Radiation Protection Dosimetry (2010), Vol. 139, No. 1 –3, pp. 371–374 Advance Access publication 11 March 2010
doi:10.1093/rpd/ncq094
ARE EXPOSURE INDEX VALUES CONSISTENT IN CLINICAL PRACTICE? A MULTI-MANUFACTURER INVESTIGATION M. L. Butler 1,*, L. Rainford 1, J. Last 1 and P. C. Brennan 2 1 School of Medicine and Medical Science, University College Dublin, Dublin, UK 2 Discipline of Medical Radiation Sciences, University of Sydney, NSW, Australia
The advent of digital radiography poses the risk of unnoticed increases in patient dose. Manufacturers have responded to this by offering an exposure index (EI) value to the clinician. Whilst the EI value is a measure of the air kerma at the detector surface, it has been recommended by international agencies as a method of monitoring radiation dose to the patient. Recent studies by the group have shown that EI values are being used in clinical practice to monitor radiation dose and assess image quality. This study aims to compare the clinical consistency of the EI value in computed radiography (CR) and direct digital radiography (DR) systems. An anthropormorphic phantom was used to simulate four common radiographic examinations: skull, pelvis, chest and hand. These examinations were chosen as they provide contrasting exposure parameters, image detail and radiation dose measurements. Four manufacturers were used for comparison: Agfa Gaevert CR, Carestream CR, Philips Digital Diagnost DR and Siemens DR. For each examination, the phantom was placed in the optimal position and exposure parameters were chosen in accordance with European guidelines and clinical practice. Multiple exposures were taken and the EI recorded. All exposure parameters and clinical conditions remained constant throughout. For both DR systems, the EI values remained consistent throughout. No significant change was noted in any examination. In both CR systems, there were noteworthy fluctuations in the EI values for all examinations. The largest for the Agfa system was a variation of 1.88– 2.21 for the skull examination. This represents to the clinician a doubling of detector dose, despite all exposure parameters remaining constant. In the Kodak system, the largest fluctuation was seen for the chest examination where the EI ranged from 2560 to 2660, representing approximately an increase of 30 % in radiation dose, despite consistent parameters. The fluctuations seen with the CR systems are most likely due to image processing delay, replacing of the imaging plate and calibration factors. Fluctuations in EI values may result in confusion to the clinician and unnecessary repeat examinations. The reliability of EI values as a feedback mechanism for CR is also questionable.
INTRODUCTION Medical radiation is the largest source of man-made radiation(1). In 2004, it was suggested that medical radiation may be responsible for approximately 1 % of the cancer in the USA, 0.6 % in the UK and up to 3 % in Japan, although there is no scientific prove for this, and these figures are expected by some authors to be even higher today due to an increase in amount and type of examination(2). Despite radiography being considered a lower dose examination in comparison to computed tomography, it is important to assume that all exposures carry some risk as long as the opposite is not proven. Computed radiography (CR) and direct digital radiography (DR) were introduced into clinical departments in the 1980s. These systems immediately offered a wide range of advantages to the patient, clinician and physicist, including post-processing capabilities and a wide exposure latitude. However, the phenomenon of exposure creep in association with the advent of CR and DR has been largely documented(3). Exposure creep occurs because with lower doses in CR and DR, noise is more apparent in the image. This is not aesthetically
pleasing to the clinician, and therefore clinicians have tended to increase dose to compensate for this. In traditional film-screen radiography, over or under exposure to the patient was immediately apparent. In CR and DR, image processing can compensate by up to 100 % for under exposure and up to 500 % for over exposure, and still produce a clinically acceptable image(4). Manufacturers have responded to this by offering an exposure index (EI) value to clinicians. This value is a measure of radiation dose to the detector and is displayed with the image to the clinician. Different manufacturers use varying methods of EI as displayed in Table 1(5). The importance of the EI value as a feedback mechanism has been highlighted by a number of international bodies(6,7). However, in a clinical setting a wide range of factors can influence the EI value such as patient size, artefacts, source to image receptor distance, collimation, centring and IP plate size(3). Whilst the EI value has been promoted as a means of quality assurance testing for best clinical practice, its clinical consistency is not apparent in any recent literature. The current study aims to assess the consistency of the EI value at constant
# The Author 2010. Published by Oxford University Press. All rights reserved. For Permissions, please email:
[email protected]
Downloaded from http://rpd.oxfordjournals.org/ at Norges Teknisk-Naturvitenskapelige Universitet on November 17, 2014
*Corresponding author:
[email protected]
M. L. BUTLER ET AL.
phantom (Radiology Support Devices, USA) was used for the hand projection, as the RANDOw phantom available has no limbs.
Table 1. EI values by manufacturer. Manufacturer Agfa
Exposure indicator
Examinations and phantom positioning
conditions and to document any fluctuations that may occur.
Four common, but contrasting, clinical examinations were chosen for the study: skull, chest, abdomen and hand. Chest examinations are the most frequent Xray procedures in the world, accounting for over 8 million examinations per annum in the UK(8). Chest radiography requires detail of both a variety of soft tissue structures such as heart and lungs, and also the need to visualise bony details. Abdominal examinations are an example of a high dose procedure, with soft tissue structures being of paramount importance. Skull X-rays, while less frequent, are still being used in Europe, in particular to look at intricate bony detail. Hand X-rays have in the UK a frequency of almost 3 million examinations per year, and are an example of a low dose examination, with accurate visualisation of bony detail being imperative(8). For each examination, the phantoms were positioned in line with clinical practice and international guidelines. For the chest, the phantom was positioned erect facing the detector, which was raised so the top was 3 cm above the skin margins above the apices. The horizontal central ray was centred in the midline at the level of T4 – T6. For the abdomen, the phantom was aligned with its median sagittal plane perpendicular to the detector and the central ray was directed in the midline at the level of the iliac crests. The phantoms head was positioned so that the interpupilary line was parallel to the film for the skull projection and the central ray was angled 208 cranially and centred to the glabella. The hand phantom was placed in the dorsi-plantar position and the central ray was directed at the head of the third metacarpal(9).
METHODOLOGY
Procedure
Equipment
For each system and each projection, the phantom was exposed 20 times. The phantom was not moved between exposures. When using the CR systems, the same imaging plate was used, as was the same digitiser. The imaging plate was processed within 2 min of exposure. The corresponding EI value for each exposure was recorded.
Carestream (formerly Kodak)
Philips
Siemens
Four manufacturers were selected to aid comparison, two CR and two DR systems. Agfa Gaevert and Carestream Health ( previously Kodak) Direct View CR500 CR systems were used, with corresponding phosphor imaging plates and digitisers. Philips Digital Diagnost and Siemens Medical Axiom Aristos FX Plus Syngo DR systems were used for comparison. Two anthropomorphic phantoms were used to produce clinically relevant images and simulate typical examination conditions. The RANDOw Phantom was used for the chest, abdomen and skull projections and the PIXYw anthropomorphic
RESULTS The results for each manufacturer are displayed in Tables 2–5. The ranges of EI values given represent the minimum and the maximum value recorded for the same examination at constant conditions.
372
Downloaded from http://rpd.oxfordjournals.org/ at Norges Teknisk-Naturvitenskapelige Universitet on November 17, 2014
Logarithmic value (lGm): the median value of the ROI histogram defines the lGm for the image. LGm is a deviation index as it compares the resultant lGm value to a reference lGm value. It is recommended to read 1.9 for all examinations. EI: This is a numerical value computed from the average code value of those areas of the image data that are used by the image processing algorithm to compute the original tonescale. It has a logarithmic relationship with the air kerma incident on the detector. It is recommended to read between 1700 and 1900 for all examinations. EI: This is inversely proportional to the air kerma incident on the image receptor. It is derived from a characteristic pixel value of the image, i.e. a pixel value that corresponds to the average detector signal representing the target area of the examination. Exposure index (EXI): exposed field is divided into 3X3 Matrix, where the central segment is the ROI. EXI is calculated as the average out the original pixel values in the central segment. EXI value is directly proportional to dose. Doubling of EXI value represents a doubling of absorbed dose at image receptor.
ARE EXPOSURE INDEX VALUES CONSISTENT IN CLINICAL PRACTICE? Table 2. Carestream CR System. Examination
EI range
Skull Chest Abdomen Hand
1630–1660 2560–2660 1820–1890 1940–1960
Examination
lGm range
Skull Chest Abdomen Hand
1.88–2.21 1.67–1.82 1.82–1.92 1.85–2.03
Table 4. Philips DR System. Examination
EI
Skull Chest Abdomen Hand
320 400 500 320
Table 5. Siemens DR System. Examination
EI
Skull Chest Abdomen Hand
487 475 309 583
DISCUSSION Overall, the results show that the EI value is not consistent in CR, and thoroughly consistent in DR. No change was noted in EI for any examination in either DR system. This suggests that only varying clinical conditions such as exposure factor manipulation, collimation, SID and patient size will change the value. Within the CR study, the largest fluctuations were seen in the Agfa system. It should be noted that an increase of 0.3 in the lGm value, indicates a doubling of dose to the detector. Therefore, in the skull examination, the fluctuations seen would indicate to the clinician that double the exposure was given then is necessary according to manufacturers guidelines. This may have an impact on resultant patient dose if
373
Downloaded from http://rpd.oxfordjournals.org/ at Norges Teknisk-Naturvitenskapelige Universitet on November 17, 2014
Table 3. Agfa CR System.
the clinician is using the lGm value as their predominant method of feedback for radiation dose. Similarly, in the Carestream system, an increase of 300 EI indicates a doubling of exposure to the plate. In the chest examination, the EI value fluctuated by 100, which would suggest an increase or decrease of one-third detector exposure. Again, this may have a clinical impact when the clinician is using the EI value to assess dose. To avoid this potential issue further education and awareness of these issues is therefore paramount. There was no evidence in this particular work to suggest that fluctuations may be examination dependant. In the Agfa system, the least fluctuation was seen in the abdomen examination and in the Carestream system it was in the hand examination. However, further investigation is necessary to clarify this issue. Processing delay may explain some fluctuations, as there is a decrease in the latent image after a certain time(10). While every imaging plate was processed within 2 min of exposure, there is the possibility that some decrease could occur in plates that would be idle longer. Further work is needed to quantify this, as it would have a significant impact for ward, neonatal and theatre radiography. One limitation of this study is that the imaging plate readers were not used at various times to investigate whether the EI consistency alters during busier periods. Also, the variability between different readers could also be assessed for any impact (10). One must query what the tolerance for variability is, and how acceptable this is. It is stated by certain companies that the EI values are accurate to within +20 %, and other agencies by +10 %(11,12). Whilst both CR systems predominantly remained within these limits, it is difficult to assess this within the logarithmic scale that is the EI value for both Carestream and Agfa. The authors feel it is unacceptable to indicate a doubling of detector dose to the clinician, with no change in dose noted. Recent work by the group has confirmed that clinicians are using this value in clinical practice, and many believe it can be used as a dose-reducing tool. Anecdotal evidence suggests that some quality assurance programs are based around the EI values. Departments may alter radiographic exposures to achieve optimum EI values to ensure optimal image quality and minimum radiation exposure. However, it is clear from this study that inconsistencies are obvious, outside of clinical technique, and caution is warranted when relying on this value. In conclusion, given the evidence presented in this study, it is imperative that clinicians are aware of the fluctuations that are occurring in the EI value in CR systems outside of clinical conditions on which they have control. The challenge is posed to the clinician to be able to assess the image quality outside of the
M. L. BUTLER ET AL.
EI value to ensure optimal clinical practice. The reliability of EI values as a feedback mechanism for CR is also uncertain. REFERENCES
374
Downloaded from http://rpd.oxfordjournals.org/ at Norges Teknisk-Naturvitenskapelige Universitet on November 17, 2014
1. Mettler, F. A. Medical radiation exposure in the U.S. in 2006: preliminary results. Health Phys. 95(5), 502–507 (2008). 2. Berrington de Gonzalez, A. and Darby, S. Risk of cancer from diagnostic x-rays: estimates from the UK and 14 other countries. Lancet 363, 345–351 (2004). 3. Willis, C. Computed radiography: a higher dose? Pediatr. Radiol. 32, 745 –750 (2002). 4. Peters, S. E. and Brennan, P. B. Digital radiography: are the manufacturers’ settings too high? Optimization of the Kodak digital radiography system with aid of the computed radiography dose index. Eur. Radiol. 12, 2381–2387 (2002). 5. Willis, C. E. Strategies for dose reduction in ordinary radiographic examinations using CR and DR. Pediatr. Radiol 34(Suppl. 3), 196–200 (2004).
6. Shepard, S. J. et al. Recommended exposure indicator for digital radiography. AAPM Report of Task Group No. 116. American Association of Physicists in Medicine (2008). 7. Stephen Amis, E. et al. American College of Radiology White Paper on Radiation Dose in Medicine. J. Am. Coll. Radiol. 4, 272– 284 (2007). 8. Hart, D. and Wall, B. F. Radiation Exposure of the UK Population from Medical and Dental X-ray examinations (Chilton: National Radiological Protection Board) (2002). 9. Swallow, R. A., Naylor, E., Roebuck, E. J. and Whitney, A. S. Clark’s Positioning in Radiography, eleventh edn. (London: Heinemann Professional Publishing Ltd) (1986). 10. Fauber, T. L., Legg, J. S. and Quinn, M. Variation in CR imaging plate readers. Radiol. Tech. 74, 15–23 (2002). 11. FCR Quality Assurance Program. Stamford (Conn: FUJIFILM Medical Systems USA Inc.) (2000). 12. Siebert, J. et al. Acceptance testing and quality control of photostimulable phosphor imaging systems. Report of Task Group 10. American Association of Physicists in Medicine (1998).
