DOI: 10.1148/radiol.2321020397
(Radiology 2004;232:115-125.)
Nationwide Evaluation of X-ray Trends Survey of Abdomen and Lumbosacral Spine Radiography1
David C. Spelic, PhD,
Richard V. Kaczmarek, MS and
Orhan H. Suleiman, PhD
1 From the Center for Devices and Radiological Health, Division of Mammography Quality and Radiation Programs, U.S. Food and Drug Administration, 1350 Piccard Dr, HFZ-240, Rockville, MD 20850. Received April 3, 2002; revision requested June 12; final revision received January 22, 2004; accepted March 3. Address correspondence to D.C.S. (e-mail: david.spelic@fda.hhs.gov).
 |
ABSTRACT
|
|---|
Results of the 1995 Nationwide Evaluation of X-ray Trends (NEXT) survey of facilities that perform diagnostic radiographic examinations of the abdomen and lumbosacral spine were compared with those of previous NEXT surveys conducted in 1987 and 1989. A clinically validated radiographic phantom was used in the 1995 survey to capture data about radiation exposure and image quality. Additional data were obtained regarding clinical techniques, facility workloads, x-ray beam quality, film processing quality, and darkroom fog. Mean skin-entrance air kerma for the abdomen examination dropped from 3.2 mGy (in 1987) to 2.8 mGy at hospitals and from 3.4 mGy (in 1989) to 3.0 mGy at nonhospital facilities. Mean skin-entrance air kerma also decreased for the lumbosacral spine examination from 3.7 mGy (in 1987) to 3.3 mGy at hospitals and from 3.8 mGy (in 1989) to 3.2 mGy at nonhospital facilities. The quality of film processing improved, although 58 (18.3%) of 317 surveyed facilities did not meet the Mammography Quality Standards Act standard for film processing quality, compared with 185 (5.9%) of 3,120 mammography facilities inspected in 1995. Finally, 181 (58.0%) of 312 surveyed facilities had darkroom fog levels greater than the Mammography Quality Standards Act standard, compared with 1,426 (16.6%) of 8,605 mammography facilities inspected in 1995.
Index terms: Abdomen, radiography, 70.11 • Dosimetry • Radiations, exposure to patients and personnel • Radiations, measurement • Screens and films • Special Reports • Spine, radiography, 33.11
|
INTRODUCTION
|
|---|
The Nationwide Evaluation of X-ray Trends (NEXT) survey is a collaborative program of the U.S. Food and Drug Administration Center for Devices and Radiological Health, the Conference of Radiation Control Program Directors, and participating state radiation-control agencies. The goal of the NEXT surveys is to document the current practice of diagnostic radiology methods associated with radiation exposure and to observe trends that arise from changes in practice and advances in radiographic technology. Each year a particular radiologic examination is selected for survey, and radiation exposure data are collected from a nationally representative sample of U.S. clinical facilities. Since its beginning in 1973, the NEXT program has surveyed the practice of a variety of radiologic examinations nationwide, including adult chest radiography (1,2), computed tomography (3,4), radiography of the abdomen (5–7), radiography of the lumbosacral spine (7–9), mammography (10,11), dental radiography (12,13), and gastrointestinal fluoroscopy (14). The first-ever survey of pediatric chest radiography practice was conducted in 1998. Since 1984, NEXT surveyors have collected radiographic exposure data by using reference phantoms that have patient-equivalent x-ray attenuation to assess the effects of automatic exposure control devices, which have been incorporated with increasing frequency into imaging equipment. Data are collected about radiographic technique factors, skin-entrance radiation exposure to the patient, radiographic image quality, the quality of film processing, and darkroom fog. Surveys for a given examination are repeated periodically to enable the detection of trends in practice over time. For each survey, facilities are randomly selected from a compilation of state databases of facilities that are likely to conduct the selected radiographic examination. Approximately 45 participating states provide radiation control personnel to conduct the surveys. Staff members at the Center for Devices and Radiological Health then compile, analyze, and publish statistical summaries of the survey results. Tabular summaries of data from each NEXT survey are published and distributed by the Conference of Radiation Control Program Directors. Currently, the Conference of Radiation Control Program Directors, the Food and Drug Administration, and the American College of Radiology provide financial support for NEXT surveys.
In 1995, a NEXT survey was conducted among hospitals and nonhospital facilities that routinely performed radiographic examinations of the abdomen and lumbosacral spine. In this article, we report the 1995 survey results and compare them with the results of similar NEXT surveys that were conducted in 1987 at hospitals and in 1989 at nonhospital facilities.
|
SURVEY
|
|---|
Sample Selection
The Conference of Radiation Control Program Directors H-4 committee on NEXT oversees the planning phases of the survey and provides liaison with the state radiation offices, while the Food and Drug Administration is responsible for training surveyors and for technical features such as equipment, procedures, and facility selection. Surveyors were given a list of randomly selected facilities in their state to survey. The sample size for a particular state was based on its population of facilities likely to perform abdomen and lumbosacral spine examinations, as a percentage of all U.S. facilities. The procedure that was used for random sample selection has been previously described (6). Each state submitted a list of nonhospital facilities. A separate list of hospitals for each state was developed from data published by the American Hospital Association (15). The 1995 sample consisted of 443 facilities, of which 202 were hospitals and 241 were nonhospitals. Hospitals were defined as clinical facilities that provided for the overnight stay of patients; this category included infirmaries at educational institutions. Facilities that provided services on an outpatient basis, such as radiology practices, multiple-specialty practices, and chiropractic offices, were categorized as nonhospitals. A total of 204 abdomen examination surveys and 320 lumbosacral spine examination surveys were returned by facilities in 44 states. Substantially fewer surveys were returned for the abdomen examination, an indication that fewer facilities routinely performed this examination. (A facility was surveyed for a particular examination only if the examination was routinely performed there.) Follow-up was conducted if it was unclear why a facility was not surveyed for both examinations. Because some states did not survey their entire sample, the number of facilities visited (n = 347) was less than the number selected (n = 443) for the survey.
Radiographic Phantom
The radiographic phantom used for the 1995 survey is representative of a standard reference patient with a height of 1.7 m, mass of 74 kg, and abdomen with an anterior-to-posterior dimension of approximately 22 cm. The schematic diagram in Figure 1 shows the geometry of the phantom. Made of polymethyl methacrylate (density, approximately 1.18 g/cm3), the phantom has an anterior-to-posterior dimension of 16.92 cm and a central raised segment (an additional 1.92 cm of lucite and 0.46 cm of aluminum) that represents the spine. The raised spine-like portion allows the capture of radiation exposure measurements for the two types of diagnostic examination with use of a single phantom. The ionization chamber is positioned far enough from the phantom to allow measurements that are nearly free of backscatter, which is typically less than 2% of the in-air exposure measurement. The development, testing, and clinical validation of this radiographic phantom have been previously reported (16).
View larger version (26K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. Schematic diagram shows the geometry of radiation exposure measurement with the phantom during the NEXT surveys. The results of phantom validation studies (16) showed that an essentially backscatter-free measurement (typically with less than 2% backscatter) was achievable with this configuration. PMMA = polymethyl methacrylate.
|
|
An image quality test tool with two sets of objects was used to assess both low-contrast sensitivity and high-contrast limiting spatial resolution. One of the object sets, an aluminum disk containing a series of holes of different depths, permitted the estimation of limiting contrast sensitivity. The diameter of the holes was 0.8 cm, large enough that the ability to see them on a radiograph did not depend on the modulation transfer characteristics of the imaging system. The other, a set of rectangular mesh patterns made of copper wire, was used to determine high-contrast limiting spatial resolution. Figure 2 shows the dimensions and configuration of the test tool. The series of wire mesh patterns was used for estimation of spatial resolution because this set of objects enabled a straightforward scoring protocol and a comparison of results with those from other NEXT surveys, if desired. The test tool described here is identical to the image quality test tool used during the 1994 NEXT survey of adult chest radiographic examinations. The protocol for the surveys of abdomen and lumbosacral spine examinations conducted in 1987 and 1989 did not include an evaluation of image quality; hence, the 1995 survey findings will serve as the baseline for future comparisons.
View larger version (41K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2. Schematic diagram of the image quality test tool shows the specifications of the holes and mesh patterns. The holes are 8 mm in diameter, and the mesh patterns are approximately 15 x 15 mm. Values inside the squares indicate the mesh count.
|
|
Low-contrast sensitivity refers to the ability of a radiographic imaging system to depict small differences in attenuation in a specified region of interest. The minimum percentage of contrast depicted by the imaging system, hereafter referred to as percent contrast, is an indicator of the ability of the system to depict objects with low relative contrast, and it decreases as the performance of the system improves. The percent contrast associated with each test tool contrast object was estimated by computing the x-ray transmission for the combination of the phantom and each test tool object with respect to the object’s background.
Survey Procedure
Each surveyor was equipped with a radiation monitor (MDH 1015; Radcal, Monrovia, Calif) with a 6-cm3 ionization chamber to measure free-in-air exposures to the phantom. Depending on the examination surveyed, the placement of the test tool was either on top of the abdomen-like section of the phantom (abdomen examination) or on top of the raised spine-like section (lumbosacral spine examination). The distance between the x-ray source and the phantom surface was measured to allow determination of the air kerma at the entrance plane of the standard reference patient. Exposure and irradiation time were then measured by using the clinical technique factors typically used at the facility for radiography in a patient with characteristics like those of the standard reference patient. A radiograph was obtained by using the phantom and the image quality test tool, and the background optical density was measured. While at the facility, the surveyor also scored the quality of depiction of the test tool on the radiograph by reporting the number of copper mesh patterns and the number of contrast holes that were visible. Scores ranged from 0 (worst possible score) to 8 (best possible score) for both sets of test objects.
Film processing quality was evaluated by using an empirical test known as the sensitometric technique for the evaluation of processing, or STEP (17). This test was used to obtain a numeric measure of the sensitometric performance of a particular film processor, called the processing speed, similar in concept to film speed for photography. If a facility processes film in accordance with the manufacturer’s recommendations, then the application of the sensitometric technique for the evaluation of processing will result in a processing speed of approximately 100. Film processors with speeds of less than 80 are considered to be underprocessing the film.
Darkroom fog is an aspect of image quality that is relatively easy to monitor and control (18). The surveyor determined the level of fog by loading the cassette with the radiographic film used routinely at the facility, performing radiography of the phantom, and then reexposing half of the film to ambient conditions in the darkroom for 2 minutes.
|
FINDINGS
|
|---|
All facilities that participated in the survey used screen-film technology for radiographic examinations. Selected findings from the 1995 NEXT survey were compared with findings from previous NEXT surveys in 1987 and 1989. Skin-entrance air kerma was computed for an anterior-to-posterior patient thickness of 23 cm to permit direct comparison with the reported results of previous NEXT surveys.
Abdomen Examination
Equipment and technique factors.—Table 1 summarizes the survey results for the abdomen examination. The results from the 1995 survey indicate improvements in many technical aspects of this examination, a substantial decrease in the use of single-phase radiographic equipment, and an increase in the use of automatic exposure control technology. In 1995, 17 (13%) of 128 hospitals surveyed used single-phase radiographic equipment, compared with 72 (29%) of 246 in 1987. Correspondingly, 38 (57%) of 67 nonhospital facilities surveyed in 1995 used single-phase equipment, compared with 67 (86%) of 78 in 1989. The percentage of facilities that used grids remained almost unchanged from values reported in 1987 and 1989; 121 (99%) of 122 hospitals and 61 (92%) of 66 nonhospital facilities in the 1995 survey reported the use of grids. The source-to-image distances measured by surveyors were tightly distributed around a mean of 103 cm, with a first quartile, median, and third quartile of 101, 102, and 105 cm, respectively. Reported peak kilovoltage values at facilities surveyed in 1995 showed little change from values reported in the earlier surveys, and the results were consistent between hospital and nonhospital facilities. Radiation exposure times decreased at facilities of both types.
Skin-entrance air kerma and image quality.—Skin-entrance air kerma was determined from exposure measurements made with the phantom by using the beam geometry and technique factors typical for a routine abdomen examination. Figure 3 shows the distribution of skin-entrance air kerma values from the 1995 abdomen survey. To determine the statistical significance of differences found between the 1995 survey results and the results of earlier NEXT surveys, two-tailed t tests were performed. Unless otherwise specified, all statistical conclusions are based on the results of t tests, with a P value of .05 or less considered to indicate a statistically significant difference. There was a statistically significant decrease by 13% (P < .05) in mean skin-entrance air kerma for the abdomen examination at hospitals, from 3.2 mGy (369 mR), reported in 1987, to 2.8 mGy (318 mR), reported in 1995. For nonhospital facilities, mean skin-entrance air kerma decreased by 12%, from 3.4 mGy (387 mR) in 1989 to 3.0 mGy (340 mR) in 1995; this decrease, however, was not statistically significant. If the results for skin-entrance air kerma from the 1995 survey are stratified according to exposure mode, nonhospital facilities using automatic exposure control had a mean skin-entrance air kerma of 2.3 mGy (269 mR), compared with 3.2 mGy (364 mR) for facilities using a manual exposure control technique. Although the two mean values differ by nearly 40%, the difference is not statistically significant (P < .10), partly because of the small numbers in these two populations (22 and 47, respectively, for facilities using automatic versus manual exposure control techniques). At hospital facilities surveyed in 1995, there was a smaller difference between values of skin-entrance air kerma for the two methods of exposure control: Hospitals using automatic exposure control had a mean skin-entrance air kerma of 2.7 mGy (307 mR), while hospitals using a manual technique had a mean skin-entrance air kerma of 3.2 mGy (371 mR).
View larger version (43K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. Bar graph shows the distributions of skin-entrance air kerma values (free in air) for abdomen examinations at hospitals and nonhospitals in the 1995 survey. Mean ± SD was 2.8 mGy ± 1.6 for hospitals and 3.0 mGy ± 2.0 for nonhospitals.
|
|
A radiograph of the phantom with the image quality test tool was obtained at each facility and was used to evaluate general image quality in typical clinical conditions. The background optical densities, compared with corresponding values obtained in the 1987 and 1989 NEXT surveys, showed little change (Table 1). Of more importance, however, is the observation that, in the 1995 survey, radiographs obtained at 31 (15%) of 204 facilities had a background optical density of less than 1.20, and radiographs obtained at 13 (6%) of 204 facilities had a background optical density of less than 0.80. Similar results were observed in the 1987 and 1989 surveys, in which 27 (12%) of 219 and eight (13%) of 62 facilities, respectively, produced phantom film images with a background optical density of less than 1.20. Low background optical densities can cause unacceptable image quality and indicate either that film processing quality is not optimal or that the radiographic technique must be adjusted. Figure 4 shows the distribution of phantom film image background optical densities from the 1995 survey.
View larger version (52K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4. Bar graph shows the distributions of background optical density values measured with the phantom for abdomen examinations at hospitals and nonhospitals. Bin range is x-axis tick label ± 0.20. Mean ± SD was 1.72 ± 0.48 for hospitals and 1.77 ± 0.69 for nonhospitals.
|
|
The surveyors also reported the numbers of image quality test objects (ie, low-contrast holes and high-contrast wire mesh patterns) that were visible on the phantom radiograph. The score for the number of visible low-contrast holes was then used to compute the corresponding percent contrast. The mean score for the number of low-contrast holes visible was 4.9 for hospitals and 4.5 for nonhospital facilities, with a possible maximum score of 8.0 (for depiction of eight holes). All 133 hospital facilities and 67 (96%) of 70 nonhospital facilities surveyed obtained images that clearly depicted at least three of the holes; 34 (26%) of 133 hospitals and 12 (17%) of 70 nonhospital facilities produced phantom film images on which six or more holes were visible. The corresponding means for percent contrast were 2.5% and 2.9% for hospital and nonhospital facilities, respectively. No hospital surveyed had a score higher than 5% for percent contrast, and only three nonhospital facilities had scores higher than 5%. Figure 5 shows the distribution of percent contrast values determined in the survey of abdomen examinations. To determine whether background optical density affected percent contrast performance, statistical testing was performed to test the hypothesis that the sample (n = 201) was drawn from a population in which there was no relationship between percent contrast and background optical density. The hypothesis that the correlation coefficient between these two parameters was zero was tested by using the statistic
with df = n – 2, where r is the correlation coefficient and n is the sample size. We found better contrast resolution with increasing background optical density on radiographs (P < .001) (Fig 6).
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5. Bar graph shows the distributions of percent contrast values for abdomen examinations at hospitals and nonhospitals. Mean ± SD was 2.5% ± 0.9 for hospitals and 2.9% ± 1.4 for nonhospitals.
|
|
View larger version (63K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6. Bar graph shows the correspondence of percent contrast to background optical density values measured with the phantom for abdomen examinations. Bin range is x-axis tick label ± 0.10.
|
|
The score for the number of copper wire mesh patterns visible on the phantom radiograph is an indicator of the spatial resolution ability of the radiographic system. These image quality test objects encompass a suitable range of spatial detail, and their depiction (or nondepiction) not only provides a snapshot of this aspect of image quality performance but also permits trend analysis as future surveys are conducted. Figure 7 shows the distribution of mesh scores for hospitals and nonhospitals. In general, as the score (ie, the number of visible mesh patterns) increases, the spatial resolution capability of the radiographic system improves. Among a possible maximum of eight mesh patterns depicted, the mean number of mesh patterns visible for hospitals (4.1) was not significantly different from that for nonhospitals (3.8). With regard to the distribution of scores, however, 10 (14%) of the 70 phantom film images obtained at nonhospital facilities showed only the two largest mesh patterns, while only two (1%) of 134 film images obtained at hospitals showed only the two largest patterns. The first-, second-, and third-largest mesh patterns correspond with spatial resolutions of 0.8, 1.2, and 1.6 lines per millimeter, respectively (Fig 2).
Lumbosacral Spine Examination
Equipment and technique factors.—Table 2 provides a summary of findings for the 1995 survey of the lumbosacral spine examination, alongside comparative results from the previous surveys of 1987 and 1989. The survey results for chiropractic facilities are included in Table 2, part B, and are presented separately in Table 2, part C. Although improvements were observed in the lumbosacral spine survey in many technical aspects, including decreased use of single-phase x-ray equipment, increased use of automatic exposure control techniques, and a substantial decrease in exposure times, the majority of private practice facilities were still using single-phase equipment and performing these examinations by using manual techniques. Peak tube potential for the lumbosacral spine examination has remained essentially constant across all facility types, as indicated by the results in Table 2. Source-to-image distances measured by surveyors were distributed similarly to those measured for the abdomen examination, with a mean of 103 cm and a first, median, and third quartile of 101, 102, and 104 cm, respectively. Exposure time results, however, showed substantial reduction across all facility types, with hospital and nonhospital facilities decreasing mean exposure time by approximately 50% and 27%, respectively (Table 2, parts A and B).
View this table:
[in this window]
[in a new window]
|
TABLE 2. Summary of NEXT Survey Results for the Lumbosacral Spine Examination at Hospitals and Nonhospital and Chiropractic Facilities
|
|
Skin-entrance air kerma and image quality.—Skin-entrance air kerma was determined from the exposure measurements taken with the phantom and by using the beam geometry and clinical technique factors for a routine lumbosacral spine examination. Figure 8 shows the distribution of skin-entrance air kerma values found for hospital and nonhospital facilities in the 1995 survey. A moderate decrease (by 12%) was found in mean skin-entrance air kerma for hospitals (P < .10), from 3.7 mGy (424 mR) in 1987 to 3.3 mGy (374 mR) in 1995, and a statistically significant decrease (by 16%) was found for all surveyed nonhospital facilities (P < .02), from 3.8 mGy (439 mR) in 1989 to 3.2 mGy (368 mR) in 1995. Among chiropractic facilities, there was little change in mean skin-entrance air kerma between 1989 and 1995. Results of a comparison of the 1995 air kerma data between facilities at which automatic exposure control was used and those at which manual exposure control was used showed no statistically significant difference: Hospitals using a manual technique had a mean skin-entrance air kerma of 3.0 mGy (343 mR) compared with 3.3 mGy (373 mR) for hospitals using automatic exposure control, and nonhospital facilities using a manual technique had a mean skin-entrance air kerma of 3.3 mGy (378 mR) compared with 2.8 mGy (317 mR) for facilities using automatic exposure control.
View larger version (38K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8. Bar graph shows the distributions of skin-entrance air kerma values for lumbosacral spine examinations at hospitals and nonhospitals in the 1995 survey. Mean ± SD was 3.3 mGy ± 1.6 for hospitals and 3.2 mGy ± 2.3 for nonhospitals.
|
|
The background optical density and image quality test tool scores from the radiograph of the phantom are good indicators of clinical image quality. For the survey of lumbosacral spine examinations, a radiograph of the image quality test tool was obtained with the tool positioned on the raised spine-like portion of the phantom. Since no test tool was incorporated into the earlier surveys in 1987 and 1989, comparisons are not possible. A statistically significant increase (P < .001) in optical density on the phantom film image was found for hospitals from 1987 to 1995, while the optical density values for nonhospital facilities showed no appreciable change. Figure 9 shows the distribution of background optical densities on the phantom film image for all facilities surveyed in 1995. As was found in the survey of abdomen examinations, a very high proportion of facilities—56 (39%) of 142 hospitals, and 80 (46%) of 176 nonhospital facilities including 29 (39%) of 74 chiropractic offices—produced film images with background optical densities of less than 1.20. These percentages, however, are improved over those observed in previous surveys: A total of 129 (60%) of 216 hospitals in 1987 and 115 (51%) of 225 nonhospital facilities in 1989 produced phantom film images with optical densities of less than 1.20.
View larger version (50K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9. Bar graph shows the distributions of background optical density values for lumbosacral spine examinations at hospitals and nonhospitals. Bin range is x-axis tick label ± 0.20. Mean ± SD was 1.34 ± 0.41 for hospitals and 1.31 ± 0.59 for nonhospitals.
|
|
The reported scores for the number of holes visible on the phantom radiograph were used to determine the corresponding percent contrast for the x-ray imaging system. Figure 10 shows the resultant distribution of values for percent contrast, for surveyed hospital and nonhospital facilities. Ten (7%) of 138 hospitals and 33 (20%) of 167 nonhospital facilities had percent contrast values greater than 5%. A closer analysis of the data for the 43 facilities with a percent contrast greater than 5% found a lower mean skin-entrance air kerma (2.8 mGy versus 3.3 mGy), significantly lower processing speed (P < .002), and significantly lower background optical densities (P < .001), compared with those for the other facilities surveyed. In fact, 31 (72%) of these 43 facilities produced phantom film images that had a mean background optical density of less than 1.00, and 14 (33%) of the 43 facilities produced images that had an optical density of less than 0.60. We performed the same t test described earlier for the abdomen examination survey data to determine whether there was a statistically significant difference between percent contrast and phantom film image background optical density (Fig 11), and we concluded that facilities with low background optical densities are significantly (P < .001) more likely to have lower image quality and, specifically, worse contrast performance.
View larger version (41K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10. Bar graph shows the distributions of percent contrast values for lumbosacral spine examinations at hospitals and nonhospitals. Mean ± SD was 3.2% ± 1.4 for hospitals and 3.7% ± 1.8 for nonhospitals.
|
|
The numbers of visible mesh patterns were slightly lower at lumbosacral spine examinations than at abdomen examinations, but, as can be seen from the distributions plotted in Figure 12, hospitals performed significantly better than nonhospitals (P < .001). The mean score for the number of mesh patterns visible on the phantom radiograph was 3.8 for hospitals and 3.3 for nonhospitals. Moreover, phantom images at 12 (9%) of 141 hospitals and at 39 (22%) of 177 nonhospitals depicted two or fewer mesh patterns. No images obtained at any facility surveyed for the lumbosacral spine examination depicted the seventh or eighth mesh patterns, which correspond to 3.9 lines per millimeter and 4.7 lines per millimeter, respectively.
View larger version (40K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12. Bar graph shows the distributions of image quality mesh scores for lumbosacral spine examinations at hospitals and nonhospitals. Mean ± SD was 3.8 ± 1.0 for hospitals and 3.3 ± 1.0 for nonhospitals.
|
|
Film Processing and Darkroom Fog
Previous NEXT surveys and related publications document trends in film processing quality, as well as the need to optimize this link in the radiographic imaging chain (19). The results from the 1995 survey, along with comparative data from previous surveys and mammography facilities inspected under the auspices of the Mammography Quality Standards Act (MQSA), are shown in Table 3. Figure 13 shows the distribution of film processing speeds for hospital and nonhospital facilities surveyed in 1995. Data for hospitals showed a statistically significant (P < .001) improvement in the quality of film processing, with an increase in mean processing speed from 88 in 1987 to 98 in 1995, whereas data for nonhospital and chiropractic facilities showed only modest improvement in mean processing speed. Of more importance, however, is the observation that, across all facility types, the relative number of facilities observed to have unacceptable processing speed (<80) decreased substantially. Figure 13 also demonstrates a noticeably broader distribution of processing speeds for nonhospital facilities compared with hospitals, an indication that hospitals may tend to make a greater effort to follow the film manufacturer’s recommendations and address quality assurance issues regarding the processor. A high level of film processing (processing speed, >120) may occur if a facility is performing extended-cycle processing. Although an extended processing cycle might improve film image contrast resolution and increase film speed for single-emulsion films such as those used in mammography, it does not yield similar benefits for double-emulsion films, and therefore it is not recommended for abdomen and lumbosacral spine examinations (20). Finally, we compared film processing quality for abdomen and lumbosacral spine examinations with that for mammography, a radiographic examination for which the Food and Drug Administration mandates a standard for film processing quality. Data from 1995 MQSA inspections show a mean processing speed of 95 (standard cycle), with 185 (5.9%) of 3,120 facilities having processing speeds of less than 80. At MQSA inspections conducted in 1997, the mean processing speed had increased to 103, and only 60 (1.0%) of 5,737 facilities had speeds of less than 80. Clearly, there is room for further improvement in film processing quality for nonmammographic examinations.
View this table:
[in this window]
[in a new window]
|
TABLE 3. Comparison of Film Processing Quality (Speed) Values Measured during the NEXT Surveys and at MQSA Inspections
|
|
View larger version (51K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13. Bar graph shows the distributions of film processing speeds at hospitals and nonhospitals in the 1995 survey. Bin range is x-axis tick label ± 5. Mean ± SD was 98 ± 14 for hospitals and 92 ± 24 for nonhospitals.
|
|
Darkroom fog was assessed by using the same film and the same darkroom conditions as were typically used for radiography at the facility. While sectors of the professional community may differ on the recommended threshold value for diagnostically important levels of fog, it can be argued that any level of fog that is visible on clinical film images is cause for corrective action. Table 4 and Figure 14 display the results of the darkroom fog test performed during the 1995 survey. No darkroom fog test was performed during the 1987 and 1989 NEXT surveys. Ninety-seven (31.1%) of 312 facilities surveyed in 1995 produced images that had darkroom fog optical densities in excess of 0.10, and the majority of facilities surveyed had fog in excess of 0.05, the MQSA standard for darkroom fog. MQSA inspection data for 1995 indicate that only 315 (3.7%) of 8,605 mammography facilities in the United States had darkroom fog levels greater than 0.10; the proportion of such facilities decreased to 161 (2.3%) of 6,872 by 1997.
View larger version (43K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 14. Bar graph shows the distribution of values for darkroom fog optical densities measured at hospitals and nonhospitals during the 1995 survey. Mean was 0.09 for hospitals and 0.12 for nonhospitals.
|
|
|
DISCUSSION
|
|---|
Radiation exposure measurements obtained at hospital and nonhospital facilities showed an appreciable reduction in patient exposure at both abdomen and lumbosacral spine examinations. The slightest reduction in patient exposure was observed in chiropractic facilities performing lumbosacral spine examinations. The dose reduction achieved at hospitals is especially noteworthy when one considers that, from 1987 to 1995, hospitals more than doubled their workloads and that these workloads are typically much larger than those of nonhospital facilities.
Many factors may have contributed to the reductions in patient skin-entrance radiation exposure. Facilities of all types reduced the use of single-phase equipment, with hospital facilities having achieved the greatest reduction. All facility types had increased the use of automatic exposure control techniques since the earlier surveys, with hospitals using automatic exposure control techniques predominantly, while nonhospital facilities, particularly chiropractic offices, continued to favor manual techniques. Hospitals achieved the greatest reduction in exposure time, a result that is consistent with the observations of reduced use of single-phase equipment and increased use of higher tube currents for both clinical examinations. Shorter exposure times can help to reduce the frequency of repeated exposures during the same examination by decreasing the occurrence of motion-induced artifacts on the initial film images.
The influence of film processing speed on patient radiation exposure is often overlooked. Processing speed is inversely related to the radiation exposure needed to maintain a constant film optical density. All facility types achieved improved quality in film processing, but the most dramatic improvement was observed for hospitals between 1987 and 1995, a period during which the percentage of facilities that were underprocessing their film (processing speed, <80) was reduced by a factor of almost 5. This translates into a potential reduction of approximately 25% in radiation exposure to the patient.
To better ascertain the effect of film processing speed on skin-entrance air kerma, we first determined the relative air kerma value for facilities at which film processing speed was less than 80 and then adjusted the skin-entrance air kerma value for these facilities to correspond with a processing speed of 80, the minimum acceptable speed. For facilities performing lumbosacral spine examinations, the average reduction in patient skin-entrance air kerma was 0.7 mGy (80 mR), and for facilities performing abdomen examinations, the average reduction was 0.6 mGy (73 mR). Given that the average skin-entrance air kerma for a routine adult posteroanterior chest radiograph, according to results from the 1994 NEXT survey, was 0.14 mGy (16.1 mR), these exposure reductions are equivalent to the radiation exposure from more than four chest radiographs. The influence of film processing quality on patient exposure clearly must be addressed and the film processing quality must be optimized at such facilities.
The effect of film processing on image quality is easy to overlook, as well. We examined the values of percent contrast separately for facilities from the lumbosacral spine survey that had processing speeds of less than 80 and for facilities with processing speeds greater than or equal to 80 (Fig 15). A t test performed for the difference between the mean values for percent contrast showed that facilities with acceptable film processing quality had significantly better percent contrast performance (P < .001). The same analysis performed for facilities that conduct abdomen examinations did not show a significant difference in percent contrast; however, there were only 26 such facilities with a processing speed below 80. We nonetheless concluded that film processing quality, as a contributing factor to good image quality, cannot be ignored.
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 15. Bar graph shows the distributions of percent contrast values for lumbosacral spine examinations at facilities with processing speeds of less than 80 (low) and facilities with processing speeds greater than or equal to 80 (normal).
|
|
Finally, the fact that a large number of facilities produced phantom film images with background optical densities below 1.20 is of concern. Proper exposure of the film is essential to obtain the best possible radiographic image consistent with the characteristics of the particular film used. The distribution of values in Figures 6 and 11 illustrates this well: Regardless of the film brand or type used at the facilities surveyed, an improvement in percent contrast occurred with increasing background optical density. It is difficult to suggest an optimal value for optical density, given the variety of film types available and the possible variations in personal preferences of practitioners. Our data, however, indicate that optical densities in the range of 1.20–2.00 and, possibly, higher (comparatively few facilities produced film images with background optical densities above 2.00) were associated with better contrast performance, compared with optical densities below 1.20. We recommend consultation with a medical physicist or with a technical representative of the film manufacturer to further narrow this range.
|
CONCLUSIONS
|
|---|
The results of the 1995 NEXT survey of U.S. hospital and nonhospital facilities that perform abdomen and lumbosacral spine radiographic examinations show a reduction in patient radiation exposure and an improvement in film processing quality since the 1987 and 1989 surveys. The different levels of performance at hospital versus nonhospital facilities are attributable in part to the radiographic equipment: Hospitals used single-phase generators less often, used lower exposure times, and used automatic exposure control techniques more often than did nonhospital facilities in the 1995 survey. Many facilities could improve the quality of their practice by paying more attention to the quality of film processing, darkroom fog, and background optical density. A high percentage of facilities produced film images that had high levels of darkroom fog and background optical density in excess of 0.10, but the results of MQSA inspections show that a reduction in darkroom fog is achievable. Facilities performing the lumbosacral spine examination that had film processing speeds of less than the acceptable level of 80 also had percent contrast values significantly lower than those at facilities that achieved acceptable film processing quality. The achievement and maintenance of high-quality film processing enable facilities to reduce patient exposure and remove a random element from the chain of parameters that affect patient radiation dose and image quality. These improvements are readily achievable and can have a positive effect on the clinical value of diagnostic radiography examinations.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank the many state radiation control personnel who devoted much of their time to conducting these surveys. We are also grateful to the Conference of Radiation Control Program Directors H-4 committee on NEXT for their efforts. We thank James E. Duff (retired), Bruce R. Fleharty, and Randolph L. Bidinger from the machine shop at the Food and Drug Administration Office of Science and Engineering Laboratories for their excellent work in producing the phantoms, test tools, and various other items used during the surveys. Finally, we thank Jane Tuttle-Kuhm for completing the exhausting task of data entry.
|
FOOTNOTES
|
|---|
The mention of commercial products, their sources, or their use in connection with this study is not to be construed as either an actual or an implied endorsement by the Food and Drug Administration.
Abbreviations: MQSA = Mammography Quality Standards Act,
NEXT = Nationwide Evaluation of X-ray Trends
|
REFERENCES
|
|---|
- Conference of Radiation Control Program Directors. Nationwide Evaluation of X-ray Trends (NEXT) Summary of 1994 Chest Survey Conference of Radiation Control Program Directors Publication No. 98–2. Frankfort, Ky: Conference of Radiation Control Program Directors, 1998.
- Kaczmarek RV, Conway BJ, Slayton RJ, Suleiman OH. Results of a nationwide survey of chest radiography: comparison with results of a previous study. Radiology 2000; 215:891-896.[Abstract/Free Full Text]
- Conference of Radiation Control Program Directors. Nationwide Evaluation of X-ray Trends (NEXT) Summary of 1990 Computed Tomography and 1991 Fluoroscopy Survey Conference of Radiation Control Program Directors Publication No. 94–2. Frankfort, Ky: Conference of Radiation Control Program Directors, 1994.
- Conway BJ, McCrohan JL, Antonsen RG, Rueter FG, Slayton RJ, Suleiman OH. Average radiation dose in standard CT examinations of the head: results of the 1990 NEXT survey. Radiology 1992; 184:135-140.[Abstract/Free Full Text]
- Conference of Radiation Control Program Directors. Nationwide Evaluation of X-ray Trends (NEXT) Tabulation and Graphical Summary of 1995 Abdomen and Lumbosacral Spine Surveys Conference of Radiation Control Program Directors Publication No. 00–2. Frankfort, Ky: Conference of Radiation Control Program Directors, 2000.
- Rueter FG, Conway BJ, McCrohan JL, Suleiman OH. Average radiation exposure values for three diagnostic radiographic examinations. Radiology 1990; 177:341-345.[Abstract/Free Full Text]
- Conference of Radiation Control Program Directors. Nationwide Evaluation of X-ray Trends (NEXT) Tabulation and Graphical Summary of Surveys 1984 through 1987 Conference of Radiation Control Program Directors Publication No. 89–3. Frankfort, Ky: Conference of Radiation Control Program Directors, 1989.
- Rueter FG, Conway BJ, McCrohan JL, Slayton RJ, Suleiman OH. Radiography of the lumbosacral spine: characteristics of examinations performed in hospitals and other facilities. Radiology 1992; 185:43-46.[Abstract/Free Full Text]
- Conference of Radiation Control Program Directors. Nationwide Evaluation of X-ray Trends (NEXT) Tabulation and Graphical Summary of 1989 Abdomen and Lumbo-sacral Spine Survey in Private Practice and Chiropractic Offices Conference of Radiation Control Program Directors Publication No. 90–3. Frankfort, Ky: Conference of Radiation Control Program Directors, 1991.
- Conway BJ, Suleiman OH, Rueter FG, Antonsen RG, Slayton RJ. National survey of mammographic facilities in 1985, 1988, and 1992. Radiology 1994; 191:323-330.[Abstract/Free Full Text]
- Suleiman OH, Spelic DC, McCrohan JL, Symonds GR, Houn F. Mammography in the 1990s: the United States and Canada. Radiology 1999; 210:345-351.[Abstract/Free Full Text]
- Conference of Radiation Control Program Directors. Nationwide Evaluation of X-ray Trends (NEXT) Tabulation and Graphical Summary of 1993 Dental Survey Conference of Radiation Control Program Directors Publication No. 96–3. Frankfort, Ky: Conference of Radiation Control Program Directors, 1996.
- Suleiman OH, Spelic DC, Conway B, Hart JC, Boyce PR, Antonsen RG, Jr. Radiographic trends of dental offices and dental schools. J Am Dent Assoc 1999; 130:1104-1110.[Abstract/Free Full Text]
- Suleiman OH, Conway BJ, Quinn P, et al. Nationwide survey of fluoroscopy: radiation dose and image quality. Radiology 1997; 203:471-476.[Abstract/Free Full Text]
- American Hospital Association. The AHA guide to the health care field Chicago, Ill: American Hospital Association, 1993.
- Conway BJ, Duff JE, Fewell TR, Jennings RJ, Rothenberg LN, Fleischman RC. A patient-equivalent attenuation phantom for estimating patient exposures from automatic exposure controlled x-ray examinations of the abdomen and lumbo-sacral spine. Med Phys 1990; 17:448-453.[CrossRef][Medline]
- Suleiman OH, Rueter FG, Antonsen RG, Conway BJ, Slayton RJ. The sensitometric technique for the evaluation of processing (STEP). Radiat Prot Dosimetry 1993; 49:105-106.[Abstract]
- Suleiman OH, Showalter CK, Gross RE, Bunge RE. Radiographic film fog in the darkroom. Radiology 1984; 151:237-238.[Abstract/Free Full Text]
- Suleiman OH, Conway BJ, Rueter FG, Slayton RJ. Automatic film processing: analysis of 9 years of observations. Radiology 1992; 185:25-28.[Abstract/Free Full Text]
- Haus AG, Jaskulski SM. The basics of film processing in medical imaging Madison, Wis: Medical Physics Publishing, 1997; 122.
This article has been cited by other articles:
|
|
|
|
 
W. E. Muhogora, N. A. Ahmed, A. Almosabihi, J. S. Alsuwaidi, A. Beganovic, O. Ciraj-Bjelac, F. K. Kabuya, A. Krisanachinda, M. Milakovic, G. Mukwada, et al.
Patient Doses in Radiographic Examinations in 12 Countries in Asia, Africa, and Eastern Europe: Initial Results from IAEA Projects
Am. J. Roentgenol.,
June 1, 2008;
190(6):
1453 - 1461.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
