| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Depts of Radiology (V.L.W., C.E.B., K.J.S., R.L.L., V.V., D.Z., P.S.D., G.A.T.), Urology (M.S., D.L.M., C.A.P.), and Orthopedics (D.Z.), Children's Hosp, Harvard Medical School, 300 Longwood Ave, Boston, MA 02115. Received Aug 23, 2004; revision requested Nov 5; revision received Mar 11, 2005; final version accepted Apr 8. Supported in part by Philips Medical Systems of North America, Children's Hosp Radiology Foundation and a Bridge Award from Harvard Medical School Minority Faculty Development Program. Address correspondence to V.L.W. (e-mail: valerie.ward{at}childrens.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
Materials and Methods: Institutional animal care and use committee approval was obtained. Vesicoureteral reflux was simulated in four pigs, and 48 VCUG studies were performed (24 with GCPFL, 24 with CFL). VCUG was performed at abdominal girths of 8–10 cm (group 1, simulates human newborn to 6-month-old infant), 12–13 cm (group 2, simulates 2–3-year-old child), and 15–17 cm (group 3, simulates 10-year-old child). An electronic device calculated total radiation exposure during fluoroscopy and image recording. With five-point ordinal scales, VCUG images were scored independently for anatomic conspicuity and overall diagnostic quality by two radiologists (radiologists A and B). An analysis of variance was used to compare radiation exposures and fluoroscopy times between GCPFL and CFL and to determine whether radiation exposure and fluoroscopy time were dependent on the pig's abdominal girth. The Pearson product-moment correlation coefficient was used to assess whether fluoroscopy time was correlated with radiation exposure. Anatomic conspicuity and diagnostic quality scores were compared by means of the Wilcoxon signed rank test.
Results: Results of analysis of variance revealed that GCPFL resulted in a significant reduction in total radiation exposure compared with CFL for each of the three groups (P < .05 for each comparison), and this reduction was most marked in the larger animals. There were no significant differences in diagnostic quality of the recorded VCUG images (P > .05). Anatomic conspicuity was not significantly different for groups 2 and 3, but there was a significantly higher score for GCPFL in group 1 for radiologist A (P = .04).
Conclusion: By using GCPFL in the performance of VCUG in a pediatric porcine model of vesicoureteral reflux, total radiation exposure can be reduced by a factor of 4.6–7.5 lower than with CFL, and diagnostic-quality images can be obtained.
© RSNA, 2006
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
The design of the fluoroscope and how the radiologist chooses to conduct the pediatric fluoroscopic examination determine the degree to which the total radiation exposure to a child is maintained at ALARA levels. Added filters in the path of the x-ray beam (5–9) and proper selection of high voltage and tube current by means of the automatic brightness control system of the fluoroscope as a function of the child's size are examples of these design feature changes (10,11). The removal of the antiscatter grid (11–13) and reduction of either fluoroscopy time or the number of recorded exposures (14–16) are operational choices that have been successful in reducing radiation exposure to patients.
Reduction in radiation exposure can also be achieved by using pulsed rather than continuous x-ray beam fluoroscopes. Conventional fluoroscopes produce a continuous fluoroscopy (CFL) x-ray beam at 30 image frames per second, with an effective exposure duration of 33 msec for each frame. In the early 1990s, fluoroscopes with a variable-rate pulsed fluoroscopy, or VRPFL, x-ray beam became available. With a VRPFL unit, the number of radiation beam pulses per second (pulse rate), and in some cases the duration of each pulse (pulse width), can be selected (17). Thus, VRPFL technology provides a method to potentially reduce patient radiation exposure (lower pulse rate) (18,19), maintain image contrast (same voltage), and improve image sharpness (exposure duration of less than 33 msec) (20,21). While VRPFL was initially available in cardiac catheterization or interventional radiology laboratories, its use has been expanded to genitourinary and gastrointestinal radiology, achieving either reduced patient radiation exposures (21–23) or improved image quality (24,25).
While the effects of various radiation dose reduction techniques have been reported and described above, there is little information available concerning measured entrance radiation exposure during the performance of clinical pediatric VCUG with fluoroscopic equipment that enables multiple radiation dose reduction techniques. Thus, the aim of this study was to compare grid-controlled variable-rate pulsed fluoroscopy (GCPFL) with CFL for the reduction of radiation exposure during VCUG in a pediatric porcine model of vesicoureteral reflux.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
Experimental Animal Model
The approval of our institutional animal care and use committee was obtained, and the investigation was conducted with the guidance of a veterinarian. Four female Hanford mini-pigs (Earle Parsons, Hadley, Mass) were chosen because of the relative ease of catheterization of the female pig urethra in comparison with the male pig urethra. Catheterization of the urethra and bladder was performed endoscopically by either a pediatric surgeon (M.S.) or a pediatric urologist (D.L.M.).
Prior to each fluoroscopic imaging session, an intravenous line was placed in the pig for the administration of fluids, medications, and intravenous contrast material. Anesthesia was induced with intramuscular injection of zolazepam (1.4 mg per kilogram of body weight), xylazine (2 mg/kg), and atropine (0.04 mg/kg). Additional anesthetic and analgesic medications were administered as follows: isoflurane (inhalational and titrated to effect), buprenorphine (0.03 mg/kg and then twice a day for 3 days), flunixamine (2 mg/kg and then once a day for 3 days), and a fentanyl patch (1 [µg · kg–1]/h). Each pig received intramuscular antibiotics (cefazolin, 40 mg/kg). Each pig underwent endotracheal intubation and was monitored throughout the imaging session. Oxygen saturation, carbon dioxide, heart rate, temperature, respiratory rate, and depth of anesthesia were recorded.
The submucosal segment of the distal pig ureter is reported to be relatively long and to prevent spontaneous vesicoureteral reflux in the normal pig (26). To simulate ureteric and renal reflux, the upper urinary tract was opacified by using intravenous contrast material (see Imaging Studies for details of this simulation of vesicoureteral reflux).
A total of four pigs were used in this study. The supine abdominal girth of each pig was measured at the level of each pig's umbilicus. As each pig grew and abdominal girth increased, they were each imaged at three different abdominal girth ranges (ie, three different VCUG imaging sessions were performed per pig as the pig's girth reached each of three specific girth ranges). The first girth range for imaging was 8–10 cm; this range simulates the abdominal girth of a newborn to 6-month-old human infant (27,28). All four pigs were imaged at this first girth range, and the VCUG images obtained in these four pigs at this range were subsequently referred to as group 1. The next girth range for imaging was 12–13 cm; this range simulates the abdominal girth of a young child (approximately 2–3 years old) (27,28). Each of the same four pigs were again imaged when their girths reached 12–13 cm, and the VCUG images obtained in these same four pigs at this girth range were subsequently referred to as group 2. Finally, the largest girth range for imaging was 15–17 cm; this range simulates the abdominal girth of an older child (approximately 10 years old) (27,28). Again, all four of the same pigs were imaged at this last girth range of 15–17 cm, and the VCUG images obtained in these same pigs at this girth range were subsequently referred to as group 3.
Fluoroscopic Unit Specifications
Two different fluoroscopes were used to perform the study. Both were modified to the extent possible to achieve ALARA exposures. The first unit was a CFL unit (Advantx fluoroscope; GE Medical Systems, Waukesha, Wis) with a 23-, 15-, and 11-cm image intensifier, an Advantx generator, and a DRS digital spot imaging device. This fluoroscope represents a state-of-the-art unit from 1995. Digital spot images and fluoroscopic images were acquired with a matrix size of 1024 and 512, respectively. The source-to-skin distance of the unit is approximately 50 cm. Two steps were taken by one of the authors (K.J.S.) to reduce radiation exposures to the pigs. A filter of 0.08 mm copper plus 1 mm aluminum thickness was added to the x-ray beam. In addition, the entrance exposure rates at the entrance plane to the image intensifier, without the antiscatter grid in place, were reduced to approximately 80% (25 µGy per minute reduced from 31 µGy per minute) of the manufacturer's recommendations, in accordance with published rationales (10) (Appendix A). When possible, the entrance exposure was further reduced to be proportional to the reciprocal of the diameter of the field of view, as opposed to the more traditional setting of proportional to the reciprocal of the square of the diameter (29). The actual entrance exposure rates set at the entrance to the image intensifier of the CFL unit in this study (expressed as absorbed dose in air, or kerma) are listed in Appendix A.
The image intensifier and television camera pickup tubes were replaced prior to the beginning of the study. These replacements restored the image quality back to its original level, when this CFL unit was new, and allowed for an equal comparison of image quality performance with the second unit, which was less than 1 year old at the beginning of this study.
The second unit was a GCPFL unit (Easy Diagnost 90/45 fluoroscope; Philips Medical Systems, Bothell, Wash) with a 23-, 17-, and 13-cm image intensifier, an SCP80 generator, a DSI digital spot imaging device, and both continuous and variable-rate pulsed fluoroscopy modes. This unit was installed in 2000. It represents a state-of-the-art GCPFL unit (17). With this unit, the switching of the x-ray beam "on and off" at the x-ray tube, instead of at the secondary of the high voltage transformer, more precisely controls the radiation beam, which results in additional reduction of radiation exposure (24). The unit has two operator-selectable exposure rates at the entrance plane to the image intensifier. Two tube potential–tube current curves for the automatic brightness control system were provided (one for adults and one for children). Digital spot images were acquired with a 1024 x 1024 matrix size. Two source-to-skin distance settings, 51 and 65 cm, were available.
Prior to its first clinical use, the manufacturer modified this GCPFL unit to optimize its use in imaging children. Specifically, the three selectable pulse rates were reduced to 7.5, 3.75, and 1.88 pulses per second. The reduced maximum pulse rates allowed the fluoroscopic matrix to be increased to 1024 in both dimensions. The two available pulse widths were decreased by half, to 5 and 10 msec. Additional filtration, consisting of 0.2 mm copper and 1 mm aluminum, was placed in the path of the x-ray beam.
The entrance exposure rates (kerma) at the entrance plane to the image intensifier, without the antiscatter grid in place, were changed (K.J.S.) relative to the manufacturer's recommendations to reduce radiation exposure to the pig while maintaining diagnostic image quality. The entrance exposure was set to be proportional to the reciprocal of the diameter of the field of view (29) in all modes of operation. The entrance exposure per pulse for the 7.5-pulse-per-second fluoroscopic mode was doubled to compensate for the added filter (29). This entrance exposure per pulse was further increased as indicated in the following equation for pulse rates of more than 5 pulses per second (30): exposure per pulse 
(30/pulse rate)0.5. The actual entrance exposure rates set on the GCPFL unit in this study (expressed as kerma) are listed in Appendix B.
The radiologists conducting the imaging were given instructions to standardize the operation of the CFL and GCPFL units. Only the largest pig girth group (group 3) was imaged with the antiscatter grid in place for both units. Each VCUG examination began with the radiologist using the largest field of view of the image intensifier, with instructions to use smaller fields of view as dictated according to the examination. Each VCUG examination was begun by using the fluoroscopy dose setting and recorded image dose setting most commonly used clinically for each unit ("medium" for both units). Each VCUG study with the GCPFL unit was begun by using a fluoroscopic pulse rate of 1.88 pulses per second, with the instructions to switch to faster pulse rates if dictated according to the examination. The pediatric fluoroscopic mode on the GCPFL unit, which operates at a pulse width of approximately 5 msec with a pediatric tube potential–tube current curve, was used to image the pigs in groups 1 and 2. The adult mode on the GCPFL unit, with a pulse width of approximately 10 msec and an adult tube potential–tube current curve, was used to image the pigs in group 3. A 65-cm source-to-skin distance was used in this study to match the source-to-skin distance used for all clinical VCUG studies performed with the GCPFL unit.
Imaging Studies
The path length of the radiation beam as it travels through the patient affects both the patient entrance exposure from ionizing radiation and the image quality. Therefore, this study was designed to evaluate both of these parameters at three different ranges of supine abdominal pig girth.
At each of the three different ranges of abdominal pig girth, an imaging session was performed. At each of these three imaging sessions, each of the four pigs underwent four VCUG examinations in random order. (Specifically, two VCUG examinations were performed independently by two experienced staff pediatric radiologists [V.L.W., C.E.B.] by using the GCPFL system, and two VCUG examinations were performed independently by the same two radiologists by using the CFL system. Hence, there were 12 VCUG studies per pig, or 48 VCUG studies total among the four pigs. At the time this study was conducted, these two staff pediatric radiologists, V.L.W. and C.E.B., had 4 and 10 years of experience performing VCUG, respectively.)
With intermittent fluoroscopic observation, iodinated cystographic contrast material (17% weight by volume solution; Cystoconray, Mallinckrodt, St Louis, Mo) was instilled by means of gravity in retrograde fashion into the pig's catheterized bladder (31). The actual bladder capacity—that is, the bladder capacity at which the pig first voided—was documented. The pig was also given an intravenous administration of contrast material (essentially intravenous pyleography) that served to simulate ureteric and renal reflux. Each pig received a total intravenous dose of 2 mL/kg of ioversol (Optiray-320; Mallinckrodt). Half of the total dose of ioversol was administered as a bolus just before the first VCUG study, and then the other half of the total dose was administered as a bolus just before the third VCUG study. These two boluses resulted in visualization of the upper tracts throughout the four VCUG studies. The four VCUG studies were performed over an average time course of 78 minutes. Any variability in contrast material visualization within the renal collecting systems between the GCPFL and CFL systems was controlled for, because each animal underwent two sequential VCUG studies with one fluoroscope and then was transferred to the other fluoroscope for the remaining two sequential VCUG studies (for a total of four VCUG studies at each imaging session). Hence, similar visualization occurred for each fluoroscope, because the time from injection was similar.
Fluoroscopic spot images were obtained and included the following: preliminary anteroposterior bladder view, preliminary anteroposterior kidney views, anteroposterior view of the early bladder filling phase, right and left oblique full bladder views of the ureterovesical junctions, anteroposterior collimated-down view(s) of the kidneys demonstrating the vesicoureteral reflux, and postvoiding anteroposterior view(s) of the bladder and kidneys. A cyclic VCUG study was performed—that is, at least three cycles of bladder filling and voiding—in the pigs in groups 1 and 2 because their girths simulated the size of young children (similar in age to the children in the investigation by Paltiel et al [32]). Cyclic examinations were not performed in the pigs in group 3, because they had an abdominal girth that simulated that of a 10-year-old child (therefore older than those children studied by Paltiel et al).
At the completion of an imaging session, each pig was monitored every 15 minutes until it had recovered from anesthesia. The recorded images were printed on laser film for analysis at a later time.
Radiation Exposure Measurements
During fluoroscopic imaging, the free-in-air radiation exposure rates and total radiation exposure were calculated at the entrance skin location of the pig. The calculations were performed with an electronic online patient exposure meter called PEMNET (an acronym for patient exposure monitoring network) (Clinical Microsystems, Arlington, Va) (10,33–36). Through an electronic interface to the generator of the fluoroscope, the PEMNET device sampled the following parameters every 5 msec: x-ray tube potential (kilovolts), x-ray tube current (milliamperes), x-ray beam pulse duration, and source-to-skin distance. PEMNET computed the free-in-air radiation exposure rate every 800 msec by using its sampled data and fitted polynomial exposure curves that quantify the unique x-ray output of the monitored x-ray unit. The sampled data and calculated radiation exposure results were periodically downloaded to a computer database capable of storing data from up to eight different imaging planes by two physicists (K.J.S. and V.V.).
The fitted coefficients for the exposure calibration curves for each individual x-ray tube monitored by PEMNET were calculated from free-in-air radiation exposure output measurements collected with a calibrated Keithley triad electrometer and model 96035B ionization chamber (Cardinal Health, Cleveland, Ohio). Fitted coefficients were required for each mode of operation (eg, continuous fluoroscopy, pulsed fluoroscopy, radiographic mode) for each individual x-ray tube. PEMNET, when properly calibrated, calculates the free-in-air radiation exposure at the entrance skin plane of the pig, located at the table top of each fluoroscope, with an accuracy within ±5% (10,33). Throughout the course of the study, the calibrated ionization chamber and triad electrometer were periodically used to verify the proper functioning and calibration of PEMNET (K.J.S. and V.V.). The exposure calculated by PEMNET will be subsequently referred to as "radiation exposure."
Image Evaluation
The recorded VCUG images from both fluoroscopy systems were displayed as hard-copy (film) images to two experienced staff pediatric radiologists (G.A.T., R.L.L., 19 and 32 years of experience interpreting VCUG images, respectively) who were not involved in the performance of the VCUG examinations. Annual physicist surveys of the equipment performance were used to verify that all high and low contrast resolution detected on each image intensifier was accurately recorded on the printed images. These radiologists evaluated the images independently and without knowledge of the fluoroscopy system used. All information identifying which fluoroscopy system was used to generate the images was masked.
The images were graded on two different five-point ordinal scales for anatomic conspicuity and overall diagnostic quality that were adapted from the literature (37). The anatomic conspicuity score was based on an ordinal scale of 1–5, on which a score of 1 equaled the best score and corresponded to all five of the following structures being seen on the images with an acceptable low level of noise: renal calyces, ureterovesical junction region, bladder contour, urethra, and bone trabecular pattern. If only four of these structures were seen with an acceptable low level of noise, then a score of 2 was assigned, and the radiologist specified which anatomic structure was not visualized. Similarly, if only three of these structures were seen, then a score of 3 was assigned, for two structures a score of 4 was assigned, and for one structure a score of 5 was assigned.
The diagnostic quality score was based on an ordinal scale of 1–5, on which a score of 1 indicated the highest diagnostic quality VCUG image; score of 2, above average diagnostic quality VCUG image; score of 3, adequate diagnostic quality VCUG image; score of 4, barely diagnostic quality VCUG; and score of 5, nondiagnostic VCUG image. The diagnostic quality score was a rank-order subjective score based on the scoring radiologists' consensus of diagnostic quality. No specific numeric criteria were used.
Statistical Analysis
Total radiation exposure, fluoroscopic radiation exposure, recorded image radiation exposure, and fluoroscopy time were compared between the GCPFL and CFL units by using a 2 x 2 within-subjects analysis of variance (38). Data for radiation exposure and fluoroscopy time were expressed as mean ± standard error. In addition, each of the four dependent variables (total radiation exposure, fluoroscopic radiation exposure, recorded image radiation exposure, and fluoroscopy time) were assessed by using repeated-measures analysis of variance separately for the GCPFL and CFL units to determine whether radiation exposure and fluoroscopy time were dependent on the pig's abdominal girth. The Pearson product-moment correlation coefficient was used to evaluate whether fluoroscopy time was correlated with radiation exposure.
Image quality and anatomic conspicuity median scores were compared between the GCPFL and CFL units for each observer (radiologists A and B) within each of the three groups of pig girth. This comparison was made by using the nonparametric Wilcoxon signed rank test for matched pairs based on the five-point ordinal scales described (39). The Wilcoxon signed rank test for matched pairs was conducted because the anatomic conspicuity score and diagnostic quality image score each represent discrete categoric data that do not follow a normal distribution. Instead, these scores represent ordinal categories each using five-point scoring scales. Therefore, this nonparametric method was used whereby a paired analysis (matched pairs) was appropriate in as much as each pig was assessed with both fluoroscopy units (CFL and GCPFL). In each abdominal girth group, each pig contributes four observations for anatomic conspicuity (ie, two anatomic conspicuity scores for CFL and two anatomic conspicuity scores for GCPFL). Also, in each abdominal girth group, each pig contributes four observations for diagnostic image quality (ie, two diagnostic image quality scores for CFL and two diagnostic image quality scores for GCPFL). Hence, this analysis is paired—that is, the same animal contributes a pair of values for CFL and GCPFL. In summary, in each abdominal girth group, the Wilcoxon signed rank test was used because each animal was assessed with two VCUG studies, by two radiologists, and with two fluoroscopy systems (CFL and GCPFL).
Analysis of the data was performed (D.Z.) by using a statistical software package (SPSS version 11.0; SPSS, Chicago, Ill). Two-tailed values of P < .05 were considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
|
|
|
|
|
There were no significant differences between the two radiologists who performed the VCUG examinations regarding total radiation exposure, fluoroscopic radiation exposure, recorded image radiation exposure, or fluoroscopy time (P > .05 for all groups).
Image Evaluation
The comparison of anatomic conspicuity scores of the recorded VCUG images between the CFL unit and the GCPFL unit (Table 2) revealed that the only significant difference was a higher median score for the GCPFL unit in group 1 for radiologist A (P = .04). There was no significant difference in the assessment of the anatomic conspicuity scores between the CFL and GCPFL units for groups 2 and 3. Examples of recorded VCUG images that received the best anatomic conspicuity score (a score of 1) are shown in Figure 2.
|
|
|
|
|
The comparison of overall diagnostic quality score (Table 3) between the recorded VCUG images from the CFL and GCPFL units showed no significant differences (P > .05 for all groups). Examples of recorded VCUG images that received the best diagnostic imaging quality score (score of 1) are shown in Figure 3.
|
|
|
|
|
Our data showed that the comparison of the anatomic conspicuity scores and the diagnostic quality scores between the CFL and GCPFL units largely resulted in nonsignificant P values. Our power analysis indicated that our sample size for each girth group was eight VCUG studies (ie, two VCUG studies per pig, per the two fluoroscopy systems, performed by two radiologists) and would provide 70% power to detect a 1.0-point difference in median anatomic conspicuity scores and diagnostic image quality scores between the fluoroscopy systems and 85% power to detect a 1.5-point difference based on median scores by using the Wilcoxon signed rank test. Therefore, with respect to detecting smaller differences in anatomic conspicuity and diagnostic image quality scores (ie, 1.0 point or less) between the two fluoroscopy systems, our sample sizes may have resulted in low power for detecting such differences. However, we believed that the sample sizes added adequate statistical power for detecting moderate differences in anatomic conspicuity or diagnostic image quality (1.5 point or greater) between the CFL and GCPFL systems. In summary, the lack of significant differences between the systems, with respect to anatomic conspicuity and diagnostic image quality scores, translates into failure to detect small differences that we believe would be considered clinically insignificant.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
Statistically significant total radiation exposure reduction with the GCPFL unit in comparison with the CFL unit occurred, because radiation exposures during fluoroscopy and during the recording of the actual VCUG images were reduced. Radiation exposure reduction with the GCPFL unit occurred during image recording, because this unit had a longer source-to-skin distance, a thicker copper filter, and modified entrance exposure settings to its image intensifier. Fluoroscopic exposure rates with the GCPFL unit were reduced because of these three factors, as well as because of the automatic brightness control system designed for children and the reduced fluoroscopic pulse rates.
We found no significant correlation between the total radiation exposure and the length of fluoroscopy time for either the GCPFL or the CFL unit. Fluoroscopy time is still thought by some radiologists to be a useful representation of radiation exposure, but the fluoroscopy time has no correlation to the number of images recorded. The fluoroscopy timer indicates the duration of time that the fluoroscopic foot pedal is depressed by the operator. This time has no correlation to the choice of fluoroscopic pulse rate of the x-ray beam (ie, 30 pulses per second versus 1.9 pulses per second). Furthermore, the fluoroscopy time has no correlation to increases in patient girth that cause the total radiation exposure to increase dramatically, as illustrated in Figure 1. Our findings underscore the fact that fluoroscopy time should not be used as an indicator of clinical radiation exposure to a child.
We were able to document diagnostic recorded image quality despite exposure reduction factors of 4.0–7.7-fold for all three groups of pig girth. Also, similar anatomic conspicuity scores were assigned to the recorded VCUG images from both fluoroscopy units, and, in fact, one radiologist even gave a higher median anatomic conspicuity score to the GCPFL unit for the group with the smallest pig girth (group 1). The implication of these image evaluation findings is that the GCPFL unit relative to the CFL unit provides comparable or somewhat improved (in smaller-sized animals) recorded images. As such, these findings warrant evaluation of this GCPFL unit in children to determine whether a similar reduction in radiation exposure can be achieved in the clinical performance of VCUG and even in other commonly performed pediatric fluoroscopic examinations.
A limitation of our study was that these pediatric-sized pigs were imaged with use of general anesthetic and all had the straightforward diagnosis of "simulated vesicoureteral reflux" (by means of intravenous administration of contrast material). Hence, these pigs remained still during imaging, and the urologic diagnosis was straightforward. These caveats likely added to the reduction in radiation exposure with both the GCPFL and CFL units. In clinical practice, a moving or uncooperative child, a child with a complicated urologic diagnosis, or a child with more than usual respiratory motion or bowel peristalsis may undergo an examination with less reduction in radiation exposure than that seen in these anesthetized pigs. However, it could be postulated that, on the basis of our study findings, the operator's ability to vary the pulse rates and pulse widths of the x-ray beam, along with other previously described modifications of the GCPFL unit, will result in even more radiation exposure reduction with the GCPFL unit than with the CFL unit.
It is possible that the fluoroscopy times were somewhat reduced by our method of simulation of vesicoureteral reflux for both the GCPFL and CFL units. Vesicoureteral reflux was simulated in the same way for all of the VCUG examinations, so if fluoroscopy time was reduced with either fluoroscopy unit, then the fluoroscopy time would be reduced similarly with both fluoroscopes and would not affect the radiation exposure results.
We believe that one of the strengths of our study was the optimization of the GCPFL unit's design for pediatric-sized pigs to reduce radiation exposure yet maintain diagnostic image quality. The GCPFL unit was treated as a system. While every exposure reduction option was considered, only those that did not sacrifice image quality were implemented. Careful consideration was given to both the fluoroscopy mode and the recorded image modes of operation. In some cases the entrance exposure to the image intensifier was increased relative to the manufacturer's recommendations. Some parameters of image quality during fluoroscopy, primarily image sharpness, were actually improved by increasing the matrix size of the displayed image and by reducing motion unsharpness by reducing the duration of each fluoroscopic pulse to freeze patient motion. These improvements in image quality "compensated" for losses in image quality owing to exposure reduction and resulted in an acceptable diagnostic image.
A second strength of our study was our electronic method for quantifying radiation exposure. PEMNET, a real-time electronic device, directly provided the total radiation exposure of each VCUG examination on the pediatric-sized pigs. Other investigators have used a dose-area-product reading that is provided on some state-of-the-art imaging equipment and have had to estimate the average collimated x-ray beam area in the patient during the examination to calculate radiation exposure (12,13,19). Without use of the PEMNET device or dose-area-product, total radiation exposure is estimated by using assumptions to determine fluoroscopy time and number of recorded images and by using a phantom model. All of these estimates, assumptions, and models introduce substantially more uncertainty into the findings. However, our use of the PEMNET device did not require these estimates and assumptions and avoided uncertainty in the calculations of total radiation exposure.
In conclusion, an optimized GCPFL unit can deliver 4.6–7.5-fold less total radiation exposure than can a CFL unit during a VCUG examination performed in a pediatric-sized pig. The overall diagnostic image quality and anatomic conspicuity of specific organs on the recorded VCUG images are comparable between the GCPFL and CFL units (and are even slightly improved with the GCPFL unit in smaller-sized animals).
Practical application: The findings of this study have allowed us to accurately quantify the amount of radiation exposure reduction associated with the performance of VCUG in pediatric-sized pigs by using a GCPFL unit compared with a CFL unit. In this pediatric porcine model of vesicoureteral reflux, an optimized GCPFL system delivered total radiation exposures that were reduced by a factor of 4.6–7.5 compared with a conventional CFL system, and diagnostic-quality recorded VCUG images were obtained. On the basis of these animal model data, GCPFL has the potential to substantially reduce the radiation exposure that children receive during the performance of common examinations such as VCUG. Furthermore, the design of the GCPFL system used in this study allows for an optimal balance between the need to reduce radiation exposure to the radiosensitive growing tissues and organs of children and the need to obtain diagnostic-quality VCUG images.
|
APPENDIX A |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
|
|
APPENDIX B |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: CFL = continuous fluoroscopy • GCPFL = grid-controlled variable-rate pulsed fluoroscopy • VCUG = voiding cystourethrography
2 Current address: Kinderchirurgische Klinik, im Dr. v. Haunerschen Kinderspital, Munich, Germany 
3 Current address: Dept of Surgery, Div of Urology, IWK Health Ctr, Halifax, Nova Scotia, Canada
See also Science to Practice in this issue
See Materials and Methods for pertinent disclosures.
Author contributions: Guarantors of integrity of entire study, V.L.W., K.J.S.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, V.L.W., K.J.S.; experimental studies, V.L.W., C.E.B., K.J.S., V.V., M.S., D.L.M., P.S.D.; statistical analysis, D.Z.; and manuscript editing, all authors
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX AAPPENDIX BReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  M. Hernanz-Schulman Science to Practice: Can Fluoroscopic Radiation Dose Be Substantially Reduced? Radiology, January 1, 2006; 238(1): 1 - 2. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||
