This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by PATEL, I Articles by KIRBY, M C PubMed PubMed Citation Articles by PATEL, I Articles by KIRBY, M C

The use of computed radiography for routine linear accelerator and simulator quality control

¹ North Western Medical Physics, Radiotherapy Department, Rosemere Cancer Centre, Lancashire Teaching Hospitals NHS Foundation Trust, Preston, ² Radiotherapy Physics Group, Department of Medical Physics and Clinical Engineering, Oxford Radcliffe Hospitals NHS Trust, Oxford, and ³ North Western Medical Physics, The Christie NHS Foundation Trust, Manchester, UK

Correspondence: I Patel, North Western Medical Physics, Radiotherapy Department, Rosemere Cancer Centre, Lancashire Teaching Hospitals NHS Foundation Trust, Preston, UK. E-mail: imran.patel{at}lthtr.nhs.uk

	Abstract

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
Computed radiography (CR) systems were originally developed for the purpose of clinical imaging, and there has been much work published on its effectiveness as a film replacement for this end. However, there has been little published on its use for routine linear accelerator and simulator quality control, and therefore we have evaluated the use of the Kodak 2000RT system with large Agfa CR plates as a replacement for film for this function. A prerequisite for any such use is a detailed understanding of the system behaviour, hence characteristics such as spatial uniformity of response, reproducibility of spatial accuracy, plate signal decay with time and the dose–response of plates were investigated. Finally, a comparison of results obtained using CR for the measurement of radiation field dimensions was made against those from radiographic film, and found to be in agreement within 0.1 mm (mean difference for high-resolution images, 0.3 mm root mean square difference) for megavoltage images and 0.3 mm (maximum difference) for simulator images. In conclusion, the CR system has been shown to be a good alternative to radiographic film for routine quality control of linear accelerators and simulators.

	Introduction

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
A major recent initiative in the UK has been the initiation and development of a national programme for integrating IT systems within the national Health Service (NHS) [1]. This programme began in 2002 and had a number of objectives, some of which were:

• to provide national standards for connectivity;
  
• to improve and enhance data quality;
  
• to have common and seamless methods of data interchange;
  
• to rationalise IT services across the NHS.
  

Part of the programme included the roll-out of picture archiving and communications systems (PACS) throughout all hospital trusts in England and Wales [1], replacing film-based systems with, wherever possible, full digital (both direct and computed) radiographic imaging for all departments, including radiotherapy.

In March 2007, as part of that PACS implementation, Rosemere Cancer Centre opted for a computed radiography (CR) solution, combined with a DICOM image store, to replace the film processor. This would provide an immediate digital imaging solution to clinical imaging for the image intensifier-based simulator and for those linear accelerators (linacs) without amorphous silicon (aSi) electronic portal imaging devices (EPIDs). However, to have a common non-film-based system for radiotherapy equipment quality assurance purposes would be more challenging. Although there is some information in the literature about the use of CR, and its effectiveness as a film replacement, for clinical imaging (including portal imaging) [2–9] and even for some dosimetric and quality control (QC) purposes [10, 11], there is little information published about its general use as a replacement for film for simple external beam radiotherapy quality control.

An alternative approach could have been to use self-developing (GafChromic) film and/or amorphous silicon EPIDs [12–17]. Although self-developing films have become more sensitive and less expensive in recent years and are a viable alternative to X-Omat V film [13], they are still a single-use consumable medium with associated cost and storage implications. EPIDs are indeed the way forward for linear accelerator QC [16], being a readily available high-quality dosimetric device, but only when installed on all treatment machines and with easily accessible software solutions.

This paper consolidates and elaborates on work done to examine CR as a routine QC replacement for film in a radiotherapy department [18]; our goal has been to examine how we could maximise the benefit from an impending filmless environment, without having to resort to a separate film processor or other film-based methods for physics use. Many of the field shapes and sizes used for both commissioning and regular QC of linear accelerators and simulators are large [14, 19] and so here we have examined the suitability of 35 x 43 cm² CR phosphor plates for a variety of QC purposes. The integration of digital imaging data into the workflow for QC (in particular its analysis, interpretation, display and storage) is not without its own challenges, and these are the subject of further work [20].

	Methods and materials

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
System description
The Kodak 2000RT CR system (Carestream Health, Rochester, NY) together with large (35 x 43 cm²) Agfa CR plates (model MD10; Agfa Healthcare, Mortsel, Belgium) were used in this study. The Kodak system consists of a phosphor plate reader with an integrated eraser, and is able to scan plates up to 35 x 43 cm². The use of this system, and the principles behind its operation and those of storage phosphors for diagnostic and megavoltage photon beams, are well described in the literature [10, 15, 21–23]. For completeness, a brief description is given here.

The active layer of the Agfa CR plate is a coating of BaSrFBr doped with Eu in a +2 ionic state. During exposure, the energy deposited in the plate produces a latent image consisting of trapped electron–hole pairs in the phosphor grains. Unlike some diagnostic CR systems, which use rigid plates, the Agfa MD10 plates are flexible, like film. The scanner feeds the CR plate using rollers under a red laser beam (658 nm) that scans the width of the plate (Figure 1). As the laser scans across the plate, the trapped electrons are stimulated, and a significant fraction returns to the lowest energy level within the phosphor, with a simultaneous release of light. The intensity of the light is proportional to the amount of radiation dose absorbed locally. Although not shown in Figure 1, the path of the plate through the reader is curved, so that plates are scanned whilst vertical, but are then turned through 90° into a horizontal output tray.

View larger version (20K): [in this window] [in a new window]   Figure 1. A schematic of the readout and light collection mechanism of the Kodak 2000RT CR system (courtesy of Carestream Health, Rochester, NY; http://www.carestreamhealth.com/radiation-oncology-beam-dosimetry-package-for-the-2000rt-cr-plus.html). PMT, photomultiplier tube.

CR data acquisition
All measurements were carried out using 6 MV X-rays (Elekta Precise Linac; Elekta Ltd, Crawley, UK) for megavoltage work using three large (35 x 43 cm²) Agfa CR plates, as these would encompass the fields normally used for multileaf collimator (MLC) QC (such as those shown in Figure 2) without modification. In accordance with the manufacturer's recommendations [24] and other published work [10, 25], the exposure of the CR plate to ambient light was minimised during irradiation, after irradiation and also during readout. This was achieved by storing, irradiating and transporting the plates in empty X-Omat V envelopes. The envelopes and plates were marked (plates on the non-phosphor side) so that a consistent and reproducible positioning of the plates was possible. The long axis of the plates was always maintained in the gun–target (longitudinal) direction with respect to the linac or simulator, and the end of the plate nearest the gantry was always fed into the scanner first for readout. The plates were placed snugly inside the X-Omat V envelope which, in turn, had positional markings so that the plate could be positioned for irradiation with a maximum variation of approximately 0.3 cm.

View larger version (87K): [in this window] [in a new window]   Figure 2. Four examples of large field images used for linear accelerator commissioning and routine quality controltesting at Rosemere Cancer Centre. (a) "Double-H" field used for examining multileaf collimator (MLC) leaf bank scaling and centring. (b) Field shape used to examine "MLC alignment". (c) An example of the "Bayouth test". (d) An example of the multisegment image called a "rocket" film.

For kilovoltage images acquired on the treatment simulator (Elekta Precise Simulator, Elekta Ltd, Crawley, UK), the CR plates were used as they would be clinically, i.e. in standard Kodak ECL cassettes (Kodak, Rochester, NY) in the film holder above the image intensifier. The same precautions with ambient light (as above) were used, but no particular post-irradiation time was set. All plates were scanned using the kilovoltage scanning mode within the Kodak Radiation Oncology Systems (KROS) clinical software (v6.0).

Reproducibility of spatial accuracy
This was determined for megavoltage images using a 0.5 cm thick Perspex sheet containing a 2-D array of ball bearings (2 mm diameter with 2.5 cm spacing) placed on a CR plate at a 100 cm source-to-plate distance. To minimise any rotation in the image and therefore any discrepancies in expected horizontal and vertical distances, the rows of ball bearings were aligned with the edges of the plate. An additional 1.0 cm of Perspex build-up was added and 40 monitor units (MUs) were delivered with a 40 x 40 cm² field centred on the plate. The plate was scanned at a 1024 pixel resolution. Intra-session reproducibility was tested by repeating the process 10 times in one session. The test was carried out on 3 separate plates in 13 sessions over a period of 10 weeks to assess intersession consistency of the spatial accuracy. Distances between ball bearings in the horizontal and vertical direction were measured on each image using the RIT113 software, and compared with expected values.

Spatial uniformity of response
The geometry of the PMTs and the laser within the reader induce a significant non-uniformity to the plate readout, which has been noted for the smaller CR plates (24 x 30 cm²) using the Kodak system [10]. The RIT113 software has a uniformity correction algorithm that utilises images acquired at an extended focus-to-surface distance (FSD) (220 cm) at 10 cm deep in a uniform water-equivalent phantom for a range of doses. As our work focused on images acquired within the isocentric plane, we used an isocentric configuration to investigate spatial non-uniformity. A 40 x 40 cm² field was used, with the plate placed at the isocentre and using 1.5 cm of Perspex build-up. Doses ranging from 0.1–2.5 Gy were delivered to each plate, with each of the three plates used on a specific linac.

Profiles were taken from these images to examine the uniformity of response in the vertical and horizontal directions. The RIT113 uniformity correction routine was evaluated by generating a uniformity correction matrix with a routine within the software and using the above images. The correction matrix was then applied to the same images. Profiles were taken from the corrected images and compared with those from the uncorrected images.

Signal decay with time
As the image formed following irradiation is in a metastable state, there is signal decay with a delay time between irradiation and readout. This change in signal after irradiation (fading) can affect the results of measurements of radiation field size, depending upon the nature of the signal decay. To investigate this effect, all three plates were irradiated isocentrically (1.5 cm Perspex build-up) with 20 x 20 cm² fields and 100 MU exposures, each on a specific linac. All three linacs were calibrated, as standard, to deliver 1 Gy per 100 MU to D_max (depth of maximum dose) for a 10 x 10 cm² field at 100 cm FSD. The linac calibrations were all within ±2% of nominal for these experiments. The plates were then read at the following times: 5 min, 12 min, 24 min, 36 min, 48 min and 60 min from the end of the exposure. No changes in field size were made between each exposure.

No uniformity correction was applied to the images but they were dose–response corrected. Each image was analysed, using the RIT113 software, by measuring the mean SU value within a 1 x 1 cm² region of interest (ROI) centred on the irradiated field. The decay rate was determined in a similar manner to that by Olch [10] by calculating the loss of SU per minute of delay time relative to the SU reading obtained from the first image, which in our case was that after a 5 min delay. Additionally, horizontal and vertical profiles were measured from each image to investigate whether the measured field size was affected by the SU decay.

Dose–response of plates
The dosimetric response of the large CR plates was measured for doses ranging from 0.1 Gy to 2.5 Gy. Two different methods were used for producing standard dose–response curves. The first was to centre a 10 x 10 cm² field on a CR plate set up at 100 cm from the source. Perspex build-up (1.5 cm) was added on top of the CR plate. For each dose delivered, the mean SU value within a 1 x 1 cm² ROI centred on the irradiated field was recorded. No uniformity correction was applied to the images. The second method used was to irradiate each quadrant of the CR plate with a 10 x 10 cm² field to different doses. All other set-up parameters were kept the same as above. For the second method, the midpoint of the total delivery time for the four beams was used as T₀, and uniformity correction was applied to these images.

Comparison with film for typical QC fields (linac)
For evaluating the use of large CR plates for routine MLC QC, simultaneous isocentric irradiations were made. Each CR plate was set up isocentrically, with X-Omat V ready-pack film (Kodak, Rochester, NY) placed in contact with it (source side) and 1.5 cm of Perspex build-up above the film. The Perspex sheet included a radio-opaque marker used to mark the central axis (crosswire) of the linac. All experiments were conducted using 6 MV X-rays and 40 MU exposures.

Several different QC fields were investigated (see Figure 2). These included: (i) a "double H" field which is used at commissioning and following major MLC changes to check the scaling and centring of the MLC leaf banks [26]; (ii) an "MLC alignment" field, where leaves at either end of the leaf banks, one from each opposite bank, are drawn out to full extension to check that the sides of the leaves are parallel to the conventional jaws [26]; (iii) a "Bayouth test" field used routinely as an MLC leaf sequence technique for examining interleaf offset and absolute position across a range of movements [27]; (iv) a "rocket" field, used routinely as a multisegment exposure to check absolute calibration of MLC leaves, backup collimators and orthogonal collimators [19]; and (v) a series of regularly shaped fields used to examine the light/X-ray field congruence for different collimator rotations.

The radiation field edges on the irradiated films were identified using a calibrated transmission densitometer (Model DT1405; Alrad Instruments, Berkshire, UK) with an aperture size of 1 mm in diameter. A 50% optical density (OD) was used to identify the leaf ends and Y diaphragms, and a 60% OD for the X backup diaphragms. The 60% OD level for the latter had previously been considered, locally, to be equivalent to the 55% dose level required for the Elekta MLC backup diaphragms [26]. The CR images were calibrated for dose using the RIT113 software, and profiles were extracted and exported to Microsoft Excel for analysis using the appropriate spatial scaling factors determined. In general, a uniformity correction was not applied to the analysis of the CR images. However, for one of each of the QC fields investigated, an analysis was conducted both with and without uniformity correction to investigate the effects.

Comparison with film for typical QC fields (simulator)
For evaluating the use of large CR plates for routine use on the simulator, a standard weekly check exposure was acquired. For the weekly checks, the collimating blades were opened up to approximately 25 x 25 cm², and the field-defining wires set to a symmetric field size of 8 x 8 cm². Identical sequential exposures (82 kV, 16 mAs) were made for the CR plates and standard Kodak T-Mat S/RA film (Kodak, Rochester, NY), without changing any simulator set-up parameters between the two. Image scaling was determined from an in-built radio-opaque reticule in the head of the simulator (Figure 3), which had been checked separately at commissioning and through routine QC. For the films, analysis was performed using a light box and ruler. For the CR images, images were scanned into the clinical KROS software in kilovoltage mode with a resolution of 2048 pixels. The distance measurement facilities within the Kodak software were used for analysis.

View larger version (78K): [in this window] [in a new window]   Figure 3. An example of an exposure used for assessing computed radiography for routine weekly simulator quality control.

	Results

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

View larger version (22K): [in this window] [in a new window]   Figure 4. Spatial uniformity of response profiles in (a) horizontal (scanning laser spot) and (b) vertical (plate feed) directions. These plots are taken isocentrically for 40 x 40 cm2 fields, 1.5 cm deep Perspex, and 40 MU (monitor unit) exposures using 6 MV X-rays. Three different plates have been used, each irradiated on a different linac. Also shown is the effect of using the RIT113 uniformity correction software for the profiles obtained with plate P3.

View larger version (9K): [in this window] [in a new window]   Figure 5. A graph showing the response of the large computed radiography plates to changes in dose. The response (in terms of raw scanner unit) for three different phosphor plates is illustrated, showing a distinct change in gradient of response at 1.5 Gy.

View larger version (32K): [in this window] [in a new window]   Figure 6. Histogram plot of spatial differences found when analysing line profiles from computed radiography (CR) images and densitometer-based measurements on film. Negative differences indicate that CR distances are shorter than film.

	Discussion

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
Reproducibility of spatial accuracy
Arguably, the spatial accuracy tests are difficult for any system that has a mechanical scanning mechanism, especially when performing measurements over the full length and width of the large plates. The scaling is determined by the motion of the scanning laser spot in the horizontal direction, and the linear motion of the plate fed through the rollers in the vertical direction; therefore, one might expect to see differences in scaling reproducibility between each. Additionally, although there is a nominal scaling factor, the system does not use this consistently, which may be regarded as a flaw, but instead detects the field of view from the scanned image and applies the 1024 (or 2048) pixels across this field of view. In this way, the pixel scaling changes are dependent upon whether large or small plates are used, in order to improve the spatial resolution for the smaller plates. For the purpose of spatial accuracy for our QC work, it is not the absolute scaling factor that is important but its reproducibility and consistency with time. If it is constant, a set pixel size can be applied in much the same way as planar digital imaging systems, and distances can be measured simply using a digital ruler [19]. If it is not constant, then a fixed scaling pattern must be used with each image, from which the scaling is calculated. This has most often been used clinically in the past with film systems where the source-to-film distance can be variable or is not known. A second requirement, especially with scanned or rolling systems, is that the scaling must be constant across the length and width of each scan/rolling direction. If not, gross errors in distance measurements can occur. For example, for an image obtained with a clinical reticule, one might assess the scaling factor from the reticule in the central part of the image. Anatomical measurements then made in a different part of the image, where the image has become "stretched" or "compressed" through the rolling mechanism, would be incorrect because the local scaling factor was different.

The tolerance quoted by the manufacturer is 1%, which is similar to that of flatbed [13] or standard film scanners (e.g. the Vidar Dosimetry Pro, Herndon, VA) [10]. For the Kodak CR system, this spatial accuracy can be assessed using a standard small (30 x 24 cm²) scaling phantom with expected 20 x 20 cm² markings scanned through the system [10, 24]. Provided that there are no scaling variations from one part of the image to the next, this tolerance is entirely acceptable for clinical use. However, for QC purposes (as with other systems [13]), it is not. For example, for a field size of 20 x 20 cm², the maximum allowable field for the treatment machine itself is 20.2 x 20.2 cm², with 0.1 cm for each collimator jaw. If the jaws were actually out of tolerance (e.g. 20.3 cm), then a 1% limit on the CR scanner might actually show that the collimators were within tolerance. From this example, one requires a tolerance of less than 0.5%. Our initial results showed that the Kodak 2000RT CR system could actually achieve this and remain consistent over long (>10 weeks) periods of time.

Unfortunately, after this initial period, we experienced problems with both the laser units and the roller mechanisms within the CR reader. The results of our spatial accuracy tests following these instances are also shown in Table 1. Some spot results were also beyond the quoted tolerance of 1%. Additionally, we experienced problems with the consistency of scaling in different parts of large plate images, for both QC and clinical work, where parts of the image would become stretched compared with others. This made the unit unacceptable at times for all types of imaging work. The plates have to be flexible to follow a curved path through the reader. Also, one may expect the plates to be more vulnerable to wear and tear when used for QC, as they are used away from the rigid protective environment of a clinical film cassette. However, we found that the plates could not stretch mechanically in any way, even after months of use, and so did not affect the scaling through any lack of physical integrity. The scaling problems have now all been resolved, and the unit has been operating consistently well for the past 10 months. However, as a result of these occurrences, special measures have been put in place for the purpose of quality assurance for the CR reader itself. For example, a daily run-up procedure has been implemented, whereby in addition to performing quick self-tests [24] an image is also acquired on the treatment simulator. This has the in-built reticule present (shown in Figure 3) and a quick visual check is made of this to look for any areas of inconsistency in scaling. For linear accelerator QC work, a special radiographic template has been designed, a schematic of which is shown in Figure 7. This consists of a 40 x 40 x 1 cm³ sheet of Perspex, inlaid (half-sunk) into which are a series of 2 mm diameter steel ball bearings. The pattern is designed so that it (i) is unique in its placement on an image, i.e. its correct orientation is easily identifiable; and (ii) places ball bearings in positions for all our QC procedures, except for the Bayouth test. The ball bearings have a constant fixed spatial separation from the larger marker at the centre, which is lined up with the central axis of the linac. Analysis of the QC images includes semi-automatic procedures integrated into our systems of work so that results are calculated for all field edges with respect to the central axis but use scaling factors that are spatially derived from the RIT113 profiles around each appropriate radiation field segment being measured [20]. These locally applied factors ensure an additional finer tolerance for spatial measurements than is achievable from the Kodak reader alone. It also helps to eliminate the effect of any local variations in spatial consistency across the large plate images.

View larger version (4K): [in this window] [in a new window]   Figure 7. A schematic of a special radiographic template designed for most linear accelerator quality control imaging using the computed radiography system.

Similar results were obtained at other exposure levels, with only slight changes in the profiles, consistent with the changes in linac beam profiles seen at different exposures, especially in the gun–target direction — the plane in which the electron beam bending occurs for the Elekta linacs. The change in uniformity of response with exposure implied that a single exposure correction matrix would not be sufficient, and this was indeed found to be the case.

Signal decay with time
The X-ray absorption mechanism in a photostimulable phosphor plate is similar to that of a conventional rare earth phosphor, such as that used in conventional film cassettes or indirect flat panel imagers, except that the light emitted is not prompt. A latent image forms consisting of trapped electron–hole pairs, which release the useful optical photon output only when the phosphor is optically stimulated — a process known as photostimulated luminescence. However, as the process involves a metastable state, there will always be a natural release of the stored charge, without stimulation, between irradiation and readout of the phosphor — a process known as fading [21]. If CR images are to be used for QC purposes, especially when trying to detect a 50% edge for determining field sizes and leaf positions, then fading may affect this information depending upon the fading mechanism involved. Where an individual symmetrical object needs to be detected on an image, such as a field-defining wire, then the effect of fading is less critical, as the central or mean position of that object will not change even though the signal to background ratio will as fading occurs.

Our results showed that the chosen time for readout was appropriate, as the decay rate had stabilised at that point. Because the decay rate itself was found to be low, it was not too critical if readout was performed a few minutes either side of the chosen delay period. There is a greater variation in results between plates than that found by Olch [10], but this is also within 0.5% of the SU signals obtained at our exposures. For a significant effect on field size or field edge determination, this would have to be between 10% and 20%. The consequent consistency of field edge results is demonstrated in the analysis of the 20 x 20 cm² field sizes used for examining the signal decay rates; these are shown in Table 4. The images have been corrected for dose–response, but not for uniformity. They show that the reported field size changes by less than 0.1 mm for a delay period of twice that used for routine QC. Therefore, a recommended delay period of 10 min is perfectly acceptable for QC purposes, and it is not critical if readout is performed a few minutes either side of this recommendation. Interestingly, the results also highlighted a "bug" within the Kodak dosimetry software. The data clearly show that the actual field sizes measured are considerably smaller than the set 20 x 20 cm² fields. This is because the scaling factor passed from the CR image file headers into the RIT113 software is that used for the smaller plates, and not the larger ones. Local scaling factors are being used in our work within the RIT113 software and in our Excel spreadsheets until the problem is resolved by the manufacturer.

View larger version (74K): [in this window] [in a new window]   Figure 8. An example of a computed radiography image obtained for the multisegment "rocket" film used for routine quality control, but which has been accidentally exposed to room light (post-irradiation) diagonally across the plate.

Comparison with film for typical QC fields (linac)
For the purpose of replacing film for QC in radiotherapy, the two most crucial investigations are perhaps (i) the examination of the spatial integrity of the CR system and its consistency, and (ii) the accuracy of field edge determination compared with current film methods. Our results in Figure 6 show that the methods that we have used here with the CR system give acceptable results. The mean and RMS differences shown in Table 3 may be considered as systematic and random errors in the measurements, respectively. These results are similar to other studies comparing point-based film analysis and scanned analysis of other detection media [13]. As the tolerance for individual collimator jaws or leaves is 1 mm or 1% for larger field sizes [14], then the CR system and our analysis methods are suitable for this purpose. Although fewer comparison points were used, slightly more accurate results were obtained using the higher (2048 pixel) scanning resolution. As the volume of QC images is small compared with that of clinical portal and simulator images, the decision was made to use a 2048 pixel resolution for all QC work. Storage of these larger image files has not been a significant issue so far.

Several CR images were also analysed after uniformity correction using the RIT113 software. The uniformity correction map was derived from images taken for the same plate as described earlier. The distances measured from the uniformity-corrected images differed by no more than 0.2 mm (maximum random error) from those measured from the same images with no uniformity correction applied. One may conclude, therefore, that it is not necessary to apply a uniformity correction for our routine QC fields. This has the added practical benefit of shortening the overall time taken for the analysis process.

As described earlier for linac QC, the CR plates are used outside of the protective environment of a film cassette, and therefore may be prone to cracking etc., especially as they are also continually flexed when passing through the reader. However, they have now been in use at our centre for over 18 months for routine QC of linacs and have been found to be robust with no signs of cracking or flaking in the phosphor coating or mechanical stretching in the backing material. This is, no doubt, partly the result of the careful and protective handling procedures that we have employed for their routine use.

Comparison with film for typical QC fields (simulator)
Most images acquired for routine QC on the simulator require the user to identify the position of field-defining wires and/or their distance with respect to a reference point, usually the central beam axis. The CR characteristics of dose–response, uniformity and fading are much less significant for this purpose, as it is individual radiographic objects rather than greyscale gradients or edges that need to be identified. Reproducibility of spatial accuracy is more important, although all of our simulator images have an in-built reticule, as illustrated in Figure 3, from which scaling factors may be determined when interrogating the images using digital rulers. Highly significant, however, is the integrity of spatial accuracy in different parts of the same image. As described earlier, any process within the readout system that causes image stretching or compression in various sections of the image may cause significant measurement errors when a single scaling factor is used, or when scaling factors are derived from other parts of the image which are not stretched or compressed. For this reason, each CR simulator image was examined along each major and minor axis using the reticule as a guide for consistency. No inconsistencies or variations were found. The comparative results we have described between hardcopy film and digital CR images were perfectly adequate. However, these results were obtained in the initial commissioning period of the CR system, i.e. before we began experiencing problems with the scanning laser and roller mechanism. Once the scanner problems were resolved, the system was again acceptable for QC checks on the simulator.

	Conclusions

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 
The use of the Kodak 2000RT system with large Agfa CR plates as a replacement for film for routine linear accelerator and simulator QC has been evaluated. We have found that the CR images may be used successfully for a wide range of QC checks for megavoltage and kilovoltage X-ray fields. The system is easy to use, giving acceptable results even without uniformity correction. Data are easily imported and analysed in commercial dosimetry software such as RIT113, and our use of this software for commissioning the CR system has been found to be essential for megavoltage work. Simple and accurate results can be made with the Kodak 2000RT CR system for QC tasks, making it acceptable as the primary tool for field analysis and/or an ideal complement to the use of aSi EPIDs for quality control.

	Acknowledgments

 
The authors are grateful to Michelle Donowsky (Carestream Health Worldwide), Phil Brennan (Carestream Health UK), William Hewitt (Carestream Health UK) and Phil Ferrar (Oncology Imaging Systems Ltd, UK) for their support and assistance during the problems experienced with the unit. We would also like to thank all the physicists and engineers at the Rosemere Cancer Centre for accepting the radical change in practice to routine quality control.

Received for publication September 29, 2008. Revision received December 4, 2008. Accepted for publication December 9, 2008.

	References

Top Abstract Introduction Methods and materials Results Discussion Conclusions References

 

1. The Department of Health. Delivering 21st century IT support for the NHS: national strategic programme. London, UK: Department of Health, 2002.
2. Wilenzick RM, Merritt CR, Balter S. Megavoltage portal films using computed radiographic imaging with photostimulable phosphors. Med Phys 1987;14:389–92.[CrossRef][Medline]
3. Gur D, Deutsch M, Fuhrman C R, Clayton PA, Weiser JC, Rosenthal MS, et al. The use of storage phosphors for portal imaging in radiation therapy: therapists' perception of image quality. Med Phys 1989;16:132–6.[CrossRef][Medline]
4. Barnea G, Navon E, Ginzburg A, Politch J, Roehrig H, Dick CE, et al. Use of storage phosphor imaging plates in portal imaging and high-energy radiography: the intensifying effect of metallic screens on the sensitivity. Med Phys 1991;18:432–8.[CrossRef][Medline]
5. Scheck RJ, Wendt T, Panzer M. Digital storage phosphor radiography for treatment verification in radiotherapy. Br J Radiol 1993;66:801–6.[Abstract/Free Full Text]
6. Fetterly KA, Hangiandreou NJ. Image quality evaluation of a desktop computed radiography system. Med Phys 2000;27:2669–79.[CrossRef][Medline]
7. Whittington R, Bloch P, Hutchinson D, Bjarngard BE. Verification of prostate treatment setup using computed radiography for portal imaging. J Appl Clin Med Phys 2002;3:88–96.[CrossRef][Medline]
8. Yamada S, Murase K. Effectiveness of flexible noise control image processing for digital portal images using computed radiography. Br J Radiol 2005;78:519–27.[Abstract/Free Full Text]
9. Roberts N, Stanley S. An observer study: comparison of computed radiography to electronic and film portal imaging. J Radiother In Practice 2006;5:197–202.[CrossRef]
10. Olch A. Evaluation of a computed radiography system for megavoltage photon beam dosimetry. Med Phys 2005;32:2987–99.[CrossRef][Medline]
11. Conte L, Bianchi C, Cassani E, Monciardini M, Mordacchini C, Novario R, et al. The use of photostimulable phosphor systems for periodic quality assurance in radiotherapy. Phys Med 2008;24:34–7.[Medline]
12. Cheung T, Butson MJ, Yu PK. Measurement of high energy x-ray beam penumbra with Gafchromic EBT radiochromic film. Med Phys 2006;33:2912–14.[CrossRef][Medline]
13. Thomas MDR, Warrington AP. GafChromic RTQA film for routine quality assurance of high-energy photon beams. Phys Med Biol 2006;51:1439–47.[CrossRef][Medline]
14. IPEM (Institute of Physics and Engineering in Medicine). Physics aspects of quality control in radiotherapy. IPEM Report 81. York, UK: IPEM, 1999. Available from: http://www.ipem.ac.uk/ipem_public/default.asp?id=360 [Accessed 25 March 2009].
15. Kirby MC, Glendinning AG. Developments in electronic portal imaging systems. Br J Radiol 2006;79:S50–65.[Abstract/Free Full Text]
16. Budgell GJ, Zhang R, Mackay RI. Daily monitoring of linear accelerator beam parameters using an amorphous silicon EPID. Phys Med Biol 2007;52:1721–33.[CrossRef][Medline]
17. Mohammadi M, Bezak E. Evaluation of MLC leaf positioning using a scanning liquid ionization chamber EPID. Phys Med Biol 2007;52:N21–33.[CrossRef][Medline]
18. Patel I, Natarajan T, Kirby MC. The use of computed radiography for megavoltage linear accelerator quality control. Int J Radiat Oncol Biol Phys 2007;69:s728
19. IPEM (Institute of Physics and Engineering in Medicine). Acceptance testing and commissioning of linear accelerators. IPEM Report 94. York, UK: IPEM, 2006.
20. Patel I, Natarajan T, Hassan S, Kirby MC. Initial experience of the implementation and integration of computed radiography into a quality control system for external beam radiotherapy. Radiother Oncol 2007;84:s247–8.
21. Rowlands JA. The physics of computed radiography. Phys Med Biol 2002;47:R123–66.[CrossRef][Medline]
22. Seibert JA, Bogucki TM, Ciona T, Huda W, Karellas A, Mercier JR, et al. Acceptance testing and quality control of photostimulable storage phosphor imaging systems. AAPM Report 93. College Park, MD: AAPM, 2006.
23. Li HH, Gonzalez AL, Ji H, Duggan DM. Does response of BaFBRI:Eu2+ storage phosphor plates exposed to megavoltage photon beams. Med Phys 2007;34:103–11.[CrossRef][Medline]
24. Kodak Product Information. Kodak 2000RT CR System. Rochester, NY, USA: Carestream Health, 2002. Available from: http://www.carestreamhealth.com/radiation-oncology-beam-dosimetry-package-for-the-2000rt-cr-plus.html [Accessed 25 March 2009].
25. Ramamurthy R, Canning CF, Scheetz JP, Farman AG. Impact of ambient lighting intensity and duration on the signal-to-noise ratio of images from photostimulable phosphor plates processed using DenOptix and Scan X systems. Dentomaxillofac Radiol 2004;33:307–11.[Abstract/Free Full Text]
26. Jordan TJ, Williams PC. The design and performance characteristics of a multileaf collimator. Phys Med Biol 1994;39:231–51.[CrossRef][Medline]
27. Bayouth JE, Wendt D, Morrill SM. MLC quality assurance techniques for IMRT applications. Med Phys 2003;30:743–50.[CrossRef][Medline]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by PATEL, I Articles by KIRBY, M C PubMed PubMed Citation Articles by PATEL, I Articles by KIRBY, M C