Quantifying Radiation Doses in Common Endoscopic Retrograde Cholangiopancreatography Scenarios: An Evaluation of Radiation Safety Measures in a Real-World Environment

Abstract

Introduction

Radiation safety training remains variable among gastroenterologists performing endoscopic retrograde cholangiopancreatography (ERCP). This study sought to ascribe dosimeter readings to various real-world ERCP scenarios to provide data supporting the 3 pillars of radiation safety: distance, time, and shielding.

Methods

An ERCP fluoroscopy unit was used to generate radiation scatter from 2 differently sized anthropomorphic phantoms. Radiation scatter was measured at various distances from the emitter, with and without a lead apron, and at various frame rates (measured in frames per second, fps) and degrees of fluoroscopy pedal actuation. An image quality phantom was used to assess resolution at various frame rates and air gaps.

Results

Increasing the distance resulted in a decrease in measured scatter (from 0.75 mR/h at 1.5 ft to 0.15 mR/h at 9 ft with the average phantom and from 50 mR/h at 1.5 ft to 3.06 mR/h at 9 ft with the large phantom). Depressing the fluoroscopy pedal less frequently, or decreasing the frame rate (ie, increasing the time per frame), resulted in a linear decrease in scatter (from 55 mR/h at 8 fps to 24.5 mR/h at 4 fps and 13.60 mR/h at 2 fps). Providing shielding through the presence of a 0.5-mm lead apron reduced scatter (from 4.10 to 0.11 mR/h with the average phantom; from 15.30 mR/h to 0.43 mR/h with the large phantom). However, decreasing the frame rate from 8 fps to 2 fps did not change the number of line pairs identified on the image phantom. A greater air gap increased the number of line pairs resolved.

Conclusions

Implementing the 3 pillars of radiation safety led to a quantifiable, clinically significant decrease in radiation scatter. The authors hope that these findings spark greater implementation of radiation safety measures among practitioners utilizing fluoroscopy.

Introduction

Medical radiation has revolutionized diagnostic and therapeutic approaches in many fields within medicine, including gastroenterology. The development of fluoroscopy has allowed the widespread availability of endoscopic procedures such as endoscopic retrograde cholangiopancreatography (ERCP). ERCP has become an important lifesaving procedure ever since the first ERCP test was performed in 1968, with an estimated 156,874 inpatient ERCPs performed in the United States in 2016. ¹ However, the increased use of such radiologic procedures has highlighted the risks of both the stochastic and deterministic effects associated with prolonged radiation exposure, such as malignancy, cataracts, hair loss, and skin or mucosal injury. ² In an effort to protect both patients and practitioners, radiation safety guidelines for ERCP have been published, including one by the American Society for Gastrointestinal Endoscopy. ³ However, radiation safety training remains highly variable within the field of gastroenterology beyond a general recommendation to adhere to the “ALARA” principle (as low as reasonably achievable). A recent survey of 159 endoscopists who perform ERCP noted that that over half (56.6%) did not receive formal training in operating their facility’s fluoroscopy system and only 33% consistently modified their fluoroscopy settings to minimize radiation. ⁴ Additionally, there is high variability in the application of fluoroscopy amongst practitioners; for example, a study attempting to establish benchmarks for fluoroscopy time found significant variability in fluoroscopy time among practitioners. ⁵ Part of the challenge in promoting radiation safety lies in the abstract nature of radiation doses during common ERCP scenarios. Thus, this study aimed to bring attention to radiation safety by ascribing discrete dosimeter readings to various clinical scenarios in ERCP in an effort to provide concrete data supporting the radiation safety principles of distance, time, and shielding.

Methods

The study did not involve any human or animal subjects so IRB approval was not needed.

Design

The study was conducted with the Radiation Safety department of Kaiser Permanente Los Angeles Medical Center, a tertiary referral hospital, using the gastroenterology department’s fixed fluoroscopy unit (Philips Easy Diagnost Fluoro System, Philips Healthcare, Andover, Massachusetts) for data generation. The study did not involve any human or animal subjects so IRB approval was not needed. The experiments were performed in the ERCP fluoroscopy suite, a room with walls measuring up to 9 ft from the emitter. This study used 2 differently sized anthropomorphic phantoms (to simulate 2 patients of differing body habitus: “average,” or 22-cm anterior/posterior diameter, and “large,” or 32-cm anterior/posterior diameter); the phantom was set up as if undergoing an actual ERCP test with a blocking drape in place. The default settings for the average phantom (RS-310 Chest/Lung and Pelvic Phantoms, Radiology Support Devices, Inc, Long Beach, California) were 15-in. magnification and a 0-in. air gap with a dose of 70 kVp and 0.4 mA. The default settings for the large phantom (Fluke RS-310 Chest/Lung and Pelvic Phantoms, Radiology Support Devices, Inc, Long Beach, California) were 15-in. magnification and a 0-in. air gap with a dose of 107 kVp and 0.8 mA. The scattered radiation was measured with an ion survey meter in milliroentgen per hour (mR/h) at 18 in. to represent the operator’s position, 72 in. to represent the nurse and anesthesiologist positions, and 108 in. to signify the room border; both shielded (ie, with the 0.5-mm lead apron present) and unshielded measurements were obtained at the levels of the thyroid (5 ft above the ground) and gonads (3 ft above the ground). A solid-state detector was used to measure the entrance exposure rate. Image quality was also assessed in a semiquantitative fashion with an image quality phantom (TOR 18FG Fluoroscopy Phantom, Leeds Test Objects Ltd., York, United Kingdom) by assessing the number of line pairs identified on a fluoroscopy image at 0-in. and 18-in. air gaps; that is, the number of bar patterns on the phantom that were visible in the fluoroscopy image was recorded at the 2 air gap distances. Data were gathered at various frame rates and various levels of magnification and air gap. See Figure 1 for a diagram of the ERCP suite.

Figure 1:

Diagram of the endoscopic retrograde cholangiopancreatography (ERCP) fluoroscopy suite used for the study. “Blocked” refers to the area of the room where the blocking drape obstructed emissions from the emitter; all distance measurements are in feet. The symbols for “Operator,” “Anesthesiologist,” and “Nurse” denote the general locations where these individuals would be located for an ERCP procedure (at 1.5 ft, 6 ft, and 9 ft, respectively). Radiation scatter measurements were taken at these locations in this study.

Results

Distance

Decreasing scatter radiation was noted at greater distances from the emitter. For the average-sized phantom undergoing fluoroscopy at the default settings (70 kVp, 0.4 mA), the unshielded scatter at the level of the thyroid was 0.75 mR/h at 1.5 ft, 0.28 mR/h at 6 ft (at the general location of the anesthesiologist), and 0.15 mR/h at 9 ft; this was a 62.7% reduction from 1.5 ft to 6 ft and a 46.5% reduction from 6 ft to 9 ft (Figure 2). The magnitude of the reduction from 6 ft to 9 ft was consistent with the inverse-square law (0.13 mR/h in this study vs the expected value of 0.15 mR/h); however, the magnitude of the reduction from 1.5 ft to 6 ft was not consistent with inverse-square law (0.47 mR/h in this study vs the expected value of 0.70 mR/h). For the large phantom undergoing fluoroscopy at the default settings (107 kVp, 0.8 mA), the unshielded scatter at the level of the thyroid was measured at 50.00 mR/h at 1.5 ft, 9.20 mR/h at 6 ft (at the location of the anesthesiologist), and 3.06 mR/h at 9 ft; this was an 81.6% reduction from 1.5 ft to 6 ft and a 66.7% reduction from 6 ft to 9 ft (Figure 3). Much as for the average phantom, the reduction in radiation scatter from 6 ft to 9 ft was close to expected value based on the inverse-square law (6.14 mR/h in this study vs the expected value of 5.11 mR/h), whereas the reduction from 1.5 ft to 6 ft was not (40.80 mR/h vs the expected value of 46.88 mR/h).

Figure 2:

Graphical representation of the effect of distance from the emitter on measured radiation with the average-sized phantom. All measurements were taken at the level of the thyroid (5 ft above the ground). mR/h = milliroentgen per hour.

Figure 3:

Graphical representation of the effect of distance from the emitter on measured radiation with the large-sized phantom. All measurements were taken at the level of the thyroid (5 ft above the ground). mR/h = milliroentgen per hour.

Time

Fluoroscopy Time

With all other settings remaining constant, fluoroscopy time showed a linear relationship between pedal actuation and scatter at 1.6 mR/h for the average-sized phantom and at 96 mR/h for the large phantom, suggesting that modulation of pedal actuation will result in a linear decrease in radiation exposure.

Frame Rate

For the same default settings, decreasing the number of frames per second (fps) during fluoroscopy resulted in a largely linear decrease in the scatter to the thyroid of the unshielded operator. The scatter was measured at 55.00 mR/h at 8 fps, 24.50 mR/h at 4 fps, and 13.60 mR/h at 2 fps; this was a roughly 2-fold decrease with each halving of the frame rate (Figure 4). Similar linear reductions were also noted with the large-sized phantom, with the scatter to the thyroid of the unshielded operator measured to be 5400 mR/h at 8 fps, 2700 mR/h at 4 fps, and 1340 mR/h at 2 fps (Figure 5).

Figure 4:

Graphical representation of the effect of decreasing frame rate (a measure of radiation exposure time) on measured radiation with the average-sized phantom. All measurements were taken at the level of the thyroid (5 ft above the ground). mR/h = milliroentgen per hour.

Figure 5:

Graphical representation of the effect of decreasing frame rate (a measure of radiation exposure time) on measured radiation with the large-sized phantom. All measurements were taken at the level of the thyroid (5 ft above the ground). mR/h = milliroentgen per hour.

Shielding

Using the default settings, shielding the gonads of the operator with a 0.5-mm lead shield resulted in a 97.2% reduction in scatter for the average-sized phantom (4.10 mR/h vs 0.11 mR/h with shielding) and a 97.2% reduction in scatter for the large-sized phantom (15.30 mR/h vs 0.43 mR/h with shielding).

Image resolution as a function of frame rate

The number of line pairs identified on the fluoroscopy image at 8 fps vs 2 fps while keeping all other factors identical (dose of 74 kVp and 0.5 mA, magnification of 15 in., and air gap of 18 in.) was unchanged at 11.

Image resolution as a function of air gap

Under a continuous frame rate and stable settings (dose of 70 kVp and 1.7 mA, magnification of 15 in.), the number of line pairs identified with a 0-in. air gap was 4 compared to 8 identified with an 18-in. air gap.

Discussion

The increased availability of non-fluoroscopic procedures such as endoscopic ultrasound and magnetic resonance cholangiopancreatography has provided reliable alternatives to diagnostic ERCP procedures (ie, to determine whether choledocholithiasis is present in an unclear clinical scenario) and has reduced some of the indications for gastrointestinal fluoroscopic procedures. ³ However, ERCP for therapeutic purposes remains one of the most frequently performed procedures within interventional endoscopy, highlighting the continued importance of radiation safety even with the advent of the aforementioned non-fluoroscopic approaches. Despite the high volume of ERCP procedures being performed today, formal education regarding radiation safety with such procedures remains inconsistent. Indeed, the United States does not have uniform federal requirements dictating radiation safety training for physicians, with state laws ranging from formal education as part of the licensure process along with biennial continuing education requirements to a lack of clear radiation safety education guidelines for X-ray operators altogether. ^3–5 Although there are data in the gastroenterology literature highlighting the efficacy of techniques such as limiting fluoroscopy time and use of shielding in reducing radiation exposure to both patients and practitioners, recent surveys in the United States and other countries suggest that key concepts of radiation safety are still not prioritized within the field of gastroenterology at both the trainee and practitioner levels. ^{4, 6–10} In contrast, cardiology has a more overt emphasis on the topic. Indeed, the American College of Cardiology highlighted the importance of radiation safety training in their Core Cardiology Training Symposium guidelines and formally recommended that all interventional cardiology training programs incorporate education on this topic. ¹¹ Given this discrepancy, this article aimed to bring attention to the efficacy of the main pillars of radiation safety (distance, time, and shielding) within the field of gastroenterology by being one of the few studies within the gastroenterology literature that provides concrete values of the impact of each measure in a real-world procedural setting.

This study clearly demonstrated the effectiveness of each measure in a clinical setting. Increasing distance from the emitter from 1.5 ft to 6 ft led to a reduction in scattered radiation of more than 60% while use of shielding with a 0.5-mm lead apron led to a 97.2% scatter radiation reduction. Of note, the distance data did not closely reflect the inverse-square law as would have been expected. As inverse-square law is limited by the assumptions that the environment is a free field (that is, no obstructions or boundaries) and that the emitter is perfectly omnidirectional, the measurements could have been affected by a variety of environmental factors including room geometry and presence of various obstacles in the room (as the room was in clinical use during the study period) and by variability with the radiation emitter itself. Decreasing fluoroscopy time by depressing the pedal less often resulted in a linear decrease in radiation exposure and decreasing frame rate led to a nearly 50% reduction in scatter with each halving of the frame rate. Also, image resolution (as measured by line pairs identified on the phantom) did not suffer with a decrease in 8 fps to 2 fps, suggesting that lower frame rates can be employed without clinically significant deterioration in image quality. Of note, this study also showed that increasing air gap may increase the number of line pairs resolved due to its physical property of geometric magnification. This may be clinically useful provided the operator is mindful of other dose-saving techniques. It is important to note that the findings reported herein can also be applied to other medical specialties that utilize fluoroscopic procedures as they are a reflection of the ALARA principles that form the cornerstone of general radiation safety. Additionally, many of the adjustable settings on the ERCP device used can be found on other fluoroscopy devices (ie, cardiac catheterization or interventional radiology machines), though image quality needs may differ from specialty to specialty.

This study does have limitations, with the primary concern being one of generalizability. As the goal of this work was to gather data representative of actual clinical scenarios, the actual degree of scatter reduction for each intervention may differ at other institutions, given various environmental factors (eg, the model of fluoroscopy equipment used, room geometry, phantom and drape positioning) unique to the facility used. Additionally, the effect of contrast media on the intensity of fluoroscopy required was not examined in this study. Despite these limitations, the general findings reported herein are consistent with previously published literature and encourage adherence to safety guidelines created by various governing bodies. ^{3, 5–8}

Given that all institutions with ERCP capabilities must (by law) have radiation safety personal protection equipment available, it seems that one of the limiting factors in more widespread adoption of the aforementioned practices is education. Reinforcement of these core safety principles through a short, standardized education course may help promote a culture of radiation safety; a study evaluating the effect on ERCP endoscopists of a 20-minute audiovisual educational presentation detailing best practices in fluoroscopy found that education was associated with a significant reduction in medial fluoroscopy time (up to 48%) and total radiation dose (up to 52%). ¹² Additionally, major gastroenterology conferences have held plenary sessions regarding the incorporation of radiation safety into daily practice; regular discussions in such settings can help keep these concepts fresh for attendees. ¹³ Borrowing a page from the radiology community, simply mandating documentation of fluoroscopy doses also has the potential to reduce radiation doses through the Hawthorne effect; Vehmas et al found that having practitioners record dose–area product measurements during fluoroscopic procedures led to a significant decrease in mean fluoroscopy time by 30.8%. ¹⁴ Beyond education and awareness, technologic advancements and modifications to the fluoroscopy machines themselves can physically limit exposure. One possible approach would be modifying machine settings to limit the duration of fluoroscopy employed with each pedal actuation. A prospective study utilizing such restrictions by limiting x-ray exposure to 3 s each time the pedal was depressed found that time-limited fluoroscopy significantly reduced fluoroscopy time by 16.4%. ¹⁵ Additionally, advances in artificial intelligence (AI) have allowed for the use of AI in automatically adjusting the shutters of a secondary collimator to help keep radiation limited to the region of interest. ¹⁶ One such device is already available in the United States (FluoroShield; Omega Medical Imaging, Sanford, Florida), and a prospective study comparing this machine to a traditional ERCP fluoroscopy machine found that the AI-enhanced system reduced scatter radiation by 59.4%. ¹⁶ Continued advancements in AI could lead to its incorporation in other aspects of fluoroscopy (eg, to automatically adjusting the frame rate or air gap to provide optimal image quality while limiting exposure), making this an exciting area of research within the realm of radiation safety.

Although radiation safety remains an issue within the field of gastroenterology, the community is showing increasing interest in both discussions regarding safety and technologic advances that have the potential to limit exposure. This study is one of the few gastroenterology studies providing clear evidence of the efficacy of general radiation safety measures in a real-world ERCP setting. The authors hope that the combination of concrete published data supporting these measures, standardized and frequent education requirements, and continued technological advances will spur positive changes in the domain of radiation safety within both the field of gastroenterology and medicine as a whole.