British Journal of Radiology 75 (2002),596-602 © 2002 The British Institute of Radiology
A practical demonstration of improved technique factors in paediatric fluoroscopy
J W Fenner, PhD1,
G D Morrison, MSc1,
J Kerry, MSc2 and
N West, MSc3
1 Department of Medical Physics and Clinical Engineering, Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF, 2 Medical Physics Department, County Hospital, Greetwell Road, Lincoln LN2 5QY and 3 Radiotherapy Physics, Derbyshire Royal Infirmary NHS Trust, London Road, Derby DE1 2QY, UK
 |
Abstract
|
|---|
Dose incurred with fluoroscopic procedures accounts for a significant proportion of medically induced diagnostic exposure. Children are particularly vulnerable and it is therefore important to minimize exposure where practicable. A recent theoretical study has highlighted the potential for X-ray equipment to produce significant dose savings during paediatric fluoroscopy without incurring loss of diagnostic image quality. This is achieved by hardening the beam with additional copper (Cu) filtration (
0.2 mm Cu) and biasing exposure factors towards low tube potential, high tube current output. In practice, this method will have limited applicability because the high powered and programmable generator characteristics required are not commonly available in installations used for paediatric imaging. However, we describe a simple experiment in which our clinical equipment was modified to approximate desired low dose performance by altering the filtration and automatic exposure control characteristics of ordinary clinical equipment in the Sheffield Children's Hospital. This enabled us to obtain significant savings in dose. We performed a comparative study (normal dose vs low dose) using water phantoms to simulate patient attenuation in the age range 0–15 years. The Leeds N2 contrast sensitivity phantom was used to provide a measure of image quality. Dosimetric measurements recorded up to 40% reduction in dose rate with only marginal loss of image quality when 0.1–0.2 mm Cu filtration was used with the modified settings. This is a strong indication that significant dose reduction is achievable on routine clinical equipment without compromising image quality. Such simple and cost effective methods of dose reduction should be considered for wider implementation.
|
Introduction
|
|---|
Fluoroscopy is an important but relatively high dose tool in radiodiagnosis [1–4], and provides real-time images of dynamic procedures. The recent work of Tapiovaara et al [5] indicates that appropriate choice of X-ray factors combined with heavy copper (Cu) filtration of a diagnostic X-ray beam can lead to significant (30–50%) dose reduction in paediatric fluoroscopy without loss of image quality. These findings have implications for general paediatric fluoroscopy since they imply that substantial dose reductions may be possible for all paediatric fluoroscopic examinations. It is clearly important to consider the practical relevance of Tapiovaara et al's work to routine clinical equipment in the NHS [6, 7], and this is the focus of this paper.
|
Background
|
|---|
Conventional wisdom and experience have demonstrated that dose reduction can always be achieved by either increasing tube potential and/or increasing filtration at the tube [8]. Unfortunately, this generally results in poorer image quality since the image intensifier receives a reduced number of higher energy photons for a satisfactory exposure, i.e. energy integrating detector, leading to increased quantum noise in the image. In contrast, Tapiovaara et al recommend the use of reduced tube potential and increased filtration to obtain significant dose reduction with insignificant loss of image quality. Their work modelled the fluoroscopic imaging chain and characterized image quality as a function of dose. This demonstrated that the ratio of image quality to dose was a maximum at approximately 50 kVp, with the beam heavily filtered to produce a narrow band spectrum.
However, the automatic exposure control (AEC) performance of many modern diagnostic X-ray units is modelled on the traditional former ethos rather than Tapiovaara et al's model, and standard X-ray equipment lacks the necessary programmability to achieve Tapiovaara et al's exposure criteria. It is prudent to enquire to what extent, and at what cost, routine equipment can be modified to operate effectively in the low dose domain. Tapiovaara et al used a highly sophisticated Monte Carlo model of their equipment to determine factors necessary for low dose/high image quality operation. Can this model be practically implemented in standard equipment commonly found in radiology departments with the facilities available to the average medical physics and radiology departments? What are the limitations of clinical fluoroscopic equipment in this regard? Are the critical parameters of the Monte Carlo model unduly compromized by the limitations of routine machine specification such that significant lower dose operation becomes unachievable in the hospital environment? It is apparent that application of Tapiovaara et al's method will be limited in practice, because most commonly available X-ray fluoroscopy installations used for paediatric imaging are not flexible enough to accommodate the suggested exposure modifications.
Being mindful of such constraints, we sought a practical implementation of the ideal low dose technique applied to an ordinary X-ray fluoroscopy system at the Sheffield Children's Hospital. With limited, simple modifications to our equipment, we were able to demonstrate significant reductions (up to 40%) in absorbed dose-rate, with negligible loss of image quality. Our results indicate that even simple modification of clinical fluoroscopy equipment can yield significant dose savings in the clinical setting.
|
Method
|
|---|
This experiment records the effect of modified AEC sensitivity and unit filtration on dose rate and image quality. Water phantoms of varying thickness were used to represent patient attenuation across a range of energies.
The experimental arrangement is illustrated in Figure 1
, with the image intensifier and water phantom above the undercouch X-ray tube (P125/30/50CR X-ray tube with a Siemens Polyphos 50 generator (Siemens, Bracknell, UK); tube installed 1989). Owing to access problems, the Cu filtration was placed on the couch beneath the phantom. The Cu sheet (purity, 99.9%; temper, half hard) was of sufficient size to fill the field of view.
View larger version (25K):
[in this window]
[in a new window]
|
Figure 1. Equipment set-up. The display monitor shows a representation of the Leeds N2 phantom; a ring of discs of diminishing contrast. The ionization chamber is inserted into the beam for dose rate measurements. Added copper filtration is on the couch since it was impractical to attach it to the tube for the purposes of this experiment.
|
|
The fluoroscopy unit has two dose rate settings, normal and high. Prior to imaging, the high dose rate performance of the fluoroscopy system was modified by the service engineer (a simple 10 minute task) to accommodate the introduction of additional Cu filtration. The normal setting was left unchanged.
Image quality was measured for a series of phantom thicknesses with a Leeds N2 contrast sensitivity phantom [9–11] (a commonly used and accepted test object) placed nearest the intensifier on top of the water block. The image intensifier was lowered to the same distance above the test object for each phantom thickness, replicating clinical practice and avoiding magnification in the image. The contrast phantom was then imaged under dynamic fluoroscopy and the number of discs visible on the display monitor recorded as an image quality score. Fluoroscopy tube potential and tube current selected by the AEC were also recorded. Ionization chambers, connected to a traceably calibrated dose rate metre (Radcal MDH1515; Radcal Corporation, Monrovia, CA), were used to measure entrance and exit doses to the water phantom, from which absorbed dose was calculated. These procedures were performed for water phantom thicknesses (±0.1 cm) of 5 cm, 10 cm, 14.7 cm and 19.7 cm, approximating attenuation in patients of ages 0–15 years [5, 12].
Finally, a complete set of data was obtained at normal fluoroscopy settings, i.e. no exposure modification, without the presence of additional Cu filtration in the beam. This provided a comparative set of data relating dose to image quality at a range of patient thickness under standard operating conditions.
|
Results
|
|---|
A selection of the exposure and image quality data collected is presented in Table 1. For simplicity, absorbed dose has been taken to be proportional to the difference between exit and entrance dose rates. However, the presence of scatter complicates dose comparison since it is a function of phantom thickness [12, 13] and, therefore, percentage changes in absorbed dose rate (comparing normal and low dose operation at the same thickness) are quoted in our final analysis.
View this table:
[in this window]
[in a new window]
|
Table 1. Example data obtained from our experiment. (a) Normal image quality and dose rate data for a range of phantom thicknesses. (b) Marginally increased image quality results from repeating exposures at high dose setting. (c) Insertion of 0.1 mm copper (Cu) filtration reduces dose significantly with image quality comparable to the normal setting. (d) Further addition of Cu thickness (0.3 mm) reduces both dose and image quality
|
|
The data of Figure 2 show the measured decrease in image quality resulting from increasing phantom thickness, with and without additional filtration. As would be expected, the highest image quality is associated with low water phantom and Cu thickness; a combination that presents the highest flux with minimal scatter to the intensifier. However, when data are presented as change of image quality related to percentage change in absorbed dose rate, as compared with data at normal settings, the dose reduction available with increased filtration becomes apparent (Figure 3). Without Cu, the high dose setting results in marginally better images, but with a 40% increase in absorbed dose over the normal setting. Contrast this with the 40% reduction in absorbed dose obtained with 0.2 mm Cu filtration, with only minimal loss of image quality. This evidence echoes the results of Tapiovaara et al; namely that with a suitable thickness Cu filtration combined with optimized exposure factors, absorbed dose during a fluoroscopic procedure can be dramatically reduced with little or no loss of image quality.
View larger version (30K):
[in this window]
[in a new window]
|
Figure 2. Measured image quality as a function of copper (Cu) filtration and water phantom thickness. The grey band is a qualitative fit to the data and highlights the trend ofdecreasing image quality with increasing Cu and phantom thickness. For reasons of clarity, no error bars are shown. Phantom thickness error approximates ±0.1 cm. Image quality error approximates ±0.5 discs. +, normal dose rate setting, no Cu;  , high dose rate setting, 0 mm Cu;  , high dose rate setting, 0.1 mm Cu;  , high dose rate setting, 0.2 mm Cu; x, high dose rate setting, 0.3 mm Cu;  , high dose rate setting, 0.4 mm Cu; •, high dose rate setting, 0.5 mm Cu.
|
|
View larger version (36K):
[in this window]
[in a new window]
|
Figure 3. Percentage change in absorbed dose rate against change in image quality compared with the same measurements at the normal setting without added filtration. The four points in each group come from the four phantom thicknesses used (5 cm, 10 cm, 14.7 cm and 19.7 cm) at each copper (Cu) thickness. Errors in percentage dose change are not shown but equate to ±4%.  , high dose rate setting, 0 mm Cu; , high dose rate setting, 0.1 mm Cu; , high dose rate setting, 0.2 mm Cu; x, high dose rate setting, 0.3 mm Cu; , high dose rate setting, 0.4 mm Cu; •, high dose rate setting, 0.5 mm Cu.
|
|
|
Discussion
|
|---|
AEC performance modification
The inflexible nature of the AEC performance curves of the Siemens unit used in this experiment is illustrated by Figure 4. The AEC adjusts the exposure factors, peak tube potential and tube current, in such a way as to provide an image of adequate contrast and brightness on the monitor screen. This is determined by monitoring the signals responsible for image intensity at the display monitor, whilst feedback mechanisms control the exposure factors, governed by the gradient of the tube potential/tube current relationship shown in Figure 4. Insufficient photon flux at the intensifier is corrected by increasing tube potential and tube current until adequate image brightness is achieved. Conversely, excessive exposure moves the factors in the reverse direction, down the tube potential/tube current line.
The rationale behind this choice of AEC performance is that soft tissue contrast is enhanced by using a softer spectrum, maximizing differences in attenuation, whilst more radio-opaque objects require higher tube potentials in an effort to increase transmission. The ideal response, proposed by Tapiovaara et al, consists of a highly filtered narrow band spectrum at approximately 50 kV, with exposure controlled predominantly by tube current. In common with most paediatric equipment, the Siemens unit used here had neither the flexibility nor the generator power to operate under such conditions. Nevertheless, we adjusted the AEC sensitivity of the high dose mode to provide satisfactory image quality with additional Cu filtration, and were able to record significant dose reductions as a result.
Comparison with the idealized dose reduction scheme of Tapiovaara et al
Tapiovaara et al's work characterized image quality as the square of the signal-to-noise ratio (SNR2) of the observed image. In practice, the SNR2 values of an image are not easily obtained. Therefore, we chose a more accessible, but semi-quantitative, measure routinely used in quality assurance; the number of dots visible on the Leeds N2 contrast sensitivity phantom. This produces a subjective and relative assessment of image quality, but is widely used and known to be reproducible [9–11].
A pragmatic approach was used to determine absorbed dose, and this was taken to be proportional to the difference between entrance and exit dose rates at the phantom. This incorporates backscatter at the entrance surface, which may increase entrance dose rate by 10–20%, and is also a function of patient thickness, compounded by inverse square law effects [12, 13]. Given these complications, we avoided reporting absolute measurements of absorbed dose and chose to make relative comparisons, charting the relative change in absorbed dose rate as a function of change of image quality. This eliminates many errors and provides a better estimate of the comparative effect of the dose reduction scheme, to first order.
Image quality and dose reduction
Figure 2
shows the steady fall in image quality associated with increasing Cu and tissue thickness. Image quality is a critical aspect of X-ray system performance in fluoroscopy, because images of adequate standard are essential to accurate diagnosis. The radiologist is looking for sufficient contrast differentiation to enable discernment of adjacent low contrast (similar attenuation) objects at a resolution sufficient to identify features that may be diagnostically important. This process is the rationale behind the design of image quality phantoms such as the N2, but its use necessarily reflects the perceptions of the observer and will therefore vary between observers. It can be argued that a more objective measure should be used. However, it is not clear to what extent purely objective measures accurately describe the clinical perception of image quality. Although there is no strong consensus on which indices of image quality should be used to permit subjective or objective comparison of systems/configuration performance [14], it has been demonstrated that by using tools such as the Leeds N2 phantom, an experienced observer can differentiate between different levels of image quality both consistently and repeatedly [15].
The experience of the diagnostic radiology physics group at Sheffield indicates that intra-observer variations in identifying the number of discs visible on image quality phantoms such as the N2 is typically ±0.5 discs, and therefore the decreased visibility of two discs over the 20 cm water phantom thickness range used in this experiment represents a significant loss of image quality. For similar reasons the ±0.5 disc wide bar on Figure 3 is quoted as "negligible" image quality change since it falls within the reproducibility error of image quality assessment. The curve is a qualitative fit to highlight the trend of the data. It demonstrates that the high dose setting without Cu results in marginally better image quality compared with normal, but at much increased dose. However, there are points lying within the marginal band that are associated with over 40% reduction in dose. These represent significant dose reduction with little loss of image quality. The majority of these points are those with 0.1 mm Cu filtration, whereas 0.2 mm filtration produces larger dose reduction but with a more demonstrable loss of image quality.
Implications
The beam hardening obtained with the use of Cu demonstrates the importance of filtration, since it dramatically reduces that proportion of photons most likely to be absorbed within the patient. However, low imaging tube potential is also important because it is responsible for reduced scatter and increased tissue contrast. In the limit, the contrary requirements of increasing filtration and decreasing tube potential meet to form a narrow, ideally monoenergetic, spectrum [16] at a particular peak tube potential that maximizes image quality and minimizes dose. This additional filtration reduces tube output and requires increased tube current to maintain adequate photon fluence at the intensifier. This has implications for generator power and may shorten tube life if applied to adult imaging, but the lower factors typical of paediatric radiology are likely to mitigate such adverse effects.
Unfortunately, the AEC of our equipment was unable to be programmed for high tube current, low tube potential output. This compromizes the achievable dose reduction. Nevertheless we were able to demonstrate dose reductions of up to 40%. We conclude that even limited implementation of the dose reduction scheme of Tapiovaara et al can be an effective measure and therefore should be considered for wider implementation. To that end we are currently seeking ethical permission to apply this technique to paediatric examinations in the clinical environment. We anticipate that a dose/image quality survey will demonstrate significant benefits.
|
Conclusion
|
|---|
This work has considered the dose reduction recommendations of Tapiovaara et al [5] and applied these principles to routine paediatric fluoroscopy in the clinical setting. Using slightly modified exposure factors, additional Cu filtration and tissue phantoms, dose rate measurements were obtained for a range of patient equivalent thicknesses. Results demonstrate that significant reductions in dose and minimal loss of image quality can be achieved with existing X-ray equipment by the addition of 0.1–0.2 mm Cu filtration to the beam. This is in agreement with Tapiovaara et al's data and provides evidence for wider use of the technique in the real clinical environment, since it is a simple and cost effective measure.
|
Acknowledgments
|
|---|
The authors are keen to acknowledge the cooperation of Dr P Broadley (Consultant Radiologist) and the support provided by Mr Stephen Howe (Superintendent Radiographer) and his staff in the Department of Paediatric Radiology at Sheffield Children's Hospital.
Received for publication March 22, 2001.
Accepted for publication January 29, 2002.
|
References
|
|---|
- Geise RA, O'Dea TJ. Radiation dose in interventional fluoroscopic procedures. Appl Radiat Isot 1999;50:173–84.[Medline]
- Ruiz Cruces R, Garcia-Granados J, Diaz Romero FJ, Hernandez Armas J. Estimation of effective dose in some digital angiographic and interventional procedures. Br J Radiol 1998;71:42–7.[Abstract]
- Schultz FW, Geleijins J, Holscher HC, Weststrate J, Zonderland HM, Zoetelief J. Radiation burden to paediatric patients due to micturating cystourethrography examinations in a Dutch children's hospital. Br J Radiol 1999;72:763–72.[Abstract]
- Persliden J, Pettersson HB, Falth-Magnusson K. Radiation dose at small intestinal biopsies in children: results of a national study. Acta Paediatr 1996;85:1042–6.[Medline]
- Tapiovaara MJ, Sandborg M, Dance DR. A search for improved technique factors in paediatric fluoroscopy. Phys Med Biol 1999;44:537–9.[Medline]
- The Ionising Radiations Regulations 1999 (Statutory Instrument 1999 No.3232). London: HMSO, 2000.
- The Ionising Radiation (Medical Exposure) Regulations 2000 (Statutory Instrument 2000 No. 1059). London: HMSO, 2000.
- Brown PH, Thomas RD, Silberberg PJ, Johnson LM. Optimisation of a fluoroscope to reduce radiation exposure in paediatric imaging. Pediatr Radiol 2000;30:229–35.[Medline]
- Hay GA, Clarke OF, Coleman BA, Cowen AR. A set of X-ray test objects for quality control in television fluoroscopy. Br J Radiol 1985;58:335–44.[Abstract]
- Cowen AR, Haywood JM, Workman A, Clarke OF. A set of X-ray test objects for image quality control in digital subtraction fluorography. I: design considerations. Br J Radiol 1987;60:1001–9.[Abstract]
- Cowen AR, Haywood JM, Workman A, Clarke OF. A set of X-ray test objects for image quality control in digital subtraction fluorography. II: application and interpretation of results. Br J Radiol 1987;60:1010–8.
- Persliden J, Sandborg M. Conversion factors between energy imparted to the patient and air collision kerma integrated over beam area in paediatric radiology. Acta Radiol 1993;34:92–8.[Medline]
- Patrocinio HJ, Bissonnette J-P, Bussiere MR, Schreiner LJ. Limiting values of backscatter factors for low-energy X-ray beams. Phys Med Biol 1996;41:239–53.[Medline]
- Marsh DM, Malone JF. Methods and materials for the measurement of subjective and objective measurements of image quality. Rad Prot Dosim 2001;94:37–42.
- Launders JH, McArdle S, Workman A, Cowen AR. Update on the recommended viewing protocol for FAXIL threshold contrast detail detectability test objects used in television fluoroscopy. Br J Radiol 1995;68:70–7.[Abstract]
- Boone JM, Seibert JA. A comparison of mono- and poly-energetic X-ray beam performance for radiographic and fluoroscopic imaging. Med Phys 1994;21:1853–63.[Medline]
This article has been cited by other articles:
|
|
|
|
 
R S Livingstone, C G Koshy, and D V Raj
Evaluation of work practices and radiation dose during adult micturating cystourethrography examinations performed using a digital imaging system
Br. J. Radiol.,
November 1, 2004;
77(923):
927 - 930.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D G W Onnasch, A Schemm, and H-H Kramer
Optimization of radiographic parameters for paediatric cardiac angiography
Br. J. Radiol.,
June 1, 2004;
77(918):
479 - 487.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
