Pediatr Radiol (2014) 44:313–321 DOI 10.1007/s00247-013-2824-9
ORIGINAL ARTICLE
Guidelines for anti-scatter grid use in pediatric digital radiography Shannon Fritz & A. Kyle Jones
Received: 10 May 2013 / Revised: 12 October 2013 / Accepted: 15 October 2013 / Published online: 27 November 2013 # Springer-Verlag Berlin Heidelberg 2013
Abstract Background Pediatric radiography presents unique challenges in balancing image quality and patient dose. Removing the anti-scatter grid reduces patient dose but also reduces image contrast. The benefit of using an anti-scatter grid decreases with decreasing patient size. Objective To determine patient thickness thresholds for antiscatter grid use by comparing scatter-to-primary ratio for progressively thinner patients without a grid to the scatter-toprimary ratio for a standard adult patient with a grid. Materials and methods We used Solid Water™ phantoms ranging in thickness from 7 cm to 16 cm to simulate pediatric abdomens. The scatter-to-primary ratio without a grid was measured for each thickness at 60 kVp, 70 kVp and 80 kVp for X-ray fields of view (FOV) of 378 cm2, 690 cm2 and 1,175 cm2 using indirect digital radiography (iDR) and computed radiography (CR). We determined thresholds for anti-scatter grid use by comparing the intersection of a fit of scatter-to-primary ratio versus patient thickness with a standard adult scatter-to-primary ratio measured for a 23-cm phantom thickness at 80 kVp with an anti-scatter grid. Dose area product (DAP) was also calculated. Results The scatter-to-primary ratio depended strongly on FOV and weakly on kVp; however DAP increased with decreasing kVp. Threshold thicknesses for grid use varied from 5 cm for a 14×17-cm FOV using iDR to 12 cm for an 8×10-cm FOV using computed radiography. Conclusions Removing the anti-scatter grid for small patients reduces patient dose without a substantial increase in scatterto-primary ratio when the FOV is restricted appropriately. S. Fritz (*) : A. K. Jones Department of Imaging Physics, The University of Texas, M. D. Anderson Cancer Center, 1400 Pressler St., Houston, TX 77030, USA e-mail: [email protected]
Radiologic technologists should base anti-scatter grid use on patient thickness and FOV rather than age. Keywords Anti-scattergrid . Scatter-to-primary ratio . Digital radiography . Abdominal radiography . Pediatric radiography . Radiation dose
Introduction Creating appropriate guidelines for pediatric radiography can be challenging. Each aspect of the imaging process must be optimized to create a diagnostic image of sufficient quality while maintaining patient dose at an acceptable level. Techniques used to improve image quality can result in increased patient dose. The use of an anti-scatter grid is one such technique. The anti-scatter grid improves image contrast by preferentially removing scattered X-rays from the X-ray beam before it reaches the image receptor. However the anti-scatter grid also removes some primary X-rays from the beam, and the mAs must be increased to maintain a constant image noise level when an anti-scatter grid is used in digital radiography. Existing guidelines offer different, and at times conflicting, advice for the use of anti-scatter grids in pediatric radiography. These include recommendations for the use of an anti-scatter grid for body parts thicker than 10 cm or when using tube potentials greater than 60 kVp [1]; for body parts greater than or equal to 15 cm in thickness [2]; for patients older than 6 months [3]; for patients other than infants or toddlers [4]; and a recommendation that anti-scatter grids are generally unnecessary in pediatric radiography [5]. One experimental study evaluated the impact of anti-scatter grid use in pediatric chest radiography using a cesium iodide (CsI) indirect flat panel detector. The authors used polymethyl methacrylate to represent a pediatric chest and compared contrast-detail
314
Pediatr Radiol (2014) 44:313–321
scores, determining that contrast-detail scores without an antiscatter grid were equivalent to those with an anti-scatter grid for pediatric chests equivalent to 5 cm or less of polymethyl methacrylate [6]. The improvement in contrast achieved when using an antiscatter grid depends strongly on the scatter-to-primary ratio, as demonstrated in the following expression for the contrast improvement ratio, K: K¼
1 þ scatter−to primary ratio 1 þ scatter−to primary ratioGrid
ð1Þ
The purpose of this study was to determine the patient thickness threshold at which the scatter-to-primary ratio in pediatric imaging without an anti-scatter grid was equal to the scatter-to-primary ratio measured for a standard adult using an anti-scatter grid for anteroposterior (AP) abdominal radiography. A standard adult was defined as a 23-cm-thick patient using a 35.6×43.2-cm (14×17-in.) field of view.
Materials and methods Set-up
where the primary transmission scatter-to-primary ratio and scatter-to-primary ratioGrid are the scatter-to-primary ratio without and with an anti-scatter grid [7]. A typical scatter-toprimary ratio for an adult abdomen using an anti-scatter grid is 0.8–1.0; without an anti-scatter grid the scatter-to-primary ratio is 3.0–4.0. This corresponds to a typical value of K of approximately 2.5. The use of an anti-scatter grid offers little improvement in contrast for radiographic examinations characterized by a low scatter-to-primary ratio. K approaches 1.0 (no contrast improvement) when scatter-to-primary ratio and scatter-to-primary ratioGrid are similar. There is a tradeoff involved when an anti-scatter grid is used, namely an increase in incident air kerma to the patient. This is necessary because the grid not only attenuates scattered X-rays but also a fraction of the primary X-ray beam. Although there is generally no agreement on by what factor the mAs must be increased when a grid is used in digital radiography, the lower limit on an increase is the reciprocal of the primary transmission (Tp) of the grid, the upper limit is the Bucky factor. The reciprocal of Tp (approximately 1.2–1.3) is often used in indirect digital radiography and computed radiography, while the Bucky factor (approximately 4.0–5.0) was used for screen-film imaging because of the need for consistent darkening of the film, to which both scattered and primary radiation contribute. Resources are available for estimating the necessary increase in technique [1]. Radiographic examinations characterized by low scatterto-primary ratio include extremities for patients of all sizes and all types of views for small patients, including many pediatric patients. Pediatric radiography is of particular importance because it is well known that pediatric patients are more radiosensitive than adults [8] and therefore it is especially important to keep the dose as low as practicable without compromising the diagnostic usefulness of the image [9]. Optimization of grid use in thoracic and abdominal pediatric imaging is of particular interest for two reasons. The first is that many radiosensitive organs are located within the thoracic and abdominal cavities. The second is that the thickness of the chest and abdomen vary widely across the pediatric population, meaning that the use of an anti-scatter grid might be indicated for some pediatric patients but not others.
The experimental set-up used throughout this study is illustrated in Fig. 1. The scatter-to-primary ratio was measured for five phantom thicknesses: 7 cm, 10 cm, 13 cm, 16 cm and 23 cm. The 23-cm thickness was chosen to represent a standard adult based on 50th percentile AP abdominal thickness given by several sources including the anthropometry software PeopleSize 2008 (Open Ergonomics, Leicestershire, UK), the LucAl abdomen phantom (developed by the Center for Devices and Radiological Health [10], and the American College of Radiology (ACR) Radiography/Fluoroscopy Accreditation abdomen phantom (equivalent to 23 cm of water) [11]). In this experiment abdominal thicknesses were simulated using Solid Water™ slabs (Gammex; RMI, Middleton, WI). An anti-scatter grid was used only for the 23-cm standard adult phantom. The scatter-to-primary ratio was measured using the graduated beam stop method [12–14]. Lead disks of 5-mm thickness and diameters of 2.5 mm, 5 mm, 7.5 mm, 10 mm, 15 mm, 17.5 mm and 20 mm were embedded in a polymethyl methacrylate sheet (Fig. 1). These disks attenuated all of the primary radiation. The beam stop array was placed on a 2.5-cm foam standoff at a distance of 71 cm from the X-ray source so the projected sizes of the disks in the image would not change with phantom thickness. The ACR recommends using 80 kVp for an average adult abdomen digital radiography exam [15] and 60–75 kVp for pediatric abdominal radiography [16], and European guidelines for pediatric abdominal radiography suggest 65– 85 kVp [3]. We chose to investigate 60 kVp, 70 kVp and 80 kVp, the last in order to compare to our standard adult technique with an anti-scatter grid. X-ray fields of view for simulating pediatric abdomen radiographs were selected from the Handbook of Selected Organ Doses for Projections Common in Pediatric Radiology [17] for AP abdomen exams. These fields of view corresponded to a 5-year-old (30×23 cm, 690 cm2) and a 1-year-old (21×18 cm, 378 cm2), referred to hereafter as “medium” and “small,” respectively. The areal dimensions of the Solid Water™ slabs available to us placed a limit on the maximum X-ray FOV that could be used to intercept the entire phantom. A maximum FOV of 35.6× 33 cm (14×13 in.) was used for the adult FOV for the indirect
Pediatr Radiol (2014) 44:313–321
315
Fig. 1 Schematic of experimental setup. Indirect digital radiography and computed radiography images were acquired with a source-to-image distance of 100 cm and 110 cm, respectively. a The beam stops were positioned on the foam standoff to maintain the same magnification factor as phantom thickness was varied. b Schematic of the beam stop array used for measuring scatter-to-primary ratio
digital radiography study using a 100-cm source-to-image distance. This large FOV (35.6×33 cm, 1,175 cm2) was therefore only 76% of the size of the standard FOV for an adult AP abdomen radiograph (35.6×43.2 cm, 1,538 cm2) as indicated in the Handbook of Selected Tissue Doses for Projections Common in Diagnostic Radiology [18]. The increase in scatter-to-primary ratio from the large FOV to the standard 35.6×43.2-cm FOV was extrapolated from a fit to the data acquired at 80 kVp for both indirect digital and computed radiography and was used to calculate scatter-toprimary ratio for the standard adult technique. The FOV at the surface of the phantom for computed radiography images was matched to the FOV at the surface of the phantom for the indirect digital radiography images. The resulting fields of view at the image receptor corresponding to large, medium and small were 39.1×36.3 cm, 33.0×25.4 cm and 23.2× 19.8 cm, respectively.
Indirect digital radiography Half of the experiments were performed using a CsI indirect flat panel digital radiography system (Definium 8000; GE Healthcare, Waukesha, WI). The half-value layer of the indirect digital radiography system was 3.0 mm aluminum at 80 kVp. We used a general-purpose 12:1 focused anti-scatter grid (Mitaya Manufacturing Co., Tokyo, Japan). The listed grid performance characteristics were 70 lines per centimeter, 12:1 grid ratio, 100-cm focus distance, strip thickness of 40 μm, lead strip material, aluminum interspace material, and carbon fiber reinforced polymer cover material. A source-to-image distance of 100 cm was used to acquire all images on the indirect digital radiography system. All regionof-interest measurements were made on the “for-processing” images (images to which only detector corrections were applied, and not image processing). The pixel values in the for-processing images from the indirect digital radiography system had a linear relationship to image receptor dose.
Computed radiography The other half of the experiments was performed on a computed radiography system using barium fluorobromide (BaFBr) screens and an FCR 5000 digitizer (Fujifilm USA, Stamford, CT). All computed radiography images were acquired using a Philips OmniDiagnost Eleva radiography/ fluoroscopy system (Philips Healthcare, Andover, MA) and the same 35.6×43.2 cm (14×17-in.) imaging plate. The halfvalue layer of the OmniDiagnost was 3.7 mm aluminum at 80 kVp. We used a general-purpose 13:1 focused anti-scatter grid (Philips Healthcare, Andover, MA). The listed grid performance characteristics were 60 lines per centimeter, 13:1 grid ratio, 120-cm focus distance, strip thickness of 36 μm, lead strip material, fiber interspace material, and carbon fiber cover. A source-to-image distance of 110 cm was used to acquire all computed radiography images. An air gap of 13 cm existed between the exit surface of the phantom and the image receptor on the OmniDiagnost system. This is compared to an air gap of 3 cm for the X-ray system used in the indirect digital radiography study, more typical for abdominal radiography. The larger air gap was a consequence of both the curvature of the patient support, to which the rigid Solid Water×™ slabs did not conform, and the geometry of the X-ray system. The plate was digitized using the Fuji Ave 4.0 test-image processing algorithm and transferred as a 10-bit image. The pixel values in the resulting images had a logarithmic relationship to image receptor dose. We decalibrated the images by measuring the characteristic function of the computed radiography system and using the inverse of the resulting function to transform images to linear image receptor dose space. All region-of-interest measurements were made on the de-calibrated images. Analysis Automatic exposure control was used to acquire an initial image for each combination of X-ray field of view, phantom
316
Pediatr Radiol (2014) 44:313–321
thickness and kVp for both the indirect digital radiography and computed radiography experiments. The fixed mAs value closest to the resulting automatic exposure control mAs value was used to acquire a series of six images for each kVp, FOV and phantom thickness combination—three without the beam stops in place and three with the beam stops in place. Images were analyzed using ImageJ v. 1.45 s (National Institutes of Health, Bethesda, MD). The signal from scatter, SS, was measured in regions of interest behind the disks in images with the beam stops in place and the total signal, ST, was measured using identical regions of interest in images without the beam stops in place. The regions of interest were half the diameter of the disks. SS and ST were then used to calculate the primary signal, SP = ST − SS. The scatter-to-primary ratio was calculated using the expression scatter − to − primary ratio ¼
SS SS ¼ S P S T −S S
ð2Þ
Because the scatter-to-primary ratio varies with beam stop diameter [19], it was necessary to determine the disk-free scatter-to-primary ratio, scatter-to-primary ratio0. This was accomplished by applying a linear fit to a plot of ln(scatterto-primary ratio) versus beam stop diameter. The intercept of this fit was equal to ln(scatter-to-primary ratio0). Exposure was measured for each automatic exposure controlled exposure using an ionization chamber (Triad Model 35050A Dosimeter, 15 cm3 ion chamber 96035B, Fluke Biomedical, Everett, WA). Exposure was converted to incident air kerma: Incident air kerma ðmGyÞ ¼ Exposure ðRÞ 8:76; and the dose area product (DAP) was calculated as the product of the incident air kerma and the FOV.
Figure 2 shows an example of the results for scatter-toprimary ratio0 versus phantom thickness for different X-ray fields of view at 80 kVp for indirect digital radiography. The standard errors of the fit intercepts used to calculate scatter-toprimary ratio0 were on the order of 10−2 (coefficient of variation less than 0.005). Systematic errors in our experiments were an order of magnitude higher than these experimental errors; these are examined in the Discussion section. The dashed line indicates scatter-to-primary ratio0 for the standard adult technique with the 23-cm phantom at 80 kVp with an anti-scatter grid and an FOVof 35.6×43.2 cm (14×17 in.). The intersection of this dashed line with the fits of scatter-to-primary ratio 0 versus phantom thickness indicates the threshold thickness for use of an anti-scatter grid for a given combination of X-ray FOV and kVp. The threshold thicknesses tended to be slightly lower for computed radiography compared to indirect digital radiography, a consequence of the different effective atomic numbers of the phosphors used in CR and iDR. Figure 3 plots threshold thicknesses versus FOV for 60 kVp, 70 kVp and 80 kVp and fits these data with power functions. Threshold thicknesses for commonly used clinical X-ray fields of view are listed in Table 1. The fit parameters, which can be used by medical physicists to calculate the threshold thickness for any FOV when constructing technique charts for pediatric abdominal radiography, are provided in Table 2. The dose area product resulting from the different experimental configurations is plotted in Figs. 4 and 5. Both figures plot the DAP calculated from the product of the measured exposure and the X-ray FOV at each kVp for all phantom thicknesses. Dose area product increased steeply as kVp decreased when using automatic exposure control.
Results The 20-mm disk was omitted in all scatter-to-primary ratio analyses because it fell partly outside the X-ray FOV for the smallest FOV evaluated. The scatter-to-primary ratio was measured using an anti-scatter grid for our standard adult and extrapolated to the full 35.6×43.2 cm (14×17 in.). Scatter-to-primary ratio was 0.78 for indirect digital radiography and 1.1 for computed radiography. The scatterto-primary ratio was higher for CR compared to iDR, a consequence of the fact that BaFBr, the most commonly used powdered CR phosphor, has a lower effective atomic number than CsI, the phosphor used in indirect DR systems. Phosphors with lower effective atomic numbers absorb scattered X-rays more efficiently than do phosphors with higher effective atomic numbers.
Fig. 2 Disk-free scatter-to-primary ratio, called scatter-to-primary ratio0, versus phantom thickness for different X-ray fields of view at 80 kVp using the indirect digital radiography system. The horizontal dashed line indicates scatter-to-primary ratio0 for the standard adult technique, the only technique for which an anti-scatter grid was used. Drop lines indicate threshold thicknesses for anti-scatter grid use
Pediatr Radiol (2014) 44:313–321
317
Fig. 3 Patient anteroposterior (AP) thickness thresholds for antiscatter grid use plotted versus field of view for each kVp for (a) indirect digital radiography (iDR) and (b) computed radiography (CR)
Discussion The results of this study indicate that for a given patient thickness the scatter-to-primary ratio is strongly dependent on X-ray FOV and weakly dependent on kVp. This is consistent with previous investigations [20]. These results reinforce the importance of tight collimation of the X-ray field Table 1 Threshold patient AP thicknesses (cm) for anti-scatter grid use for common clinical fields of view (inches) iDR
60 kVp 70 kVp 80 kVp
CR
to the anatomy of interest. Collimation has the added benefit of reducing DAP and therefore patient dose. The decision to use an anti-scatter grid should be based on the smaller of the X-ray FOV or patient dimensions. For example our results indicated that when using our large X-ray field of view, the patient thickness threshold for grid use was 6 cm or 7 cm. However, consider the case of an Table 2 Fit parameters from Fig. 3. These parameters can be used to construct technique charts for pediatric abdominal radiography. The fitted equation is y = ax b , where y = AP thickness threshold (cm) and x = FOV (cm2)
14×17a
10×12
8×10
14×17
10×12
8×10
iDR
5 5 5
9 9 8
12 11 11
6 5 6
9 9 9
13 12 12
a
b
1,250 1,024 1,020
−0.746 −0.720 −0.727
a
14×17 in.=35.6×43.2 cm; 10×12 in.=25.4×30.5 cm; 8×10 in.=20.3× 25.4 cm
AP anteroposterior, iDR indirect digital radiography, CR computed radiography
60 kVp 70 kVp 80 kVp
CR a 1,326 1,392 820
b −0.745 −0.760 −0.679
AP anteroposterior, iDR indirect digital radiography, CR computed radiography
318
Pediatr Radiol (2014) 44:313–321
Fig. 4 Dose area product (DAP) as a function of kVp for indirect digital radiography (iDR) for the (a) large, (b) medium and (c) small X-ray field of view (FOV)
uncooperative child, where the collimation must be broad so that the anatomy of interest is not missed. Because the patient is the primary source of scatter, the scatter-to-primary ratio is not appreciably increased if the collimated X-ray FOVextends beyond the patient anatomy. There was a slight increase in the threshold thickness for anti-scatter grid use with decreasing kVp (Fig. 3) (Table 1); however, there was a much steeper associated dose penalty (Figs. 4 and 5). For a constant receptor dose, dose area product increases with decreasing kVp when using automatic exposure control, especially for thicker patients. Because scatter-to-primary ratio is weakly dependent on kVp it is important to consider the cost of reducing kVp to reduce scatter. The AP threshold thicknesses and transverse X-ray field dimensions in this study correspond to children across a wide age range. A study by Kleinman et al. [21] in 2010 demonstrated the large degree of overlap in child sizes from toddlers to teens. For example, in the current study the transverse dimension of the small FOV in the patient plane
was approximately 16 cm. According to the Kleinman [21] study, this corresponds to the mean transverse dimension of a 1-year-old child, but children up to age 10 may share this transverse dimension. In the current study, the threshold thickness at 80 kVp for the small FOV for indirect digital radiography was 14 cm. This corresponds to the mean AP thickness of a 4-year-old child but also includes children up to 14 years old who are smaller than average. These large variations mean that it is difficult to base technique factor selection, including use of anti-scatter grids, on patient age. Instead, the patient should be measured along the appropriate dimensions before a radiographic exam and technical factors should be selected based on this information. This study had several limitations. First, our patient phantom dimensions were limited by the areal dimensions of the Solid Water™ slabs available to us. In many cases the extent of the phantom, particularly in the transverse dimension, exceeded the dimensions of actual patients. Therefore it is likely that some X-rays that scattered out of the field of view and normally would not have reached the detector scattered
Pediatr Radiol (2014) 44:313–321
319
Fig. 5 Dose area product (DAP) as a function of kVp for computed radiography (CR) for the (a) large, (b) medium and (c) small X-ray field of view (FOV)
back into the FOV and contributed to Ss. This would increase the scatter-to-primary ratio slightly compared to phantoms with areal dimensions closer to those of actual patients. In the adult case, our maximum X-ray FOV was limited by the phantom dimensions, which required us to extrapolate to find the scatterto-primary ratio0 for the full adult FOV. However considering the gradual, predictable increase in scatter-to-primary ratio at large FOV [22], the error introduced was small. Second, there was a slight difference between the source-to-image distance used and the grid focus distance for the computed radiography system used in this study. Based on the magnitude of this difference, the geometry and dimensions of our beam stop array, and the ratio of the grid [22], we have estimated that the signal from primary radiation was underestimated by 3– 5%, depending on the exact position of the beam stops. This would result in an overestimation of scatter-to-primary ratio0 by a similar factor for our standard adult technique. In the end this would result in overestimation of threshold thicknesses by approximately 1 cm in some cases. Third, an air gap of 13 cm existed between the exit surface of the phantom and the image
receptor on the OmniDiagnost system used to acquire computed radiography images. The presence of this air gap would be expected to result in a slight decrease in the scatterto-primary ratio for each of our measurements, with larger decreases for thicker phantoms [23]. The impact of the air gap was expected to be an underestimation of threshold thicknesses by approximately 5%. Although we cannot be sure what the sum of these two effects was, it is unlikely that the threshold thicknesses presented here are in error by more than +/− 5%. Finally, the characteristics of the antiscatter grids used for the indirect digital radiography and computed radiography experiments were slightly different. These results have not been validated in clinical practice, and in light of the study limitations presented here, our recommendations should be validated prior to clinical use. Another important consideration is that pediatric radiologists are often willing to tolerate images of marginal diagnostic quality in an effort to minimize patient dose. Therefore, our choice to benchmark scatter-to-primary ratio using a standard adult technique may have resulted in a scatter-to-primary ratio
320
that was lower than the value of scatter-to-primary ratio that can be tolerated in pediatric radiography. However, our results can be adapted to any choice of baseline scatter-to-primary ratio.
Conclusion It is important that X-ray technologists be educated about the use of anti-scatter grids, including the costs and benefits of removing the grid. Technique charts should include clear guidelines for removing the anti-scatter grid when imaging patients of specific dimensions. Technologists should know which X-ray systems in their department have anti-scatter grids that can be removed and be instructed on how to remove the grid, which may require the selection of specific imaging presets at the control console. There is currently a recommendation from the Image Gently campaign that manufacturers develop technology to allow automatic removal of the anti-scatter grid, as opposed to manual removal, to aid the technologist and improve workflow [24]. When appropriate, pediatric patients should either be protocoled to rooms that have removable anti-scatter grids or undergo tabletop imaging. The data presented in this study can be used to create technique charts for pediatric abdominal radiography. Technique charts should emphasize the importance of measuring patient dimensions, either using a prior cross-sectional imaging exam or physical calipers, rather than relying on patient age to determine proper technique. The medical physicist should work closely with the radiologist to determine what image quality is required for specific procedures and if removing the anti-scatter grid is appropriate given the required image quality and potential impact on patient dose [9]. In many indications for abdominal radiography lower contrast is acceptable, for example high subject contrast tasks such as visualizing air-fluid levels. Thus, the anti-scatter grid might be removed for a wider range of pediatric sizes than indicated by our data. It would be beneficial to repeat this study for other common pediatric radiographic examinations where no consensus guidelines exist for anti-scatter grid use, for example chest radiography and fluoroscopy, and to repeat these experiments using other image receptors such as amorphous selenium. Conflicts of interest None.
References 1. Carlton RR, Adler AM (2012) The grid. In: Principles of radiographic imaging: an art and a science, 5th edn. Delmar Cengage Learning, Clifton Park, pp 257–272
Pediatr Radiol (2014) 44:313–321 2. Alzen G, Benz-Bohm G (2011) Radiation protection in pediatric radiology. Dtsch Arztebl Int 108:407–414 3. European Commission (1996) European guidelines on quality criteria for diagnostic images in paediatrics. ftp://ftp.cordis. lu/pub/fp5-euratom/docs/eur16260.pdf. Accessed May 1, 2013 4. American College of Radiology (2008) ACR-SPR practice guidelines for general radiography. http://www.acr.org/~/ media/95AB63131AD64CC6BB54BFE008DBA30C.pdf. Accessed May 1, 2013 5. International Atomic Energy Agency (2013) Safety report series No. 71, radiation protection in paediatric radiology. http://www-pub.iaea. org/books/iaeabooks/8727/Radiation-Protection-in-PaediatricRadiology. Accessed May 1, 2013 6. Ween B, Olstad M, Jakobsen JA et al (2009) Pediatric digital chest radiography, comparison of grid versus non-grid techniques. Eur J Radiography 1:201–206 7. IEC International Standard 60627 (2001) Diagnostic X-ray imaging equipment—characteristics of general purpose and mammographic antiscatter grids. International Electrochemical Commission, ISBN 28318-5953-0 8. Hall EJ (2002) Lessons we have learned from our children: cancer risks from diagnostic radiology. Pediatr Radiol 32: 700–706 9. Amis ES Jr, Butler PF, Applegate KE et al (2007) American College of Radiology white paper on radiation dose in medicine. J Am Coll Radiol 4:272–284 10. Conway BJ, Duff JE, Fewell TR et al (1990) A patient-equivalent attenuation phantom for estimating patient exposures from automatic exposure controlled X-ray examinations of the abdomen and lumbosacral spine. Med Phys 17:448–453 11. American College of Radiology (2008) ACR Practice Guideline for Diagnostic Reference Levels in Medical X-ray Imaging. http://www. acr.org/~/media/796DE35AA407447DB81CEB5612B4553D.pdf. Accessed May 1, 2013 12. Yaffe M, Fenster A, Johns HE (1977) Xenon ionization detectors for fan beam computed tomography scanners. J Comput Assist Tomogr 1:419–428 13. Floyd CE Jr, Lo JY, Chotas HG et al (1991) Quantitative scatter measurement in digital radiography using a photostimulable phosphor imaging system. Med Phys 18:408–413 14. Fetterly KA, Schueler BA (2007) Experimental evaluation of fiber-interspaced antiscatter grids for large patient imaging with digital x-ray systems. Phys Med Biol 52: 4863–4880 15. American College of Radiology (2012) ACR-AAPM-SIIM practice guideline for digital radiography. http://www.acr.org/~/media/ 3E08C87AD6E6498D9E19769E5E5E390D.pdf. Accessed May 1, 2013 16. American College of Radiology (2011) ACR-SPR guideline for the performance of abdominal radiography. http://www.acr.org/~/media/ 79A594819BBD4631A7E31404DAA66EF6.pdf. Accessed May 1, 2013 17. Rosenstein M, Beck TJ, Warner GG (1979) Handbook of selected organ doses for projections common in pediatric radiology. HEW Publication FDA pp. 79–8079 18. Rosenstein M (1988) Handbook of selected tissue doses for projections common in diagnostic radiology. HEW Publication FDA pp. 89–8031 19. Maher KP (1993) Comparison of scatter measurement techniques in digital fluoroscopy. Phys Med Biol 38:1977–1983 20. Yaffe MJ, Johns PC (1983) Scattered radiation in diagnostic radiology—magnitudes, effects, and methods of reduction. J Appl Photogr Eng 9:184–195
Pediatr Radiol (2014) 44:313–321 21. Kleinman PL, Strauss KJ, Zurakowski D et al (2010) Patient size measured on CT images as a function of age at a tertiary care children’s hospital. AJR Am J Roentgenol 194:1611–1619 22. Aichinger H, Dierker J, Joite-Barfuß S et al (2012) Scattered radiation. In: Radiation exposure and image quality in X-Ray diagnostic radiology, 2nd edn. Springer, Berlin, pp 53–66
321 23. Curry TS III, Dowdey JE, Murry RC Jr (1990) Christensen’s physics of diagnostic radiology, 4th edn. Lippincott, Williams, and Wilkins, Philadelphia 24. Goske MJ, Charkot E, Herrmann T et al (2011) Image Gently: challenges for radiologic technologists when performing digital radiography in children. Pediatr Radiol 41:611–619
ORIGINAL ARTICLE
Guidelines for anti-scatter grid use in pediatric digital radiography Shannon Fritz & A. Kyle Jones
Received: 10 May 2013 / Revised: 12 October 2013 / Accepted: 15 October 2013 / Published online: 27 November 2013 # Springer-Verlag Berlin Heidelberg 2013
Abstract Background Pediatric radiography presents unique challenges in balancing image quality and patient dose. Removing the anti-scatter grid reduces patient dose but also reduces image contrast. The benefit of using an anti-scatter grid decreases with decreasing patient size. Objective To determine patient thickness thresholds for antiscatter grid use by comparing scatter-to-primary ratio for progressively thinner patients without a grid to the scatter-toprimary ratio for a standard adult patient with a grid. Materials and methods We used Solid Water™ phantoms ranging in thickness from 7 cm to 16 cm to simulate pediatric abdomens. The scatter-to-primary ratio without a grid was measured for each thickness at 60 kVp, 70 kVp and 80 kVp for X-ray fields of view (FOV) of 378 cm2, 690 cm2 and 1,175 cm2 using indirect digital radiography (iDR) and computed radiography (CR). We determined thresholds for anti-scatter grid use by comparing the intersection of a fit of scatter-to-primary ratio versus patient thickness with a standard adult scatter-to-primary ratio measured for a 23-cm phantom thickness at 80 kVp with an anti-scatter grid. Dose area product (DAP) was also calculated. Results The scatter-to-primary ratio depended strongly on FOV and weakly on kVp; however DAP increased with decreasing kVp. Threshold thicknesses for grid use varied from 5 cm for a 14×17-cm FOV using iDR to 12 cm for an 8×10-cm FOV using computed radiography. Conclusions Removing the anti-scatter grid for small patients reduces patient dose without a substantial increase in scatterto-primary ratio when the FOV is restricted appropriately. S. Fritz (*) : A. K. Jones Department of Imaging Physics, The University of Texas, M. D. Anderson Cancer Center, 1400 Pressler St., Houston, TX 77030, USA e-mail: [email protected]
Radiologic technologists should base anti-scatter grid use on patient thickness and FOV rather than age. Keywords Anti-scattergrid . Scatter-to-primary ratio . Digital radiography . Abdominal radiography . Pediatric radiography . Radiation dose
Introduction Creating appropriate guidelines for pediatric radiography can be challenging. Each aspect of the imaging process must be optimized to create a diagnostic image of sufficient quality while maintaining patient dose at an acceptable level. Techniques used to improve image quality can result in increased patient dose. The use of an anti-scatter grid is one such technique. The anti-scatter grid improves image contrast by preferentially removing scattered X-rays from the X-ray beam before it reaches the image receptor. However the anti-scatter grid also removes some primary X-rays from the beam, and the mAs must be increased to maintain a constant image noise level when an anti-scatter grid is used in digital radiography. Existing guidelines offer different, and at times conflicting, advice for the use of anti-scatter grids in pediatric radiography. These include recommendations for the use of an anti-scatter grid for body parts thicker than 10 cm or when using tube potentials greater than 60 kVp [1]; for body parts greater than or equal to 15 cm in thickness [2]; for patients older than 6 months [3]; for patients other than infants or toddlers [4]; and a recommendation that anti-scatter grids are generally unnecessary in pediatric radiography [5]. One experimental study evaluated the impact of anti-scatter grid use in pediatric chest radiography using a cesium iodide (CsI) indirect flat panel detector. The authors used polymethyl methacrylate to represent a pediatric chest and compared contrast-detail
314
Pediatr Radiol (2014) 44:313–321
scores, determining that contrast-detail scores without an antiscatter grid were equivalent to those with an anti-scatter grid for pediatric chests equivalent to 5 cm or less of polymethyl methacrylate [6]. The improvement in contrast achieved when using an antiscatter grid depends strongly on the scatter-to-primary ratio, as demonstrated in the following expression for the contrast improvement ratio, K: K¼
1 þ scatter−to primary ratio 1 þ scatter−to primary ratioGrid
ð1Þ
The purpose of this study was to determine the patient thickness threshold at which the scatter-to-primary ratio in pediatric imaging without an anti-scatter grid was equal to the scatter-to-primary ratio measured for a standard adult using an anti-scatter grid for anteroposterior (AP) abdominal radiography. A standard adult was defined as a 23-cm-thick patient using a 35.6×43.2-cm (14×17-in.) field of view.
Materials and methods Set-up
where the primary transmission scatter-to-primary ratio and scatter-to-primary ratioGrid are the scatter-to-primary ratio without and with an anti-scatter grid [7]. A typical scatter-toprimary ratio for an adult abdomen using an anti-scatter grid is 0.8–1.0; without an anti-scatter grid the scatter-to-primary ratio is 3.0–4.0. This corresponds to a typical value of K of approximately 2.5. The use of an anti-scatter grid offers little improvement in contrast for radiographic examinations characterized by a low scatter-to-primary ratio. K approaches 1.0 (no contrast improvement) when scatter-to-primary ratio and scatter-to-primary ratioGrid are similar. There is a tradeoff involved when an anti-scatter grid is used, namely an increase in incident air kerma to the patient. This is necessary because the grid not only attenuates scattered X-rays but also a fraction of the primary X-ray beam. Although there is generally no agreement on by what factor the mAs must be increased when a grid is used in digital radiography, the lower limit on an increase is the reciprocal of the primary transmission (Tp) of the grid, the upper limit is the Bucky factor. The reciprocal of Tp (approximately 1.2–1.3) is often used in indirect digital radiography and computed radiography, while the Bucky factor (approximately 4.0–5.0) was used for screen-film imaging because of the need for consistent darkening of the film, to which both scattered and primary radiation contribute. Resources are available for estimating the necessary increase in technique [1]. Radiographic examinations characterized by low scatterto-primary ratio include extremities for patients of all sizes and all types of views for small patients, including many pediatric patients. Pediatric radiography is of particular importance because it is well known that pediatric patients are more radiosensitive than adults [8] and therefore it is especially important to keep the dose as low as practicable without compromising the diagnostic usefulness of the image [9]. Optimization of grid use in thoracic and abdominal pediatric imaging is of particular interest for two reasons. The first is that many radiosensitive organs are located within the thoracic and abdominal cavities. The second is that the thickness of the chest and abdomen vary widely across the pediatric population, meaning that the use of an anti-scatter grid might be indicated for some pediatric patients but not others.
The experimental set-up used throughout this study is illustrated in Fig. 1. The scatter-to-primary ratio was measured for five phantom thicknesses: 7 cm, 10 cm, 13 cm, 16 cm and 23 cm. The 23-cm thickness was chosen to represent a standard adult based on 50th percentile AP abdominal thickness given by several sources including the anthropometry software PeopleSize 2008 (Open Ergonomics, Leicestershire, UK), the LucAl abdomen phantom (developed by the Center for Devices and Radiological Health [10], and the American College of Radiology (ACR) Radiography/Fluoroscopy Accreditation abdomen phantom (equivalent to 23 cm of water) [11]). In this experiment abdominal thicknesses were simulated using Solid Water™ slabs (Gammex; RMI, Middleton, WI). An anti-scatter grid was used only for the 23-cm standard adult phantom. The scatter-to-primary ratio was measured using the graduated beam stop method [12–14]. Lead disks of 5-mm thickness and diameters of 2.5 mm, 5 mm, 7.5 mm, 10 mm, 15 mm, 17.5 mm and 20 mm were embedded in a polymethyl methacrylate sheet (Fig. 1). These disks attenuated all of the primary radiation. The beam stop array was placed on a 2.5-cm foam standoff at a distance of 71 cm from the X-ray source so the projected sizes of the disks in the image would not change with phantom thickness. The ACR recommends using 80 kVp for an average adult abdomen digital radiography exam [15] and 60–75 kVp for pediatric abdominal radiography [16], and European guidelines for pediatric abdominal radiography suggest 65– 85 kVp [3]. We chose to investigate 60 kVp, 70 kVp and 80 kVp, the last in order to compare to our standard adult technique with an anti-scatter grid. X-ray fields of view for simulating pediatric abdomen radiographs were selected from the Handbook of Selected Organ Doses for Projections Common in Pediatric Radiology [17] for AP abdomen exams. These fields of view corresponded to a 5-year-old (30×23 cm, 690 cm2) and a 1-year-old (21×18 cm, 378 cm2), referred to hereafter as “medium” and “small,” respectively. The areal dimensions of the Solid Water™ slabs available to us placed a limit on the maximum X-ray FOV that could be used to intercept the entire phantom. A maximum FOV of 35.6× 33 cm (14×13 in.) was used for the adult FOV for the indirect
Pediatr Radiol (2014) 44:313–321
315
Fig. 1 Schematic of experimental setup. Indirect digital radiography and computed radiography images were acquired with a source-to-image distance of 100 cm and 110 cm, respectively. a The beam stops were positioned on the foam standoff to maintain the same magnification factor as phantom thickness was varied. b Schematic of the beam stop array used for measuring scatter-to-primary ratio
digital radiography study using a 100-cm source-to-image distance. This large FOV (35.6×33 cm, 1,175 cm2) was therefore only 76% of the size of the standard FOV for an adult AP abdomen radiograph (35.6×43.2 cm, 1,538 cm2) as indicated in the Handbook of Selected Tissue Doses for Projections Common in Diagnostic Radiology [18]. The increase in scatter-to-primary ratio from the large FOV to the standard 35.6×43.2-cm FOV was extrapolated from a fit to the data acquired at 80 kVp for both indirect digital and computed radiography and was used to calculate scatter-toprimary ratio for the standard adult technique. The FOV at the surface of the phantom for computed radiography images was matched to the FOV at the surface of the phantom for the indirect digital radiography images. The resulting fields of view at the image receptor corresponding to large, medium and small were 39.1×36.3 cm, 33.0×25.4 cm and 23.2× 19.8 cm, respectively.
Indirect digital radiography Half of the experiments were performed using a CsI indirect flat panel digital radiography system (Definium 8000; GE Healthcare, Waukesha, WI). The half-value layer of the indirect digital radiography system was 3.0 mm aluminum at 80 kVp. We used a general-purpose 12:1 focused anti-scatter grid (Mitaya Manufacturing Co., Tokyo, Japan). The listed grid performance characteristics were 70 lines per centimeter, 12:1 grid ratio, 100-cm focus distance, strip thickness of 40 μm, lead strip material, aluminum interspace material, and carbon fiber reinforced polymer cover material. A source-to-image distance of 100 cm was used to acquire all images on the indirect digital radiography system. All regionof-interest measurements were made on the “for-processing” images (images to which only detector corrections were applied, and not image processing). The pixel values in the for-processing images from the indirect digital radiography system had a linear relationship to image receptor dose.
Computed radiography The other half of the experiments was performed on a computed radiography system using barium fluorobromide (BaFBr) screens and an FCR 5000 digitizer (Fujifilm USA, Stamford, CT). All computed radiography images were acquired using a Philips OmniDiagnost Eleva radiography/ fluoroscopy system (Philips Healthcare, Andover, MA) and the same 35.6×43.2 cm (14×17-in.) imaging plate. The halfvalue layer of the OmniDiagnost was 3.7 mm aluminum at 80 kVp. We used a general-purpose 13:1 focused anti-scatter grid (Philips Healthcare, Andover, MA). The listed grid performance characteristics were 60 lines per centimeter, 13:1 grid ratio, 120-cm focus distance, strip thickness of 36 μm, lead strip material, fiber interspace material, and carbon fiber cover. A source-to-image distance of 110 cm was used to acquire all computed radiography images. An air gap of 13 cm existed between the exit surface of the phantom and the image receptor on the OmniDiagnost system. This is compared to an air gap of 3 cm for the X-ray system used in the indirect digital radiography study, more typical for abdominal radiography. The larger air gap was a consequence of both the curvature of the patient support, to which the rigid Solid Water×™ slabs did not conform, and the geometry of the X-ray system. The plate was digitized using the Fuji Ave 4.0 test-image processing algorithm and transferred as a 10-bit image. The pixel values in the resulting images had a logarithmic relationship to image receptor dose. We decalibrated the images by measuring the characteristic function of the computed radiography system and using the inverse of the resulting function to transform images to linear image receptor dose space. All region-of-interest measurements were made on the de-calibrated images. Analysis Automatic exposure control was used to acquire an initial image for each combination of X-ray field of view, phantom
316
Pediatr Radiol (2014) 44:313–321
thickness and kVp for both the indirect digital radiography and computed radiography experiments. The fixed mAs value closest to the resulting automatic exposure control mAs value was used to acquire a series of six images for each kVp, FOV and phantom thickness combination—three without the beam stops in place and three with the beam stops in place. Images were analyzed using ImageJ v. 1.45 s (National Institutes of Health, Bethesda, MD). The signal from scatter, SS, was measured in regions of interest behind the disks in images with the beam stops in place and the total signal, ST, was measured using identical regions of interest in images without the beam stops in place. The regions of interest were half the diameter of the disks. SS and ST were then used to calculate the primary signal, SP = ST − SS. The scatter-to-primary ratio was calculated using the expression scatter − to − primary ratio ¼
SS SS ¼ S P S T −S S
ð2Þ
Because the scatter-to-primary ratio varies with beam stop diameter [19], it was necessary to determine the disk-free scatter-to-primary ratio, scatter-to-primary ratio0. This was accomplished by applying a linear fit to a plot of ln(scatterto-primary ratio) versus beam stop diameter. The intercept of this fit was equal to ln(scatter-to-primary ratio0). Exposure was measured for each automatic exposure controlled exposure using an ionization chamber (Triad Model 35050A Dosimeter, 15 cm3 ion chamber 96035B, Fluke Biomedical, Everett, WA). Exposure was converted to incident air kerma: Incident air kerma ðmGyÞ ¼ Exposure ðRÞ 8:76; and the dose area product (DAP) was calculated as the product of the incident air kerma and the FOV.
Figure 2 shows an example of the results for scatter-toprimary ratio0 versus phantom thickness for different X-ray fields of view at 80 kVp for indirect digital radiography. The standard errors of the fit intercepts used to calculate scatter-toprimary ratio0 were on the order of 10−2 (coefficient of variation less than 0.005). Systematic errors in our experiments were an order of magnitude higher than these experimental errors; these are examined in the Discussion section. The dashed line indicates scatter-to-primary ratio0 for the standard adult technique with the 23-cm phantom at 80 kVp with an anti-scatter grid and an FOVof 35.6×43.2 cm (14×17 in.). The intersection of this dashed line with the fits of scatter-to-primary ratio 0 versus phantom thickness indicates the threshold thickness for use of an anti-scatter grid for a given combination of X-ray FOV and kVp. The threshold thicknesses tended to be slightly lower for computed radiography compared to indirect digital radiography, a consequence of the different effective atomic numbers of the phosphors used in CR and iDR. Figure 3 plots threshold thicknesses versus FOV for 60 kVp, 70 kVp and 80 kVp and fits these data with power functions. Threshold thicknesses for commonly used clinical X-ray fields of view are listed in Table 1. The fit parameters, which can be used by medical physicists to calculate the threshold thickness for any FOV when constructing technique charts for pediatric abdominal radiography, are provided in Table 2. The dose area product resulting from the different experimental configurations is plotted in Figs. 4 and 5. Both figures plot the DAP calculated from the product of the measured exposure and the X-ray FOV at each kVp for all phantom thicknesses. Dose area product increased steeply as kVp decreased when using automatic exposure control.
Results The 20-mm disk was omitted in all scatter-to-primary ratio analyses because it fell partly outside the X-ray FOV for the smallest FOV evaluated. The scatter-to-primary ratio was measured using an anti-scatter grid for our standard adult and extrapolated to the full 35.6×43.2 cm (14×17 in.). Scatter-to-primary ratio was 0.78 for indirect digital radiography and 1.1 for computed radiography. The scatterto-primary ratio was higher for CR compared to iDR, a consequence of the fact that BaFBr, the most commonly used powdered CR phosphor, has a lower effective atomic number than CsI, the phosphor used in indirect DR systems. Phosphors with lower effective atomic numbers absorb scattered X-rays more efficiently than do phosphors with higher effective atomic numbers.
Fig. 2 Disk-free scatter-to-primary ratio, called scatter-to-primary ratio0, versus phantom thickness for different X-ray fields of view at 80 kVp using the indirect digital radiography system. The horizontal dashed line indicates scatter-to-primary ratio0 for the standard adult technique, the only technique for which an anti-scatter grid was used. Drop lines indicate threshold thicknesses for anti-scatter grid use
Pediatr Radiol (2014) 44:313–321
317
Fig. 3 Patient anteroposterior (AP) thickness thresholds for antiscatter grid use plotted versus field of view for each kVp for (a) indirect digital radiography (iDR) and (b) computed radiography (CR)
Discussion The results of this study indicate that for a given patient thickness the scatter-to-primary ratio is strongly dependent on X-ray FOV and weakly dependent on kVp. This is consistent with previous investigations [20]. These results reinforce the importance of tight collimation of the X-ray field Table 1 Threshold patient AP thicknesses (cm) for anti-scatter grid use for common clinical fields of view (inches) iDR
60 kVp 70 kVp 80 kVp
CR
to the anatomy of interest. Collimation has the added benefit of reducing DAP and therefore patient dose. The decision to use an anti-scatter grid should be based on the smaller of the X-ray FOV or patient dimensions. For example our results indicated that when using our large X-ray field of view, the patient thickness threshold for grid use was 6 cm or 7 cm. However, consider the case of an Table 2 Fit parameters from Fig. 3. These parameters can be used to construct technique charts for pediatric abdominal radiography. The fitted equation is y = ax b , where y = AP thickness threshold (cm) and x = FOV (cm2)
14×17a
10×12
8×10
14×17
10×12
8×10
iDR
5 5 5
9 9 8
12 11 11
6 5 6
9 9 9
13 12 12
a
b
1,250 1,024 1,020
−0.746 −0.720 −0.727
a
14×17 in.=35.6×43.2 cm; 10×12 in.=25.4×30.5 cm; 8×10 in.=20.3× 25.4 cm
AP anteroposterior, iDR indirect digital radiography, CR computed radiography
60 kVp 70 kVp 80 kVp
CR a 1,326 1,392 820
b −0.745 −0.760 −0.679
AP anteroposterior, iDR indirect digital radiography, CR computed radiography
318
Pediatr Radiol (2014) 44:313–321
Fig. 4 Dose area product (DAP) as a function of kVp for indirect digital radiography (iDR) for the (a) large, (b) medium and (c) small X-ray field of view (FOV)
uncooperative child, where the collimation must be broad so that the anatomy of interest is not missed. Because the patient is the primary source of scatter, the scatter-to-primary ratio is not appreciably increased if the collimated X-ray FOVextends beyond the patient anatomy. There was a slight increase in the threshold thickness for anti-scatter grid use with decreasing kVp (Fig. 3) (Table 1); however, there was a much steeper associated dose penalty (Figs. 4 and 5). For a constant receptor dose, dose area product increases with decreasing kVp when using automatic exposure control, especially for thicker patients. Because scatter-to-primary ratio is weakly dependent on kVp it is important to consider the cost of reducing kVp to reduce scatter. The AP threshold thicknesses and transverse X-ray field dimensions in this study correspond to children across a wide age range. A study by Kleinman et al. [21] in 2010 demonstrated the large degree of overlap in child sizes from toddlers to teens. For example, in the current study the transverse dimension of the small FOV in the patient plane
was approximately 16 cm. According to the Kleinman [21] study, this corresponds to the mean transverse dimension of a 1-year-old child, but children up to age 10 may share this transverse dimension. In the current study, the threshold thickness at 80 kVp for the small FOV for indirect digital radiography was 14 cm. This corresponds to the mean AP thickness of a 4-year-old child but also includes children up to 14 years old who are smaller than average. These large variations mean that it is difficult to base technique factor selection, including use of anti-scatter grids, on patient age. Instead, the patient should be measured along the appropriate dimensions before a radiographic exam and technical factors should be selected based on this information. This study had several limitations. First, our patient phantom dimensions were limited by the areal dimensions of the Solid Water™ slabs available to us. In many cases the extent of the phantom, particularly in the transverse dimension, exceeded the dimensions of actual patients. Therefore it is likely that some X-rays that scattered out of the field of view and normally would not have reached the detector scattered
Pediatr Radiol (2014) 44:313–321
319
Fig. 5 Dose area product (DAP) as a function of kVp for computed radiography (CR) for the (a) large, (b) medium and (c) small X-ray field of view (FOV)
back into the FOV and contributed to Ss. This would increase the scatter-to-primary ratio slightly compared to phantoms with areal dimensions closer to those of actual patients. In the adult case, our maximum X-ray FOV was limited by the phantom dimensions, which required us to extrapolate to find the scatterto-primary ratio0 for the full adult FOV. However considering the gradual, predictable increase in scatter-to-primary ratio at large FOV [22], the error introduced was small. Second, there was a slight difference between the source-to-image distance used and the grid focus distance for the computed radiography system used in this study. Based on the magnitude of this difference, the geometry and dimensions of our beam stop array, and the ratio of the grid [22], we have estimated that the signal from primary radiation was underestimated by 3– 5%, depending on the exact position of the beam stops. This would result in an overestimation of scatter-to-primary ratio0 by a similar factor for our standard adult technique. In the end this would result in overestimation of threshold thicknesses by approximately 1 cm in some cases. Third, an air gap of 13 cm existed between the exit surface of the phantom and the image
receptor on the OmniDiagnost system used to acquire computed radiography images. The presence of this air gap would be expected to result in a slight decrease in the scatterto-primary ratio for each of our measurements, with larger decreases for thicker phantoms [23]. The impact of the air gap was expected to be an underestimation of threshold thicknesses by approximately 5%. Although we cannot be sure what the sum of these two effects was, it is unlikely that the threshold thicknesses presented here are in error by more than +/− 5%. Finally, the characteristics of the antiscatter grids used for the indirect digital radiography and computed radiography experiments were slightly different. These results have not been validated in clinical practice, and in light of the study limitations presented here, our recommendations should be validated prior to clinical use. Another important consideration is that pediatric radiologists are often willing to tolerate images of marginal diagnostic quality in an effort to minimize patient dose. Therefore, our choice to benchmark scatter-to-primary ratio using a standard adult technique may have resulted in a scatter-to-primary ratio
320
that was lower than the value of scatter-to-primary ratio that can be tolerated in pediatric radiography. However, our results can be adapted to any choice of baseline scatter-to-primary ratio.
Conclusion It is important that X-ray technologists be educated about the use of anti-scatter grids, including the costs and benefits of removing the grid. Technique charts should include clear guidelines for removing the anti-scatter grid when imaging patients of specific dimensions. Technologists should know which X-ray systems in their department have anti-scatter grids that can be removed and be instructed on how to remove the grid, which may require the selection of specific imaging presets at the control console. There is currently a recommendation from the Image Gently campaign that manufacturers develop technology to allow automatic removal of the anti-scatter grid, as opposed to manual removal, to aid the technologist and improve workflow [24]. When appropriate, pediatric patients should either be protocoled to rooms that have removable anti-scatter grids or undergo tabletop imaging. The data presented in this study can be used to create technique charts for pediatric abdominal radiography. Technique charts should emphasize the importance of measuring patient dimensions, either using a prior cross-sectional imaging exam or physical calipers, rather than relying on patient age to determine proper technique. The medical physicist should work closely with the radiologist to determine what image quality is required for specific procedures and if removing the anti-scatter grid is appropriate given the required image quality and potential impact on patient dose [9]. In many indications for abdominal radiography lower contrast is acceptable, for example high subject contrast tasks such as visualizing air-fluid levels. Thus, the anti-scatter grid might be removed for a wider range of pediatric sizes than indicated by our data. It would be beneficial to repeat this study for other common pediatric radiographic examinations where no consensus guidelines exist for anti-scatter grid use, for example chest radiography and fluoroscopy, and to repeat these experiments using other image receptors such as amorphous selenium. Conflicts of interest None.
References 1. Carlton RR, Adler AM (2012) The grid. In: Principles of radiographic imaging: an art and a science, 5th edn. Delmar Cengage Learning, Clifton Park, pp 257–272
Pediatr Radiol (2014) 44:313–321 2. Alzen G, Benz-Bohm G (2011) Radiation protection in pediatric radiology. Dtsch Arztebl Int 108:407–414 3. European Commission (1996) European guidelines on quality criteria for diagnostic images in paediatrics. ftp://ftp.cordis. lu/pub/fp5-euratom/docs/eur16260.pdf. Accessed May 1, 2013 4. American College of Radiology (2008) ACR-SPR practice guidelines for general radiography. http://www.acr.org/~/ media/95AB63131AD64CC6BB54BFE008DBA30C.pdf. Accessed May 1, 2013 5. International Atomic Energy Agency (2013) Safety report series No. 71, radiation protection in paediatric radiology. http://www-pub.iaea. org/books/iaeabooks/8727/Radiation-Protection-in-PaediatricRadiology. Accessed May 1, 2013 6. Ween B, Olstad M, Jakobsen JA et al (2009) Pediatric digital chest radiography, comparison of grid versus non-grid techniques. Eur J Radiography 1:201–206 7. IEC International Standard 60627 (2001) Diagnostic X-ray imaging equipment—characteristics of general purpose and mammographic antiscatter grids. International Electrochemical Commission, ISBN 28318-5953-0 8. Hall EJ (2002) Lessons we have learned from our children: cancer risks from diagnostic radiology. Pediatr Radiol 32: 700–706 9. Amis ES Jr, Butler PF, Applegate KE et al (2007) American College of Radiology white paper on radiation dose in medicine. J Am Coll Radiol 4:272–284 10. Conway BJ, Duff JE, Fewell TR et al (1990) A patient-equivalent attenuation phantom for estimating patient exposures from automatic exposure controlled X-ray examinations of the abdomen and lumbosacral spine. Med Phys 17:448–453 11. American College of Radiology (2008) ACR Practice Guideline for Diagnostic Reference Levels in Medical X-ray Imaging. http://www. acr.org/~/media/796DE35AA407447DB81CEB5612B4553D.pdf. Accessed May 1, 2013 12. Yaffe M, Fenster A, Johns HE (1977) Xenon ionization detectors for fan beam computed tomography scanners. J Comput Assist Tomogr 1:419–428 13. Floyd CE Jr, Lo JY, Chotas HG et al (1991) Quantitative scatter measurement in digital radiography using a photostimulable phosphor imaging system. Med Phys 18:408–413 14. Fetterly KA, Schueler BA (2007) Experimental evaluation of fiber-interspaced antiscatter grids for large patient imaging with digital x-ray systems. Phys Med Biol 52: 4863–4880 15. American College of Radiology (2012) ACR-AAPM-SIIM practice guideline for digital radiography. http://www.acr.org/~/media/ 3E08C87AD6E6498D9E19769E5E5E390D.pdf. Accessed May 1, 2013 16. American College of Radiology (2011) ACR-SPR guideline for the performance of abdominal radiography. http://www.acr.org/~/media/ 79A594819BBD4631A7E31404DAA66EF6.pdf. Accessed May 1, 2013 17. Rosenstein M, Beck TJ, Warner GG (1979) Handbook of selected organ doses for projections common in pediatric radiology. HEW Publication FDA pp. 79–8079 18. Rosenstein M (1988) Handbook of selected tissue doses for projections common in diagnostic radiology. HEW Publication FDA pp. 89–8031 19. Maher KP (1993) Comparison of scatter measurement techniques in digital fluoroscopy. Phys Med Biol 38:1977–1983 20. Yaffe MJ, Johns PC (1983) Scattered radiation in diagnostic radiology—magnitudes, effects, and methods of reduction. J Appl Photogr Eng 9:184–195
Pediatr Radiol (2014) 44:313–321 21. Kleinman PL, Strauss KJ, Zurakowski D et al (2010) Patient size measured on CT images as a function of age at a tertiary care children’s hospital. AJR Am J Roentgenol 194:1611–1619 22. Aichinger H, Dierker J, Joite-Barfuß S et al (2012) Scattered radiation. In: Radiation exposure and image quality in X-Ray diagnostic radiology, 2nd edn. Springer, Berlin, pp 53–66
321 23. Curry TS III, Dowdey JE, Murry RC Jr (1990) Christensen’s physics of diagnostic radiology, 4th edn. Lippincott, Williams, and Wilkins, Philadelphia 24. Goske MJ, Charkot E, Herrmann T et al (2011) Image Gently: challenges for radiologic technologists when performing digital radiography in children. Pediatr Radiol 41:611–619
