the importance of cost containment by utilizing carefully planned diagnostic modalities and evidence-based therapies.

In coupling the frequency of CT imaging for patients with necrotizing pancreatitis to the increasing population incidence of acute pancreatitis, the potential risk of radiation exposure will continue to be a significant public health issue. This issue is especially important in younger patients undergoing repeated CT examinations. Attempts at increasing the awareness of this issue are ongoing.6062 Although the widespread adoption of CT imaging represents the most important advancement in the history of diagnostic imaging, strategies to reduce radiation exposure remain crucial. These include ensuring the use of automatic exposure-control software,63,64 replacing outdated scanners, and a simple reduction in the overall number of CT examinations when possible. This may be achieved, in part, by using alternative modalities such as MRI as well as limiting surveillance to single-phase studies. Equally important, a careful assessment of the likelihood of altering a patient’s clinical management based on the results of a given CT scan is essential.


We would like to acknowledge Karl Mockler for the technical assistance.

Copyright information

© The Society for Surgery of the Alimentary Tract 2010