If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Surgeon radiation dose during complex fluoroscopically guided interventions (FGIs) has not been well studied. We sought to characterize radiation exposure to surgeons during FGIs based on procedure type, operator position, level of operator training, upper vs lower body exposure, and addition of protective shielding.
Optically stimulable, luminescent nanoDot (Landauer, Inc, Glenwood, Ill) detectors were used to measure radiation dose prospectively to surgeons during FGIs. The nanoDot dosimeters were placed outside the lead apron of the primary and assistant operators at the left upper chest and left lower pelvis positions. For each case, the procedure type, the reference air kerma, the kerma-area product, the relative position of the operator, the level of training of the fellow, and the presence or absence of external additional shielding devices were recorded. Three positions were assigned on the right-hand side of the patient in decreasing relative proximity to the flat panel detector (A, B, and C, respectively). Position A (main operator) was closest to the flat panel detector. Position D was on the left side of the patient at the brachial access site. The nanoDots were read using a microSTARii medical dosimetry system (Landauer, Inc) after every procedure. The nanoDot dosimetry system was calibrated for scattered radiation in an endovascular suite with a National Institute of Standards and Technology traceable solid-state radiation detector (Piranha T20; RTI Electronics, Fairfield, NJ). Comparative statistical analysis of nanoDot dose levels between categories was performed by analysis of variance with Tukey pairwise comparisons. Bonferroni correction was used for multiple comparisons.
There were 415 nanoDot measurements with the following case distribution: 16 thoracic endovascular aortic repairs/endovascular aneurysm repairs, 18 fenestrated endovascular aneurysm repairs (FEVARs), 13 embolizations, 41 lower extremity interventions, 10 fistulograms, 13 visceral interventions, and 3 cerebrovascular procedures. The mean operator effective dose for FEVARs was higher than for other case types (P < .03), 20 μSv at position A and 9 μSv at position B. For all case types, position A (9.0 μSv) and position D (20 μSv) received statistically higher effective doses than position B (4 μSv) or position C (0.4 μSv) (P < .001). However, the mean operator effective dose for position D was not statistically different from that for position A. The addition of the lead skirt significantly decreased the lower body dose (33 ± 3.4 μSv to 6.3 ± 3.3 μSv) but not the upper body dose (6.5 ± 3.3 μSv to 5.7 ± 2.2 μSv). Neither ceiling-mounted shielding nor level of fellow training affected operator dose.
Surgeon radiation dose during FGIs depends on case type, operator position, and table skirt use but not on the level of fellow training. On the basis of these data, the primary operator could perform approximately 12 FEVARs/wk and have an annual dose <10 mSv, which would not exceed lifetime occupational dose limits during a 35-year career. With practical case loads, operator doses are relatively low and unlikely to exceed occupational limits.
Fluoroscopically guided interventions (FGIs) are frequently the preferred treatment for many vascular conditions. FGIs are increasing not only in number but also in complexity, thus requiring higher radiation doses to complete.
However, individual case data are not obtainable from these dosimeters. Operators must then infer their exposure dose from the recorded fluoroscopic dose metrics: reference air kerma (RAK) and kerma-area product (KAP). The RAK is the kinetic energy released in the medium at the interventional reference point, which is located 15 cm along the beam axis toward the focal spot from isocenter.
KAP is a measure of the total X-ray energy output of the X-ray tube and therefore a good approximation of the total X-ray energy absorbed by the patient. KAP has been shown to be a better predictor of operator dose and stochastic risk to both patient and operator compared with RAK.
Without feedback regarding radiation exposure after individual cases, the operator is unable to amend practice patterns to lower future dose. This study sought to characterize radiation exposure to surgeons during complex FGIs based on procedure type, operator position, level of fellow training, upper vs lower body exposure, and addition of external ceiling- or table-mounted shielding. A second aim of this study was to enable operators to estimate personal exposure dose after every FGI on the basis of these parameters.
Optically stimulable, luminescent nanoDot (Landauer, Inc, Glenwood, Ill) detectors were used to measure radiation dose prospectively to surgeons during FGIs. All FGIs performed in an Allura Xper FD20 hybrid room (Philips Healthcare, Andover, Mass) from March through June 2014 were included. The nanoDot dosimeters were placed outside the lead apron of the primary and assistant operators at both the left upper chest and left lower pelvis positions. Institutional Review Board approval and consent of the patient were waived for this project because it was quality improvement, not meeting the definition of human subject research, and intended only for the improvement of local processes. For each case, procedure type, RAK, KAP, relative position of the operator, level of training of the fellow (first- or second-year fellow), and presence or absence of additional shielding devices were recorded. The RAK is the kinetic energy released in the medium at the international reference point, which is a stationary point along the central X-ray beam (Fig 1). The KAP is the total energy output of the machine.
Three positions were assigned at the fluoroscopic table on the opposite side of the fluoroscopic C-arm in decreasing relative proximity to the flat panel detector (A, B, and C, respectively), with position A (main operator) closest to the flat panel detector (Fig 2). Position D was on the same side of the table as the fluoroscopic C-arm and for all cases represented the left brachial access site. The primary and assistant operators documented the percentage of the case spent in each position. An operator was assigned a position when >50% of the case was spent in that location. If the case was split with each operator spending close to 50% of the time in both position A and position B, those data points were excluded from the analysis. The primary operator (position A) controlled the fluoroscopy pedal in all cases.
The nanoDots were read using a microSTARii medical dosimetry system (Landauer, Inc) after every procedure. The nanoDot dosimetry system was calibrated for scattered radiation in an endovascular suite with a National Institute of Standards and Technology traceable solid-state radiation detector (Piranha T20; RTI Electronics, Fairfield, NJ). The calibration was performed with the C-arm in posteroanterior geometry centered on a 28.5-cm stack of 30 × 30-cm polymethyl methacrylate to simulate radiation scatter from a patient positioned on the table. The thickness of the polymethyl methacrylate was sufficient to drive the fluoroscope to typical operating parameters (80 kVp, digital acquisitions). The nanoDots were positioned 1 m perpendicular from the fluoroscopy table at approximately position A, at elevations of 25 cm above and below the fluoroscopy table surface. The T20 detectors were placed adjacent to the nanoDots, and exposures were made at radiation levels corresponding to 1, 2, and 5 Gy RAK to confirm measurement linearity.
Operator effective dose was determined using a modified Niklason algorithm, which used the nanoDot reading at the left upper chest and assumed a thyroid shield. Effective dose = 0.03 × the badge dose at the upper chest.
Comparative statistical analysis of nanoDot dose levels between categories was performed by analysis of variance with Tukey pairwise comparisons. Bonferroni correction was used for multiple comparisons.
There were 415 nanoDot measurements recorded for 114 FGIs during 4 months in a single hybrid room. The case distribution was as follows: 16 thoracic endovascular aortic repairs/endovascular aneurysm repairs, 18 fenestrated endovascular aneurysm repairs (FEVARs), 13 embolizations, 41 lower extremity interventions, 10 fistulograms, 13 visceral interventions, and 3 cerebrovascular procedures. Of those cases, one lower extremity intervention, four visceral interventions, and three FEVARs were excluded from the analysis because the primary operator spent close to 50% in both positions A and B. Position D was present in only eight cases, five of which were FEVARs. Position C was present in only six cases and had a dose of zero except for one high-dose FEVAR. Position C, when present, was a junior resident.
The mean operator effective dose for all cases was 9 μSv for position A, 4 μSv for position B, 0.4 μSv for position C, and 20 μSv for position D (Table I). For all case types, positions A and D received statistically higher effective doses than positions B and C. However, the mean operator effective dose for position D was not statistically different from that for position A. The mean operator effective dose for FEVARs was statistically higher than for other case types (P < .03), 20 μSv at position A and 9 μSv at position B (Fig 3).
KAP and RAK correlated similarly with operator effective dose (Fig 4). For all cases, the regression coefficient of operator effective dose vs KAP was 25 ± 2 × 10−6 μSv/mGy-cm2 for position A and 14 ± 2 × 10−6 μSv/mGy-cm2 for position B. The regression coefficient also depended on case type (Table II). Fistulograms and lower extremity procedures had an operator effective dose/KAP rate of approximately twice that of the other procedures (Fig 5). Cerebrovascular interventions had a 10-fold higher operator effective dose/KAP in position A than in position B; however, there were only six of these procedures. For fistulograms and visceral interventions, the rate of operator effective dose per KAP is the same for positions A and B.
Table IIRegression coefficients for operator dose vs kerma-area product (KAP) by procedure type
The table-mounted shielding skirt, when used, was placed adjacent to positions A to C on the opposite side of the fluoroscopic C-arm. The addition of the table-mounted shielding skirt significantly decreased the lower body dose (1.0 to 0.2 mGy) but not the upper body dose (0.3 to 0.2 mGy) (Fig 6). The use of ceiling-mounted shielding did not affect nanoDot dose. No difference in operator dose at position A or position B was observed on the basis of the level of fellow training.
Sources of radiation exposure for the interventionalist and staff include the primary X-ray beam, X-ray tube leakage radiation, and scattered radiation. The primary beam is in the direct line between the X-ray tube and the detector. Rarely, the operator's hand can be in the primary beam, resulting in doses in the range of 5 to 20 mGy/h.
Avoidance of the primary beam is paramount for the interventionalist.
Predominant operator exposure is from leakage and scatter radiation. Leakage radiation levels are regulated to a maximum of 0.88 mGy air kerma per hour measured at 1 m from the X-ray tube at maximum peak kilovoltage and current. Under typical conditions, the leakage rate is lower than this level, in the range of 0.001 to 0.01 mGy/h.
Scattered radiation levels are generally much more significant. Scatter radiation is produced within the patient's tissue when it is exposed to the primary X-ray beam. These X rays travel in all directions originating at the patient, with the highest level of scatter radiation occurring at the entry point of the patient's skin. Scattered radiation air kerma rates adjacent to the patient range from 1 to 10 mGy/h.
In this study, FEVARs generated the highest operator dose to both A and B positions. This is because FEVARs are complex procedures that often involve acute gantry angulations and prolonged operative time and have been shown to generate higher RAK and KAP levels compared with other endovascular procedures.
who found an average operator effective dose of 38 μSv per FEVAR. They also noted significant variability, with the most experienced operator receiving an average dose of 68 μSv and more junior interventionalists receiving 172 μSV per FEVAR. All the FEVARs in this series were performed by one experienced interventionalist, and therefore no variations of dose from learning curves were seen.
described a scatter “radiation cloud” that causes significant variation in dose, depending on the operator's position around the perimeter of the angiographic table. We found that the primary operator in position A receives approximately twice the dose of the assistant operator in position B. This is similar to what Haqqani et al found when irradiating cadavers. In this work, the assistant operator (position B) received 23% to 46% of the main operator's dose (position A).
In our project, although the majority of procedures demonstrated a two to one dose difference between positions A and B, we did note some deviation from this distribution in certain procedures. For example, cerebrovascular interventions had a 10 times higher operator dose in position A than in position B. However, there were so few of these cases performed in the study period that no firm conclusions can be drawn from these data.
Fistulograms were performed with the patient's arm extended at 90 degrees from the fluoroscopic table. These procedures had similar operator doses for both positions A and B. This is likely related to the clustering of both operators around the arm and therefore the X-ray beam due to the nature of the procedure. For fistulograms, the positions do not strictly adhere to the same A and B positions as other procedures do, and the two operators are more accurately represented by two A positions because one operator was above and one operator was below the arm, making both positions equally close to the X-ray source. Visceral interventions also had similar A and B doses, which may be the result of sharing of positions between the primary and assistant operators, although more cases are needed to confirm this assumption.
When present, position D is a high-dose position. Few cases, however, used position D, and therefore it is not possible to discriminate between doses received at position A vs position D. Position C doses were undetectable except for one high-dose FEVAR. This can be explained by the dose reduction related to the inverse square law: as X rays exit the source, there is an exponential decrease in the number of X rays per unit area as the distance from the source increases (1/r2).
However, in this series, RAK and KAP correlated similarly to operator dose. These two dose metrics are related to each other because KAP = RAK multiplied by the area of the collimated X-ray beam. If the use of collimation and magnification is similar between operators and procedure type, then RAK and KAP are expected to be similar. We chose to use KAP for the analysis to remain consistent with the literature and to allow comparison with other studies. Our results indicate that fistulograms and lower extremity procedures have approximately twice the rate of operator dose per KAP compared with the other procedures, including FEVARs. This is likely because these procedure types are focused on extremities, and therefore the reabsorption of scatter radiation by the patient is diminished because of the smaller body part examined compared with procedures focused on the thoracoabdominal area. This is consistent with the study of Ingwersen et al,
who found that upper and lower extremity FGIs resulted in higher operator radiation doses compared with percutaneous coronary interventions. Although these FGIs have high rates of dose to the operator per KAP, the overall operator dose for these procedures remained low in this study because they tended to be less complex with lower procedural doses. The calculated regression coefficients provide operators a method to estimate their personal dose after an FGI based on position and procedure type.
External shield placement has been shown to reduce operator dose by up to 80% in interventional cardiology procedures.
In the current study, however, the ceiling-mounted external shielding did not decrease operator dose. This is likely the result of inconsistent or inappropriate use of the shield. To be effective, the shield must be routinely positioned between the patient's skin entrance of the X-ray beam and the operator. We noted significant variability between operators and cases in how the external shield was employed. Interventionalists should be reminded to adjust the shield with changes in gantry angulation and position of the patient. We agree that a more consistent use of the ceiling-mounted shield should result in decreased operator dose.
Alternatively, the addition of the table-mounted skirt did result in a significant decrease to the operator's lower body but not upper body dose. Previous work has demonstrated that the highest scatter is below the fluoroscopy table, and accordingly, the highest operator dose is to the lower body.
irradiated a phantom to simulate FGIs, they found that the highest operator exposure rates were at the waist and knee levels without the use of the lower body table-mounted skirt. With the table-mounted skirt employed, the exposure rates at the waist dropped off dramatically. The upper body dose, however, did not decrease as significantly. The scatter from the primary beam hitting the patient reaches the operator's lower body directly but must travel through the patient and is exponentially attenuated for the upper body, resulting in differences in upper and lower body operator dose measurements.
Ocular dose and the risk of cataract formation are growing concerns in complex endovascular procedures. Recent recommendations from the International Atomic Energy Agency suggest lowering of ocular occupational limits from 150 mSv per year to 20 mSv per year.
The nanoDot dose at the left upper chest is essentially equal to ocular dose. In this study, ocular dose limits would be exceeded without the use of leaded eye protection. However, with leaded eye protection, the risks are reduced well below occupational dose limits, and the predominant concern remains staying within annual limits for effective dose.
We found no difference in operator dose at position A or position B based on level of fellow training when the fellow was the primary operator (position A). This is consistent with the findings of Patel et al,
who demonstrated no difference in operator dose when the trainee was acting as the primary operator during endovascular aneurysm repairs. Although previous reports have suggested that learning curves play a role in radiation dose,
it is possible that we found no difference because all of our fellows receive the same radiation safety training and are equally equipped to apply techniques to limit dose. Alternatively, the results could have been confounded by the attending surgeon's spending closer to 50% of the time in position A. Operator position was assigned either A or B when the operator was in that position for >50% of the case. If the case was split 50/50, the data from that case were excluded. However, some cases may have been close to a 50%/50% split for positions A and B, which could mask differences between the two positions and potentially eliminate any difference between first-year and second-year fellows.
Another study limitation that may have confounded the data is that the study was relatively small, with some case types having few procedures and only a few surgeons included. In addition, operating surgeons were aware that their radiation exposures were being examined. Changes in practice patterns mitigating operator dose (ie, increased use of the table-mounted skirt) may have occurred. It is also possible that operators paid extra attention to limiting dose during FGIs, and therefore the operator exposure rates may be lower than they would have been had the surgeons not been aware of dose monitoring.
The patient's body mass index and the case's complexity are both important determinants of the RAK and KAP of a procedure.
This study was limited because we did not control for either variable. This led to the high observed variability within procedure types and large error bars on the regression coefficients. Not controlling for these variables could also have confounded any potential differences in level of operator training on operator dose. Assuming a normally distributed patient population across body mass indexes and case complexity, our results should be generalizable to other practitioners performing similar cases.
Another limitation of the study is the calculation of effective dose. The effective dose estimations would have been more accurate with the addition of a dosimeter under the lead apron at the waist level. The nanoDots, however, were not sensitive enough to detect radiation at this location except for the highest dose procedures. Therefore, we used the modified Niklason algorithm that relies solely on the upper body dosimeter. As a result, operator effective doses did not accurately reflect the use of the table skirt because the table skirt was shown to affect only the lower body dosimeter in our study.
Surgeon radiation dose during complex FGIs depends on case type, operator position, and table skirt use but not on the level of fellow training. With advancements in image processing and noise reduction technology as in the new Allura Clarity system (Philips Healthcare), both the patient's and operator's dose can be lowered while superior image quality is maintained.
New image processing and noise reduction technology allows reduction of radiation exposure in complex electrophysiologic interventions while maintaining optimal image quality: a randomized clinical trial.
Until this technology is widely available, our data can help guide practitioners who are operating on standard fluoroscopic systems. On the basis of these data, the primary operator could perform approximately 12 FEVARs per week and have an annual dose <10 mSv, which would not exceed lifetime occupational dose limits during a 35-year career. Excluding FEVARs, with the case mix described here, the primary operator could perform roughly 40 complex endovascular procedures per week and stay well within regulatory limits. With practical case loads, operator doses are relatively low and unlikely to exceed occupational limits.
Conception and design: MK, GA, JG
Analysis and interpretation: MK, GA, JG, JA
Data collection: MK, RV, CT
Writing the article: MK, GA, JG
Critical revision of the article: MK, GA, JG, RV, JA, CT
Final approval of the article: MK, GA, JG, RV, JA, CT
New image processing and noise reduction technology allows reduction of radiation exposure in complex electrophysiologic interventions while maintaining optimal image quality: a randomized clinical trial.
The CME exam for this article can be accessed at http://www.jvascsurg.org/cme/home.
Author conflict of interest: none.
The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest.