- Email: [email protected]
Exposure Creep in Computed Radiography
Exposure Creep in Computed Radiography: A Longitudinal Study Dale J. Gibson, BAppSc, Robert A. Davidson, PhD, MAppSc(MI) Purpose: Exposure creep is the gradual increase in x-ray exposures over time that results in increased radiation dose to the patient. It has been theorized as being a phenomenon that results from the wide-exposure latitude of computed radiography (CR) and direct/indirect digital radiography (DR). This project evaluates radiographic exposures over 43 months to determine if exposure creep exists and if measures can be applied to halt or reverse exposure creep trends. Methods: Exposure indices were initially recorded over 29 months between August 2007 and December 2009 from the intensive and critical care unit (ICCU) and the emergency department (ED) departments where manual CR exposures were used. The data from this period were then assessed and the exposure indexes (EI) values from the radiographic images were compared to the radiology department criteria of EI values between 1400 to 1800 as being in the optimal exposure range. EI values below this were considered underexposed and over this as overexposed. An intervention was required to be used in ICCU and implemented in January 2010 to halt a noted trend of overexposure. The EI value for each chest x-ray (CXR) was recorded in the patients’ ICCU records and was to be used by radiologic technologists/radiographers in determine exposure factors in subsequent CXR. After the intervention, EI values were recorded and evaluated for an additional 15 months between February 2010 and March 2011. Results: Between August 2007 and December 2009, 17,678 ICCU CXR images and 69,327 ED x-ray examinations were evaluated for over- and underexposure. A trend was noted in ICCU that showed a significant increase (P = .023) in EI values from the beginning to the end of the evaluation. No such trend was seen in the ED EI values (P = .120). After the intervention in ICCU, the overexposure trend was halted. Conclusions: Exposure creep has been show to exist. It is surmised that this occurs where judgment to determine the correct radiographic exposure factors is needed when taking into account a large range of patient sizes. It has also been shown that providing radiologic technologists/radiographers with previous EI values for the same x-ray examination can halt a trend of exposure creep. Key Words: Exposure creep; computed radiography; digital radiography; exposure indices. ªAUR, 2012
C
omputed radiography (CR) and direct/indirect digital radiography (DR) have many benefits over film/screen (FS) radiography. One of the greatest advantages is the wider dynamic range or latitude that CR/ DR exhibits (1,2). Radiologic technologists/radiographers are able to produce diagnostic images using a greater range of exposure factors (3). This eliminates one of the primary reasons in FS radiography for repeating an examination: an overexposure to the image plate. The phenomenon of exposure creep is described in previous studies, though its existence has yet to be shown (4,5). Exposure creep is the gradual increase over time of the radiologic technologists/radiographers ‘‘usual’’ exposures Acad Radiol 2012; 19:458–462 From the Flinders Medical Centre, Bedford Park, South Australia (D.J.G.) and Charles Sturt University, School of Dentistry and Health Sciences, Boorooma Street, Wagga Wagga, New South Wales 2678, Australia (D.J.G., R.A.D.). Received August 3, 2011; accepted December 2, 2011. Address correspondence to: R.A.D. e-mail: [email protected] ªAUR, 2012 doi:10.1016/j.acra.2011.12.003
458
for a given radiographic anatomical projection. Exposure creep commenced with the introduction of digital radiographic techniques such as CR and DR. As CR and DR became more widespread, radiologic technologists/ radiographers have learned that overexposure in these imaging modalities actually improves image quality. In FS radiography, over- or underexposure resulted in an unacceptable image. When choosing radiographic exposure factors for CR and DR, particularly manual exposures, the tendency of some radiologic technologists/radiographers is to choose toward the high exposure levels because it is more likely to result in a better quality image. Lower radiographic exposure factors can result in noisy or lower quality images. CR and DR images that are over or underexposed image can still be displayed on a monitor with adequate brightness and contrast because of the large dynamic range of the digital detectors and the ability to manipulate window widths and levels. These higher radiographic exposure factors then become the norm. A second aspect of exposure creep can be attributed to the manufacturer’s recommendations for correct exposures. Every
Academic Radiology, Vol 19, No 4, April 2012
manufacturer has a system in place to provide a numerical indication of the exposure reaching the imaging plate and hence provide an indication as to whether the correct radiographic exposure was used. These exposure indices provide an indication of image quality as well as a measure of the correct exposure (6,7). The naming of exposure indices varies between manufacturers. Carestream (formerly Kodak) named theirs the Exposure Index (EI) value, FujiFilm the S value, and AgfaGeavert the log mean (lgM) value (8). Carestream’s recommendation is to achieve an EI range of 1500–1800. AgfaGeavert’s lgM is logarithmic formula and FujiFilm is an inverse value in which greater exposure to the imaging plate results in a lower S value. Radiologic technologists/radiographers moving from one manufacturer to another could struggle with the different systems. When using Carestream’s system, an EI range of 1500–1800 is a generous range given that an increase of 300 is a doubling of exposure. A small increase in AgfaGeavert’s lgM is also a large dose increase. It has also been argued that these recommended levels may be set too high by the manufacturers in the first place (5). Using CR, significant artifacts are not experienced until the imaging plate is exposed 500 times more or 100 times less than the exposure used for FS (4,5). The manufacturers’ exposure indicators provide a guideline for radiology staff to make them aware of the dose delivered to the patient and the dose that reaches the imaging plate (IP). Film/screen radiography provides feedback in that a black film is overexposed and a white film is underexposed. Reviewing the exposure indicators in CR and DR is the only tool available to check that the correct exposure factors were used in the examination. It is therefore imperative that these exposure indicators influence a radiologic technologist/radiographer practice. Exposure creep can result when normal practice does not include a routine review of exposure indictors. At Flinders Medical Centre (FMC), a major public hospital in Adelaide, Australia, questions were raised over the trend for the radiographers to use higher than recommended exposures. It was decided to review the EI trends in two sections of the radiology department. The section selected, intensive and critical care unit (ICCU) is equipped with a Carestream CR 850 CR digitizer unit and the emergency department (ED) uses a Carestream CR975 Multiloader (Carestream Health Australia [formally Kodak Australia], Melbourne, Victoria). The same types of IPs are used in both areas. In ICCU, approximately 7000 chest radiographic (CXR) examinations are performed each year. The CXR examinations are performed using a dedicated mobile x-ray unit (GE AMX4, GE Healthcare, Sydney, Australia) that requires the setting of manual radiographic exposure factors. In the ED, more than 25,000 x-ray examinations on a wide range of anatomical areas are performed each year. The purpose of this study was to evaluate the existence of exposure creep in manual exposure radiographic examinations in the ICCU and in the CR room of the ED at FMC. If the evaluation showed the existence of exposure creep, an
EXPOSURE CREEP IN COMPUTED RADIOGRAPHY
intervention would be planned to determine if the exposure creep could be halted or reversed. METHODS Inclusion Criteria
Exposure creep is expected to occur where radiologic technologists/radiographers set manual, non-automatic control (AEC) radiographic exposures based on judgment of patient size. The two sections within the FMC Radiology Department where manual exposure setting predominate are the ICCU department—where CXR are performed with a GE AMX4—and the ED, which focuses on extremity radiography and bed-bound examinations that did not use AEC radiographic exposures. Other radiology sections where AEC are commonly used were excluded from this study. Data Collection
A longitudinal study of EI measurements was proposed to monitor the exposures used in the ICCU and ED departments. Only EI measurements from manual exposures were included. All CXR exposures in the ICCU are manually selected and examinations in the ED that were performed using the AEC were excluded. In all examinations, the standard departmental routines and protocols were used. All equipment used was calibrated and serviced at the beginning and throughout the study by the manufacturer’s technicians through the normal service schedules for each machine. The Kodak (Carestream) Total Quality Tool was used annually to measure EI accuracy and cassette screen performance. No other equipment variations were made during the study. Data were initially collected from August 2007 for more than 28 months. In January 2010, an intervention was put in place in the ICCU department and data continue to be measured both in the ED and the ICCU. Determination of Optimal, Under-, and Overexposures
Peters and Brennan (5) found that there is evidence that manufacturers may be recommending higher exposure indicators than is necessary. Previous discussion with staff of the CR manufacturer, Carestream, determined that an acceptable EI range for appropriate image quality was 1500 to 1800. After a review by radiologists and radiographers of CR image quality across the radiology department, a lower level EI of 1400 was deemed as providing acceptable diagnostic image quality. For this project, an EI range of 1400 to 1800 was therefore used as the optimal exposure range; radiographic images with an EI <1400 were considered underexposed and radiographic images with an EI >1800 were considered overexposed. Using this range with the department accepted low of 1400 and the manufacturer’s recommendation of 1800 as the high, 1600 was selected as the target EI for all planar radiographic examinations. 459
GIBSON AND DAVIDSON
Academic Radiology, Vol 19, No 4, April 2012
Figure 1. Plot of intensive and critical care unit chest x-ray indicating optimal, over-, and underexposed exposure indexes percentages between August 2007 and December 2009.
Intervention
After the evaluation of the ICCU data from August 2007 to December 2009, an intervention was proposed to determine if exposure creep had truly occurred. A method was decided to determine if a simple measure could halt or reverse the trend seen in the evaluation of the ICCU EI data. Before the intervention, radiographers recorded the radiographic factors of tube kilovoltage (kVp) and current (mAs) and the patient’s position, supine or erect, in the ICCU patient’s records. No comment was recorded on EI or image quality. In January 2010, it was decided to modify what radiographers recorded in the ICCU patients’ records and to include the EI value of the CXR performed as well as kVp, mAs, and position. Before subsequent CXRs on the same ICCU patient, radiographers were instructed to note the EI of the previous CXR in determining if modification of the radiographic exposure factors was needed to keep in the optimal exposure range. Approval for the study was gained from FMC Radiology Department. No ethical considerations arose from this project because only deindentified exposure data from ICCU and ED radiographic images were collected.
RESULTS Over the entire study period (August 2007–March 2011), the FMC Radiology Department undertook 488,073 radiologic examinations. During the 28 months from August 2007 to December 2009, 17,678 ICCU CXR images and 69,327 ED x-ray examinations were evaluated for over- and underexposure. EI measures for these images were recorded. Data were collated on a monthly basis with an average number of 610 ICCU and 2391 ED measurements per month. Over- and underexposure EI values were determined as an underexposure being an EI value <1400 and overexposure being an EI value >1800. An optimal exposure had an EI value between 1400 and 1800. The percentage of the total monthly ICCU CXR exams of optimal, over-, and underex460
posure EI values between August 2007 and December 2010 are shown in Figure 1. Linear regression lines are shown to highlight the trends in optimal, over-, and underexposure data. The percentage of the total monthly ED manual exposure exams of optimal, over-, and underexposure EI values from between 2007 and December 2010 are shown in Figure 2. Linear regression lines have also been displayed to show trends. On reviewing both the ICCU and ED data in January 2010, the ICCU data showed an increasing trend for overexposure. The intervention, discussed previously, was implemented. After the intervention in ICCU, data was collected for 14 months (February 2010–March 2011). The same criteria were used for determine optimal, under-, and overexposure as that used before the intervention. The total number of EI values for ICCU CXR was 6397. During October and November 2010, technical issues arose that affected the data collection; as a result, these months have been excluded from the data. The percentage of the total monthly ICCU CXR exams of optimal, over-, and underexposure EI values between February 2010 and March 2011 is shown in Figure 3. Linear regression lines have been included to show trend. Calculations of P values were made to confirm the significance of the changes from the beginning of the measurement period to the end of the measure period.
DISCUSSION Preintervention
Between August 2007 and January 2010, the EI for ICCU CXR showed a trend in which there was a gradual reduction in percentage of optimal exposures. The optimal percentage values, as determined from the regression analysis, have reduced from approximately 61% to 51% (Fig 1). The data show a significance decrease (P = .030) in optimal exposure percentages between the EI data at the beginning of the measurement period and the end of the measurement period.
Academic Radiology, Vol 19, No 4, April 2012
EXPOSURE CREEP IN COMPUTED RADIOGRAPHY
Figure 2. Plot of emergency department x-ray examinations indicating optimal, over-, and underexposed exposure index percentages between August 2007 and December 2009.
Figure 3. Plot of ICCU CXR indicating optimal, over-, and underexposed exposure index percentages between February 2010 and March 2011.
During the same period, there also has been a corresponding increase in overexposures of ICCU CXR. The overexposed percentage values, as determined from the regression analysis, increased from 25% to 38%. There is a significant increase (P = .023) in the percentage of CXR that were overexposed from the beginning of the evaluation period to the end of the evaluation. During the same measurement period, the EI from the ED showed no significant change for either optimal exposure (P = .115) or overexposure (P = .120). The trend lines seen in Figure 2 depict this lack of change. A trend was seen in the ICCU EI for an increase in the percentage of EI indicating an overexposure and a similar decrease in percentage of an overexposure. This trend was seen under conditions in which patient size could not be measured for infection control reasons. The trend of increasing EI was not seen in the ED. An influencing factor was considered in that generally only a single examination CXR is performed in ICCU, whereas a broad range of examinations are undertaken in the ED. It is also known that EI
values can be influenced by other factors other than exposure settings, namely source to image distance, collimation size, area of cassette used, time to process, and the algorithm chosen (3). Intensive care patients are usually difficult to position and have numerous attachments to consider. All images are processed in the CR digitizer using ‘portable chest’ algorithm. Examinations in the ED are more controlled, often have multiple images on the one IP, use a variety of algorithms and are more likely processed in a timely manner. Most of the manual exposure examinations performed in ED are of the extremities. These areas have a smaller variance in size from one patient to another, compared to that of the torso variations experienced in ICCU. It is therefore likely that radiographers reflect more accurately the exposure charts provided than they do in ICCU where the variation is greater. The combination of these factors is most likely to have caused the differences between ICCU and ED examinations. Although a trend of increasing overexposure is seen from the data, the determination of exposure creep, as discussed by Warren-Forward et al (4), cannot be proved from these 461
GIBSON AND DAVIDSON
data alone. To determine the existence of exposure creep, a halting or reversal of the trend following a change of radiographer behavior would prove its existence. Postintervention
The trend of an increase in overexposure in the ICCU was seen in late 2009. After a meeting of senior radiographic staff, it was decided to implement an intervention. The intervention required radiographers to record the EI value of the CXR taken into the patients’ records, along with the previously included radiographic factors of kVp, mAs, and the patient’s position. Radiographers were aware of the optimal EI range. On undertaking a later CXR on the same patient, the radiographer could then use the EI of the previous CXR as a guide to whether the listed radiographic factors were within the optimal range and then adjust the radiographic factors before the next CXR exposure if needed. The EI values collected from the ICCU between February 2010 and March 2011 show a decline of the trend of an increase in overexposure that was seen before the intervention. There has been a marginal increase in the percentage of optimal exposure (P = .042); however, there has been no significant decrease in percentage change of overexposure (P = .081) from the start of the evaluation period to the end of current evaluation. The trend for an increasing percentage of overexposed and for a decreasing percentage of optimally exposed CXR in ICCU between August 2007 and December 2009 has been halted, but not reversed. This halting of the trends shows that before the intervention, exposure creep existed in ICCU. Although this evaluation of EI values and trend does not show a cause of the exposure creep, it is suggested that the exposure creep in ICCU can be attributed to radiographers having to guess at the correct exposure factor in an anatomical area that exhibits a large range of sizes. This is supported by the fact that when a second CXR examination was performed on the same patient, a previous indication of optimal or over exposure from the recorded EI value was available in determining the next set of radiographic exposure factors.
462
Academic Radiology, Vol 19, No 4, April 2012
CONCLUSION Exposure creep when using CR has been show to exist. It is surmised that this occurs where judgment to determine the correct radiographic exposure is needed to take into account a large range of patient sizes. This study also demonstrates the importance of monitoring the behaviors of radiographers and their relative perceptions of ‘‘normal’’ exposures. By having a system in place that radiographers can refer to during clinical practice, radiographers can cross check their own tendencies toward exposure settings. It has also been shown that providing radiographers with additional information, in this case an indication of a correct exposure on the same patient can halt a trend of exposure creep. ACKNOWLEDGMENT The authors would like to thank Ms Liza Ricote of Flinders Medical Centre for her help in the project; Professor Patrick Brennan of The University of Sydney for his assistance and advice and to the radiographic staff of Flinders Medical Centre. REFERENCES 1. Walsh C, Gorman D, Byrne P, et al. Quality assurance of computed and digital radiography systems. Radiat Prot Dosimetry 2008; 129:271–275. 2. Cowen AR, Davies AG, Kengyelics SM. Advances in computed radiography systems and their physical imaging characteristics. Clin Radiol 2007; 62: 1132–1141. 3. Davidson R, Sim J. Computed radiography and dosimetry: some practical tips for dose optimization procedures. J Med Imag Radiat Sci 2008; 39: 109–114. 4. Warren-Forward HM, Arthur L, Hobson L, et al. An assessment of exposure indices in computed radiography for the posterior-anterior chest and the lateral lumbar spine. BJR 2007; 80:26–31. 5. Peters SE, Brennan PC. Digital radiography: are the manufacturers’ settings too high? Optimisation of the Kodak digital radiography system with aid of the computed radiography dose index. Eur Radiol 2002; 12:2381–2387. 6. Tsalafoutas IA, Blastaris GA, Moutsatsos AS, et al. Correlation of image quality with exposure index and processing protocol in a computed radiography system. Radiat Prot Dosimetry 2008; 130:162–171. 7. Brindhaban A, Al Khalifah K, Al Wathiqi G, et al. Effect of x-ray tube potential on image quality and patient dose for lumbar spine computed radiography examinations. Australas Phys Eng Sci Med 2005; 28:216–222. 8. Butler ML, Rainford L, Last J, et al . Are exposure index values consistent in clinical practice? A multi-manufacturer investigation. Radiat Prot Dosimetry 2010; 139:371–374.
C
omputed radiography (CR) and direct/indirect digital radiography (DR) have many benefits over film/screen (FS) radiography. One of the greatest advantages is the wider dynamic range or latitude that CR/ DR exhibits (1,2). Radiologic technologists/radiographers are able to produce diagnostic images using a greater range of exposure factors (3). This eliminates one of the primary reasons in FS radiography for repeating an examination: an overexposure to the image plate. The phenomenon of exposure creep is described in previous studies, though its existence has yet to be shown (4,5). Exposure creep is the gradual increase over time of the radiologic technologists/radiographers ‘‘usual’’ exposures Acad Radiol 2012; 19:458–462 From the Flinders Medical Centre, Bedford Park, South Australia (D.J.G.) and Charles Sturt University, School of Dentistry and Health Sciences, Boorooma Street, Wagga Wagga, New South Wales 2678, Australia (D.J.G., R.A.D.). Received August 3, 2011; accepted December 2, 2011. Address correspondence to: R.A.D. e-mail: [email protected] ªAUR, 2012 doi:10.1016/j.acra.2011.12.003
458
for a given radiographic anatomical projection. Exposure creep commenced with the introduction of digital radiographic techniques such as CR and DR. As CR and DR became more widespread, radiologic technologists/ radiographers have learned that overexposure in these imaging modalities actually improves image quality. In FS radiography, over- or underexposure resulted in an unacceptable image. When choosing radiographic exposure factors for CR and DR, particularly manual exposures, the tendency of some radiologic technologists/radiographers is to choose toward the high exposure levels because it is more likely to result in a better quality image. Lower radiographic exposure factors can result in noisy or lower quality images. CR and DR images that are over or underexposed image can still be displayed on a monitor with adequate brightness and contrast because of the large dynamic range of the digital detectors and the ability to manipulate window widths and levels. These higher radiographic exposure factors then become the norm. A second aspect of exposure creep can be attributed to the manufacturer’s recommendations for correct exposures. Every
Academic Radiology, Vol 19, No 4, April 2012
manufacturer has a system in place to provide a numerical indication of the exposure reaching the imaging plate and hence provide an indication as to whether the correct radiographic exposure was used. These exposure indices provide an indication of image quality as well as a measure of the correct exposure (6,7). The naming of exposure indices varies between manufacturers. Carestream (formerly Kodak) named theirs the Exposure Index (EI) value, FujiFilm the S value, and AgfaGeavert the log mean (lgM) value (8). Carestream’s recommendation is to achieve an EI range of 1500–1800. AgfaGeavert’s lgM is logarithmic formula and FujiFilm is an inverse value in which greater exposure to the imaging plate results in a lower S value. Radiologic technologists/radiographers moving from one manufacturer to another could struggle with the different systems. When using Carestream’s system, an EI range of 1500–1800 is a generous range given that an increase of 300 is a doubling of exposure. A small increase in AgfaGeavert’s lgM is also a large dose increase. It has also been argued that these recommended levels may be set too high by the manufacturers in the first place (5). Using CR, significant artifacts are not experienced until the imaging plate is exposed 500 times more or 100 times less than the exposure used for FS (4,5). The manufacturers’ exposure indicators provide a guideline for radiology staff to make them aware of the dose delivered to the patient and the dose that reaches the imaging plate (IP). Film/screen radiography provides feedback in that a black film is overexposed and a white film is underexposed. Reviewing the exposure indicators in CR and DR is the only tool available to check that the correct exposure factors were used in the examination. It is therefore imperative that these exposure indicators influence a radiologic technologist/radiographer practice. Exposure creep can result when normal practice does not include a routine review of exposure indictors. At Flinders Medical Centre (FMC), a major public hospital in Adelaide, Australia, questions were raised over the trend for the radiographers to use higher than recommended exposures. It was decided to review the EI trends in two sections of the radiology department. The section selected, intensive and critical care unit (ICCU) is equipped with a Carestream CR 850 CR digitizer unit and the emergency department (ED) uses a Carestream CR975 Multiloader (Carestream Health Australia [formally Kodak Australia], Melbourne, Victoria). The same types of IPs are used in both areas. In ICCU, approximately 7000 chest radiographic (CXR) examinations are performed each year. The CXR examinations are performed using a dedicated mobile x-ray unit (GE AMX4, GE Healthcare, Sydney, Australia) that requires the setting of manual radiographic exposure factors. In the ED, more than 25,000 x-ray examinations on a wide range of anatomical areas are performed each year. The purpose of this study was to evaluate the existence of exposure creep in manual exposure radiographic examinations in the ICCU and in the CR room of the ED at FMC. If the evaluation showed the existence of exposure creep, an
EXPOSURE CREEP IN COMPUTED RADIOGRAPHY
intervention would be planned to determine if the exposure creep could be halted or reversed. METHODS Inclusion Criteria
Exposure creep is expected to occur where radiologic technologists/radiographers set manual, non-automatic control (AEC) radiographic exposures based on judgment of patient size. The two sections within the FMC Radiology Department where manual exposure setting predominate are the ICCU department—where CXR are performed with a GE AMX4—and the ED, which focuses on extremity radiography and bed-bound examinations that did not use AEC radiographic exposures. Other radiology sections where AEC are commonly used were excluded from this study. Data Collection
A longitudinal study of EI measurements was proposed to monitor the exposures used in the ICCU and ED departments. Only EI measurements from manual exposures were included. All CXR exposures in the ICCU are manually selected and examinations in the ED that were performed using the AEC were excluded. In all examinations, the standard departmental routines and protocols were used. All equipment used was calibrated and serviced at the beginning and throughout the study by the manufacturer’s technicians through the normal service schedules for each machine. The Kodak (Carestream) Total Quality Tool was used annually to measure EI accuracy and cassette screen performance. No other equipment variations were made during the study. Data were initially collected from August 2007 for more than 28 months. In January 2010, an intervention was put in place in the ICCU department and data continue to be measured both in the ED and the ICCU. Determination of Optimal, Under-, and Overexposures
Peters and Brennan (5) found that there is evidence that manufacturers may be recommending higher exposure indicators than is necessary. Previous discussion with staff of the CR manufacturer, Carestream, determined that an acceptable EI range for appropriate image quality was 1500 to 1800. After a review by radiologists and radiographers of CR image quality across the radiology department, a lower level EI of 1400 was deemed as providing acceptable diagnostic image quality. For this project, an EI range of 1400 to 1800 was therefore used as the optimal exposure range; radiographic images with an EI <1400 were considered underexposed and radiographic images with an EI >1800 were considered overexposed. Using this range with the department accepted low of 1400 and the manufacturer’s recommendation of 1800 as the high, 1600 was selected as the target EI for all planar radiographic examinations. 459
GIBSON AND DAVIDSON
Academic Radiology, Vol 19, No 4, April 2012
Figure 1. Plot of intensive and critical care unit chest x-ray indicating optimal, over-, and underexposed exposure indexes percentages between August 2007 and December 2009.
Intervention
After the evaluation of the ICCU data from August 2007 to December 2009, an intervention was proposed to determine if exposure creep had truly occurred. A method was decided to determine if a simple measure could halt or reverse the trend seen in the evaluation of the ICCU EI data. Before the intervention, radiographers recorded the radiographic factors of tube kilovoltage (kVp) and current (mAs) and the patient’s position, supine or erect, in the ICCU patient’s records. No comment was recorded on EI or image quality. In January 2010, it was decided to modify what radiographers recorded in the ICCU patients’ records and to include the EI value of the CXR performed as well as kVp, mAs, and position. Before subsequent CXRs on the same ICCU patient, radiographers were instructed to note the EI of the previous CXR in determining if modification of the radiographic exposure factors was needed to keep in the optimal exposure range. Approval for the study was gained from FMC Radiology Department. No ethical considerations arose from this project because only deindentified exposure data from ICCU and ED radiographic images were collected.
RESULTS Over the entire study period (August 2007–March 2011), the FMC Radiology Department undertook 488,073 radiologic examinations. During the 28 months from August 2007 to December 2009, 17,678 ICCU CXR images and 69,327 ED x-ray examinations were evaluated for over- and underexposure. EI measures for these images were recorded. Data were collated on a monthly basis with an average number of 610 ICCU and 2391 ED measurements per month. Over- and underexposure EI values were determined as an underexposure being an EI value <1400 and overexposure being an EI value >1800. An optimal exposure had an EI value between 1400 and 1800. The percentage of the total monthly ICCU CXR exams of optimal, over-, and underex460
posure EI values between August 2007 and December 2010 are shown in Figure 1. Linear regression lines are shown to highlight the trends in optimal, over-, and underexposure data. The percentage of the total monthly ED manual exposure exams of optimal, over-, and underexposure EI values from between 2007 and December 2010 are shown in Figure 2. Linear regression lines have also been displayed to show trends. On reviewing both the ICCU and ED data in January 2010, the ICCU data showed an increasing trend for overexposure. The intervention, discussed previously, was implemented. After the intervention in ICCU, data was collected for 14 months (February 2010–March 2011). The same criteria were used for determine optimal, under-, and overexposure as that used before the intervention. The total number of EI values for ICCU CXR was 6397. During October and November 2010, technical issues arose that affected the data collection; as a result, these months have been excluded from the data. The percentage of the total monthly ICCU CXR exams of optimal, over-, and underexposure EI values between February 2010 and March 2011 is shown in Figure 3. Linear regression lines have been included to show trend. Calculations of P values were made to confirm the significance of the changes from the beginning of the measurement period to the end of the measure period.
DISCUSSION Preintervention
Between August 2007 and January 2010, the EI for ICCU CXR showed a trend in which there was a gradual reduction in percentage of optimal exposures. The optimal percentage values, as determined from the regression analysis, have reduced from approximately 61% to 51% (Fig 1). The data show a significance decrease (P = .030) in optimal exposure percentages between the EI data at the beginning of the measurement period and the end of the measurement period.
Academic Radiology, Vol 19, No 4, April 2012
EXPOSURE CREEP IN COMPUTED RADIOGRAPHY
Figure 2. Plot of emergency department x-ray examinations indicating optimal, over-, and underexposed exposure index percentages between August 2007 and December 2009.
Figure 3. Plot of ICCU CXR indicating optimal, over-, and underexposed exposure index percentages between February 2010 and March 2011.
During the same period, there also has been a corresponding increase in overexposures of ICCU CXR. The overexposed percentage values, as determined from the regression analysis, increased from 25% to 38%. There is a significant increase (P = .023) in the percentage of CXR that were overexposed from the beginning of the evaluation period to the end of the evaluation. During the same measurement period, the EI from the ED showed no significant change for either optimal exposure (P = .115) or overexposure (P = .120). The trend lines seen in Figure 2 depict this lack of change. A trend was seen in the ICCU EI for an increase in the percentage of EI indicating an overexposure and a similar decrease in percentage of an overexposure. This trend was seen under conditions in which patient size could not be measured for infection control reasons. The trend of increasing EI was not seen in the ED. An influencing factor was considered in that generally only a single examination CXR is performed in ICCU, whereas a broad range of examinations are undertaken in the ED. It is also known that EI
values can be influenced by other factors other than exposure settings, namely source to image distance, collimation size, area of cassette used, time to process, and the algorithm chosen (3). Intensive care patients are usually difficult to position and have numerous attachments to consider. All images are processed in the CR digitizer using ‘portable chest’ algorithm. Examinations in the ED are more controlled, often have multiple images on the one IP, use a variety of algorithms and are more likely processed in a timely manner. Most of the manual exposure examinations performed in ED are of the extremities. These areas have a smaller variance in size from one patient to another, compared to that of the torso variations experienced in ICCU. It is therefore likely that radiographers reflect more accurately the exposure charts provided than they do in ICCU where the variation is greater. The combination of these factors is most likely to have caused the differences between ICCU and ED examinations. Although a trend of increasing overexposure is seen from the data, the determination of exposure creep, as discussed by Warren-Forward et al (4), cannot be proved from these 461
GIBSON AND DAVIDSON
data alone. To determine the existence of exposure creep, a halting or reversal of the trend following a change of radiographer behavior would prove its existence. Postintervention
The trend of an increase in overexposure in the ICCU was seen in late 2009. After a meeting of senior radiographic staff, it was decided to implement an intervention. The intervention required radiographers to record the EI value of the CXR taken into the patients’ records, along with the previously included radiographic factors of kVp, mAs, and the patient’s position. Radiographers were aware of the optimal EI range. On undertaking a later CXR on the same patient, the radiographer could then use the EI of the previous CXR as a guide to whether the listed radiographic factors were within the optimal range and then adjust the radiographic factors before the next CXR exposure if needed. The EI values collected from the ICCU between February 2010 and March 2011 show a decline of the trend of an increase in overexposure that was seen before the intervention. There has been a marginal increase in the percentage of optimal exposure (P = .042); however, there has been no significant decrease in percentage change of overexposure (P = .081) from the start of the evaluation period to the end of current evaluation. The trend for an increasing percentage of overexposed and for a decreasing percentage of optimally exposed CXR in ICCU between August 2007 and December 2009 has been halted, but not reversed. This halting of the trends shows that before the intervention, exposure creep existed in ICCU. Although this evaluation of EI values and trend does not show a cause of the exposure creep, it is suggested that the exposure creep in ICCU can be attributed to radiographers having to guess at the correct exposure factor in an anatomical area that exhibits a large range of sizes. This is supported by the fact that when a second CXR examination was performed on the same patient, a previous indication of optimal or over exposure from the recorded EI value was available in determining the next set of radiographic exposure factors.
462
Academic Radiology, Vol 19, No 4, April 2012
CONCLUSION Exposure creep when using CR has been show to exist. It is surmised that this occurs where judgment to determine the correct radiographic exposure is needed to take into account a large range of patient sizes. This study also demonstrates the importance of monitoring the behaviors of radiographers and their relative perceptions of ‘‘normal’’ exposures. By having a system in place that radiographers can refer to during clinical practice, radiographers can cross check their own tendencies toward exposure settings. It has also been shown that providing radiographers with additional information, in this case an indication of a correct exposure on the same patient can halt a trend of exposure creep. ACKNOWLEDGMENT The authors would like to thank Ms Liza Ricote of Flinders Medical Centre for her help in the project; Professor Patrick Brennan of The University of Sydney for his assistance and advice and to the radiographic staff of Flinders Medical Centre. REFERENCES 1. Walsh C, Gorman D, Byrne P, et al. Quality assurance of computed and digital radiography systems. Radiat Prot Dosimetry 2008; 129:271–275. 2. Cowen AR, Davies AG, Kengyelics SM. Advances in computed radiography systems and their physical imaging characteristics. Clin Radiol 2007; 62: 1132–1141. 3. Davidson R, Sim J. Computed radiography and dosimetry: some practical tips for dose optimization procedures. J Med Imag Radiat Sci 2008; 39: 109–114. 4. Warren-Forward HM, Arthur L, Hobson L, et al. An assessment of exposure indices in computed radiography for the posterior-anterior chest and the lateral lumbar spine. BJR 2007; 80:26–31. 5. Peters SE, Brennan PC. Digital radiography: are the manufacturers’ settings too high? Optimisation of the Kodak digital radiography system with aid of the computed radiography dose index. Eur Radiol 2002; 12:2381–2387. 6. Tsalafoutas IA, Blastaris GA, Moutsatsos AS, et al. Correlation of image quality with exposure index and processing protocol in a computed radiography system. Radiat Prot Dosimetry 2008; 130:162–171. 7. Brindhaban A, Al Khalifah K, Al Wathiqi G, et al. Effect of x-ray tube potential on image quality and patient dose for lumbar spine computed radiography examinations. Australas Phys Eng Sci Med 2005; 28:216–222. 8. Butler ML, Rainford L, Last J, et al . Are exposure index values consistent in clinical practice? A multi-manufacturer investigation. Radiat Prot Dosimetry 2010; 139:371–374.