New Exposure Indicators for Digital Radiography Simplified for Radiologists and Technologists

Transcription

1 Medical Physics and Informatics Technical Innovation Don et al. New Simplified Exposure Indicators Medical Physics and Informatics Technical Innovation Steven Don 1 ruce R. Whiting 2 Lois Jo Rutz 3 ruce K. pgar 4 Don S, Whiting R, Rutz LJ, pgar K Keywords: computed radiography, digital radiography, dose, exposure index DOI: /JR Received February 2, 2012; accepted after revision March 23, Electronic Radiology Laboratory, Mallinckrodt Institute of Radiology, St. Louis Children s Hospital, Washington University School of Medicine, 510 S Kingshighway, St. Louis, MO ddress correspondence to S. Don (dons@mir.wustl.edu). 2 Department of Radiology, University of Pittsburgh, Pittsburgh, P. 3 Gammex, Middleton, WI. 4 gfa HealthCare, Greenville, SC. JR 2012; 199: X/12/ merican Roentgen Ray Society New Exposure Indicators for Digital Radiography Simplified for Radiologists and Technologists OJECTIVE. The purpose of this article is to educate radiologists and technologists about the clinically relevant portion of the new digital radiography standards. CONCLUSION. oth the International Electrotechnical Commission (IEC standard ) and the merican ssociation of Physicists in Medicine (PM Task Group 116) have developed similar standards for monitoring exposure in digital radiography to eliminate proprietary and confusing terminology. Radiologists and technologists will need to learn three new terms exposure index, target exposure index, and deviation index to understand the new standards. F ilm-screen radiography has a direct relationship between exposure and image density (Fig. 1), giving immediate visual feedback regarding over- or underexposure. Quality control is straightforward: Checking the waste bin reveals the number of underexposed or overexposed examinations. Digital radiography overcomes the dynamic range limitations of film-screen radiography because the receptor responds linearly to exposure, with a dynamic range on the order of 100 times that of film-screen radiography [1]. Image processing will adjust the image and still produce acceptable gray scale (Fig. 1), but the trade-off is lack of visual feedback regarding exposure. ll manufacturers systems measure the receptor exposure. Fuji coined the term S number, mimicking film-screen radiography speed, to measure receptor exposure in computed radiography in the mid 1980s [2], but eventually other manufacturers developed their own proprietary reporting terminology (Table 1). In the Fuji system, the exposure indicator is inversely related to receptor exposure, whereas Carestream and gfa systems have a direct logarithmic relationship. Other manufacturers introduced additional indicators. In a department that uses only one manufacturer s systems, the radiologists and technologists can learn a single set of parameters, but in departments with multiple manufacturers systems, it can be difficult to remember all the exposure values that denote a correct exposure range. Digital radiography image processing can compensate for under- and overexposure and still produce ideal gray scale. Underexposed images appear noisy (quantum mottle), whereas overexposed images contain negligible noise. Radiologists prefer images without noise, so there is a tendency over time to increase exposure (i.e., dose creep ) [3]. dditionally, there are currently no common quality-control programs to monitor exposure indicators and other factors for documentation of proper technique for each examination. Significant challenges for radiologists and technologists to maintain optimal image quality (i.e., diagnostic-quality images) at a reasonably low patient exposure include the following: manufacturer-specific exposure indicators, preference for noiseless images, and lack of immediate visual feedback for under- or overexposure. New Standards Developed The need for terminology standardization to eliminate confusion has been expressed by many organizations. The conference on the as low as reasonably achievable (LR) concept in pediatric computed radiography and digital radiography, which was held in Houston, TX, in 2004, urged the standardization of exposure indicators, with exposure feedback [3]. Van Metter and Yorkston [4, 5] proposed a framework for a universal receptor exposure measure in 2005 [4], which they subsequently tested in 2006 [5]. Their elements of an ideal metric include a clearly defined and calibrated system that is independent of the manufactur- JR:199, December

2 Don et al. TLE 1: Selected Manufacturer Exposure Values Manufacturer Value Symbol Units Formula Fuji S value S 200/mR Carestream Exposure index EI Millibels (mbel) [1000 log(mr)] gfa Log of median value lgm els log(mr) a Note mr = milliroentgens. a For a speed-class system of 200. er and the technology. It must be robust, consistent, and simple. The International Electrotechnical Commission (IEC) and the merican ssociation of Physicists in Medicine (PM) have been working separately on exposure value standardization. oth efforts involve collaboration among physicists, manufacturers, and the Medical Imaging and Technology lliance. IEC standard was published in 2008 [6], and the report of PM Task Group 116 was published in 2009 [7]. Even though these standards are not legislatively mandated, some manufacturers have already adopted them and others will likely follow. Radiologists and technologists must become familiar with these standards to optimally use new digital radiography equipment. (For simplicity, the IEC terminology will be used here and the differences clarified in the next section.) There are three important aspects of the new standards from a user s viewpoint: exposure index (EI), target exposure index (EI T ), and deviation index (DI). ecause this is a new standard, there are limited published articles available discussing the terminology and its use [8, 9]. EI is a measure of radiation in the relevant image region on the receptor. It is a function of the region of interest (ROI) selected by the digital radiography workstation for the examination type, image processing, and exposure used. Normally, the ROI is determined automatically through an analysis of the image. Image analysis techniques vary by manufacturer, but their overall goal is to select a clinically appropriate ROI for proper image processing. If the ROI selection is done incorrectly, either by the system or operator intervention, then the EI will be incorrect (Fig. 2). EI is calibrated using specified beam conditions [6] and is linearly related to receptor exposure; doubling the ms setting doubles the EI (Fig. 3). It is a relative exposure measure within a type of examination and not a patient dose indicator. EI is dependent on the beam spectrum (Fig. 3). Thus, one must be careful when comparing two examinations using different kvp settings or types of examinations. EI T is the reference exposure obtained when an image is optimally exposed. It may be set by the manufacturer or by the local imaging center and is dependent on the body part, view, procedure, and imaging receptor. DI quantifies how much the actual EI varies from the EI T, defined by the following formula: DI = 10log EI EI T In an ideal situation, where EI and EI T are the same, DI will be zero. DI value of ± 1.0 corresponds to one ms step on a typical x-ray generator console [7]. DI values of 1.0 and 3.0 correspond to 26% and 100% overexposure, respectively. Conversely, DI values of 1.0 and 3.0 correspond to 20% and 50% underexposure, respectively. The DI value gives immediate feedback to the technologist about the adequacy of the exposure. The PM has made TLE 2: Deviation Index and Use With Clinical Images initial recommendations on the interpretation of DI for clinical use [7] (Table 2 and Figs. 4 7). These recommendations may be used as starting points, but each facility should determine their own action points based on their clinical needs and department requirements. Differences etween IEC and PM Standards From the perspective of the radiologist and the technologist, there are few differences between the IEC and PM standards. However, the following are significant differences between the respective standards: PM reports the EI in terms of microgray, whereas IEC uses a unitless measure that multiples the PM value by 100; PM reports the DI with one significant digit of precision (e.g., 1.3), whereas IEC does not specify precision. PM has pledged to work toward adoption of the IEC definitions to create a universal standard. Exposure Index and Patient Dose EI is a measure of radiation exposure on the image receptor. It is not a measure of patient dose. There are many factors that need to be known to estimate patient dose [10] (Table 3 and Figs. 7 and 8). Some radiography units record generator technique factors in the DICOM header. y the coupling of the Deviation Index Exposure ction > 3 > 2 overexposure Report to management, repeat if image burned out Overexposure Repeat if image burned out 0.5 to 0.5 Target range 1.0 to 3.0 Underexposed Consult radiologist for repeat < 3.0 Underexposed Repeat Note Modified from Shepard et al. [7]. Recommendations are subject to revision by the merican ssociation of Physicists in Medicine (PM). Contact PM for the current recommendations. TLE 3: Factors Needed for Estimation of Patient or Organ Dose Group Technique Factors Patient factors Organ factors Factor Entrance x-ray beam quality kvp and total filtration Use of a grid Distance of patient from source Positioning (e.g., anteroposterior, posteroanterior, or lateral) Entrance skin exposure ge (if pediatric patient) Target organ whole body or specific organ (e.g., thyroid or breast) rea of the entrance beam covering the organ of interest Depth of the organ of interest Thickness of non soft-tissue structures overlaying the organ of interest ackscatter factor, which is a function of the irradiated area 1338 JR:199, December 2012

3 New Simplified Exposure Indicators technique factors with a dose-area product meter, dose could be estimated. Discussion The new standard terminology (EI, EI T, and DI) is helpful in eliminating proprietary terms, thus reducing confusion for radiologists and technologists. DI gives immediate feedback to the radiologist and technologist about the adequacy of the technique for each image. Recommendations for corrective action when the technologist notes a DI value that is too high or too low are suggested (e.g., reviewing the examination with the radiologist) (Table 2). ecause these are new standards, there are factors that need to be addressed to optimize patient imaging. First, an objective EI T for common examinations needs to be established on the basis of image quality metrics and not just subjective values set by a manufacturer or a local imaging center. Second, quality assurance programs are needed, which use exported measures that are recorded in DICOM structured reports and can be input into the Integrating the Healthcare Enterprise Radiation Exposure Monitoring profile [11]. Thus, not only can individual examinations with too much or too little exposure be Fig. 1 Comparison of film-screen and digital radiography systems and exposure feedback. (Reprinted with permission from Seibert J, Morin RL. The standardized exposure index for digital radiography: an opportunity for optimization of radiation dose to the pediatric population. Pediatr Radiol 2011; 41: ), Traditional film-screen systems use overall film density as exposure indicator, giving direct feedback to technologist regarding exposure. s technique increases, image progressively darkens. Note that although Hunter and Driffield curve is fixed, histogram shifts to right., Digital radiography uses image processing to adjust gray scale with signal. Direct visual cues (dark vs light) are lost regarding exposure. Note that as histogram shifts to right, image processing shifts as well. identified but, by monitoring over time, systematic trends can also be identified and corrected. The data from the Integrating the Healthcare Enterprise Radiation Exposure Monitoring profile can be used to establish national benchmarks. The merican College of Radiology has established the CT Dose Index Registry to benchmark CT; a similar program could be undertaken for digital radiography. cknowledgments We thank Michael Flynn and J. nthony Seibert for reviewing the manuscript. References 1. Willis CE. Computed radiography: a higher dose? Pediatr Radiol 2002; 32: Sonoda M, Takano M, Miyahara J, Kato H. Computed radiography utilizing scanning laser stimulated luminescence. Radiology 1983; 148: Willis CE, Slovis TL. The LR concept in pediatric CR and DR: dose reduction in pediatric radiographic exams a white paper conference executive summary. Pediatr Radiol 2004; 34(suppl 3):S162 S Van Metter R, Yorkston J. Toward a universal definition of speed for digitally acquired projection images. In: Medical imaging 2005: physics of medical imaging. San Diego, C: SPIE, 2005: Van Metter R, Yorkston J. pplying a proposed definition for receptor dose to digital projection images. In: Medical imaging 2006: physics of medical imaging. San Diego, C: SPIE, 2006: International Electrotechnical Commission (IEC). Medical electrical equipment: exposure index of digital x-ray imaging systems. Part 1. Definitions and requirements for general radiography. International standard IEC Geneva, Switzerland: IEC, Shepard SJ, Wang J, Flynn M, et al. n exposure indicator for digital radiography: PM Task Group 116 executive summary. Med Phys 2009; 36: Cohen MD, Markowitz R, Hill J, Huda W, abyn P, pgar. Quality assurance: a comparison study of radiographic exposure for neonatal chest radiographs at 4 academic hospitals. Pediatr Radiol 2012; 42: Cohen MD. Quality assurance: potential use for the newly described exposure index in clinical practice. J m Coll Radiol 2010; 7: McCollough CH, Schueler. Calculation of effective dose. Med Phys 2000; 27: O Donnell K. Radiation exposure monitoring: a new IHE profile. Pediatr Radiol 2011; 41: JR:199, December

4 Don et al. Fig. 2 Images show effect of modifying relevant image region. Target exposure index (EI T ) for adult chest radiograph is currently set at 300 for DXG CR system (gfa)., In this example, relative image region for image processing and exposure index (EI) calculation was done automatically. Calculated EI is 264, slightly lower than EI T of 300, yet well within normal variation with deviation index (DI) of 0.6., Red regions of interest (ROIs) over lungs were selected by user and EI increased to 375 with DI of 1.0. Selecting only lungs eliminated some of soft-tissue density and raised EI by manual processing, but because region selected was appropriate for diagnostic task, EI was within acceptable range. In example of green ROI in abdomen, which is incorrectly selected region for chest radiography, EI is 64 and DI is 6.7. ecause area selected was significantly lower in exposure than lungs, there is dramatic change in EI and DI. In example of yellow ROI in free-in-air exposure, which also is incorrectly selected region for chest radiography, EI is 1924 and DI is 8.1; yellow ROI selected was outside diagnostic area and has very little attenuation. Exposure Index kvp mgy, EI ~1000 (RQ-5) 90 kvp mgy, EI ~ kvp mgy, EI ~ kvp mgy, EI ~ Microgray Fig. 3 Graph shows calibration of exposure index (EI) for specific spectrum (dark blue line). HVL = half-value layer; RQ-5 = calibrated radiation quality with aluminum filtration standard. Variations from this condition will change exposure index. Note that EI has linear relationship with receptor exposure. Fig. 4 Images show effects of varying ms on exposure index (EI) and deviation index (DI). Neonatal Chest Phantom (Gammex) was used as test object. Images were obtained on DXG CR system (gfa) using manufacturer s exposure-monitoring quality assurance software with visual feedback. Target exposure index is 450., Radiograph exposure of 60 kvp and 1 ms. EI is 479 and DI is 0.3, well within accepted range. Color bar is green, indicating diagnostic-quality image., Radiograph exposure of 60 kvp and 2.5 ms. EI increased to 1258, and DI increased to 4.5, indicating higher than acceptable exposure. Color bar is yellow, indicating image flagged for review. C, Radiograph exposure of 60 kvp and 0.25 ms. EI decreased to 102, and DI decreased to 6.4. Noise is visible. Color bar is red, indicating repeat examination may be needed pending radiologist s review. C 1340 JR:199, December 2012

5 New Simplified Exposure Indicators Fig. 7 Clinical image of left forearm using Ysio digital radiography system (Siemens Healthcare). Exposure index was 1780 and target exposure index was set at 400; deviation index was 6.5. lthough image was overexposed, there was no saturation and no need to repeat. Thus, patient received more exposure than necessary, but image is visually acceptable with no indication of overexposure. Fig. 5 Clinical image of pediatric chest using Ysio digital radiography system (Siemens Healthcare). Exposure index was 102 and target exposure index was set at 250; even though deviation index was 3.8, image was of diagnostic quality and examination not repeated. Hence, it is important not to rely solely on deviation index but to review image with radiologist before repeating imaging. Fig. 6 Clinical image of soft-tissue neck examination using Ysio digital radiography system (Siemens Healthcare). Exposure index was 12 and target exposure index was set at 250; deviation index was Noise is excessive and image was repeated. Such low exposure could be explained by use of automatic exposure control for child who was too small for and not centered over chamber. Fig. 8 Clinical image of pediatric chest using Ysio digital radiography system (Siemens Healthcare). Exposure index (EI) was 301 with target exposure index set at 250; deviation index (DI) was 0.8. In image, grid lines are present, which should not be used with this age and thickness of patient. dditionally, collimation could have been tighter and eliminated exposure to left extremity (ignoring positioning). Thus, patient dose was excessive, even though EI and DI were in acceptable range. JR:199, December