Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bassett, L. W.
Right arrow Articles by Feig, S. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bassett, L. W.
Right arrow Articles by Feig, S. A.
(Radiology. 2000;215:698-702.)
© RSNA, 2000

Breast Imaging

Reasons for Failure of a Mammography Unit at Clinical Image Review in the American College of Radiology Mammography Accreditation Program1

Lawrence W. Bassett, MD, Dione M. Farria, MD, MPH, Swati Bansal, MS, Marybeth A. Farquhar, RN, MSN, Pamela A. Wilcox, MBA and Stephen A. Feig, MD

1 From the Iris Cantor Center for Breast Imaging, University of California Los Angeles School of Medicine (L.W.B.); the Breast Imaging Center, Thomas Jefferson University, Philadelphia, Pa (D.M.F.); the American College of Radiology, Reston, Va (S.B., M.A.F., P.A.W.); and the Department of Radiology, Mount Sinai School of Medicine, New York, NY (S.A.F.). Received June 18, 1999; revision requested July 14; revision received September 1; accepted September 14. Address correspondence to D.M.F., Mallinckrodt Institute of Radiology, Washington University Medical Center, 510 S Kingshighway Blvd, St Louis, MO 63110 (e-mail: farriad@mir.wustl.edu).


   Abstract
TOP
 Abstract
Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To identify the most common deficiencies in the quality of mammograms submitted for clinical image evaluation (evaluation of image from actual patient referred for mammography).

MATERIALS AND METHODS: In 1997, the American College of Radiology Mammography Accreditation Program reviewed clinical images for 2,341 mammography units. For each mammography unit, the facility submitted bilateral mediolateral oblique and craniocaudal mammograms obtained in a woman with fatty breasts and a woman with dense breasts. Images were reviewed independently by two experienced radiologists. Reviewers listed the general categories and specific deficiencies that led to a decision to fail the unit that produced the clinical images.

RESULTS: Of the 2,341 mammography units, 1,034 (44%) failed the clinical image evaluation process. Of 6,128 categories cited by reviewers as deficient, 1,250 (20%) involved problems in positioning; 944 (15%), exposure; 887 (14%), compression; 806 (13%), sharpness; 785 (13%), contrast; 703 (11%), labeling; 465 (8%), artifacts; and 288 (5%), noise. A significantly higher proportion of failures was attributed to positioning deficiencies for fatty breasts than for dense breasts (P = .028). Higher proportions of failures in dense breasts were related to compression (P < .001) and exposure (P < .001) deficiencies.

CONCLUSION: Common problems in clinical image quality have been identified. This information should be useful for educators and facilities striving to improve the quality of mammography.

Index terms: Breast radiography, quality assurance, 00.11, 00.93, 00.99 • Quality assurance


   Introduction
TOP
 Abstract
 Introduction
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
The most common reason that facilities do not succeed in their first application to the American College of Radiology (ACR) Mammography Accreditation Program (MAP) criteria is unsatisfactory clinical image review (image from actual patient referred for mammography). The Mammography Quality Standards Act (MQSA) of 1992 mandates that all approved accrediting bodies conduct "a review of clinical images from each facility accredited . . . not less often than every 3 years which review will be made by qualified practicing physicians" (1). The final regulations of MQSA, which went into effect in April 1999, specify that the clinical image review should address breast positioning, compression, exposure level, contrast, sharpness, artifacts, and examination identification (image labeling) (2). Due to the variations in body habitus in patients and their ability to cooperate, it is not possible to attain ideal breast positioning and compression in all women (3). Therefore, facilities are requested to submit what they consider to be their best representative images. Each facility submits images for each mammography unit.

In ACR MAP, there are more than 110 radiologists who conduct peer reviews of clinical images that are submitted for accreditation. Two independent clinical image reviewers assess each set of mammograms. In the event of disagreement, the outcome is determined by a third reviewer. All ACR MAP clinical image reviewers meet the following standards: (a) an active practice in which 50% or more of the practice is dedicated to breast imaging, (b) member of the ACR, (c) MQSA-qualified interpreting physician, (d) American Board of Radiology certification, (e) practice in an accredited facility, and (f) recommendation for the position from a current clinical image reviewer. The performance of all reviewers is monitored on a quarterly basis. These reports evaluate consistency among reviewers and provide each reviewer with an assessment of his or her reviews for the most recent quarter. Reviewers are informed of their individual rates and whether their individual rates are within 1 SD of the mean rate for all reviewers. Corrective action is initiated for reviewers who score outside of the acceptable range.

Clinical image evaluation has been a component of the ACR MAP since the inception of ACR MAP in 1987 (4). In the 3 years before this article was written, ACR MAP had reviewed clinical images from nearly 11,300 facilities, which represents approximately 94% of mammography facilities nationwide (5). We conducted a retrospective analysis of all ACR MAP clinical image evaluations in 1997 to identify the types and numbers of deficiencies that cause a mammography unit to fail the clinical image–evaluation process.


   MATERIALS AND METHODS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
We conducted a secondary data analysis based on ACR MAP clinical image reviews completed in 1997. Clinical images for 2,341 individual mammography units were submitted for review in 1997. A mammography unit is defined as a single machine. A facility may have more than one unit. The 2,341 individual mammography units may include units evaluated more than once if there was an initial deficiency and repeat evaluation in 1997. For each mammography unit, a facility submitted two sets of images. One set included the standard mediolateral oblique and craniocaudal views in a woman with primarily dense breasts; the other set included the standard mammographic views in a woman with fatty breasts. The fatty tissue composition includes the entirely fatty breast or the breast with scattered fibroglandular densities. The dense breast composition includes the heterogeneously dense breast or the extremely dense breast (6).

Image quality was reviewed for the following eight image quality categories: positioning, compression, exposure, contrast, sharpness, noise, artifacts, and labeling. For each of these categories, specific deficiencies were noted and recorded by ACR MAP reviewers for all submitted sets of clinical images. Thus, there could be multiple deficiencies in any of the categories (Table 1).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Potential Deficiencies in Image Quality Categories Used in the Clinical Image–Evaluation Process
 
For the purposes of this analysis, we defined "deficiency" as any problem within a category that contributed to a category score of 1, 2, or 3 on a five-point scale (1 = worst, 5 = best). A score of 1 or 2 indicated an important image defect; a score of 3 was reserved for mildly deficient or marginally acceptable images. A score of 1 or 2 in a single category or a score of 3 in multiple categories typically led to failure. Failure was defined as the requirement for resubmission of any set of clinical images, in fatty or dense breasts, for evaluation to demonstrate adequate image quality.

To identify the most common categories with problems that led to a failed clinical image review, we tabulated the frequencies with which each of the eight general image quality categories contributed to a decision of a failure grade. We further identified and tabulated the specific types and frequencies of deficiencies within each of these eight categories. To identify image quality problems specific to fatty versus dense breasts, we compared image quality categories identified as contributing to failure for these two breast compositions.

Using the test of significance of proportions (7), we identified the categories where the percentages of failure were significantly different (P < .05) in fatty versus dense breasts. We also used the Spearman rank correlation coefficient to determine whether failure in one category significantly correlated with failure in other categories (7). Since clinical image reviewers do not fail a mammography unit in a facility for isolated artifact or image-labeling deficiencies, we also tabulated all serious deficiencies in these two categories, regardless of whether a mammography unit in a facility failed the review process. This additional information was a more accurate reflection of the frequency of serious labeling and artifact problems in the sample. A "serious" problem was defined as an image quality category that received a "1" or "2" rating on a scale of 1 to 5 (1 = worst, 5 = best).


   RESULTS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Of the 2,341 mammography units that were evaluated for accreditation in 1997, 1,254 failed one or more of the following areas on their evaluation: clinical image quality, acceptable imaging of simulated masses and calcifications in a test phantom, processor quality control, and radiation dose. Clinical image quality was the most common. Of the 1,254 units that failed, 1,034 had failed the clinical image quality review, including 951 (76%) that failed only the clinical image review process and 83 (7%) that failed the clinical image review process and had serious problems in other areas.

Table 2 reports the frequency of problems in the eight image quality categories that were judged to be serious enough to lead to failure. Several categories were usually cited as responsible for a failed clinical image evaluation. Overall, individual categories were cited 6,128 times as reasons for failure. The most common image quality category leading to failure was positioning, which affected 1,250 sets of submitted images, followed by exposure, compression, sharpness, and contrast. In the 1,250 sets of images with problems in positioning, there were 3,400 specific positioning deficiencies cited by reviewers (Table 3).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Frequency of Image Quality Category Problems in Fatty versus Dense Breasts
 

View this table:
[in this window]
[in a new window]

  		TABLE 3. Specific Deficiencies in Positioning in Failed Sets of Clinical Images
 
In the 944 sets of images with exposure problems, there were 1,403 specific exposure deficiencies, including 642 (46%) for generalized underexposure, 556 (40%) for inadequate x-ray penetration of dense areas, 131 (9%) for generalized overexposure, and 74 (5%) for other reasons. In 887 sets of images with compression problems, reviewers cited 1,480 total deficiencies, including 869 (59%) for inadequate separation of parenchymal tissues, 344 (23%) for patient motion, 238 (16%) for nonuniform exposure levels, and 29 (2%) for other deficiencies.

There were 806 sets of submitted images with sharpness problems. In the 806 sets of images, there were 1,514 specific deficiencies related to image sharpness, including 889 (59%) for poor delineation of linear structures, 489 (32%) for unsharpness of feature margins, and 136 (9%) for blurring of microcalcifications. In 785 sets of images with contrast problems, there were 964 total contrast deficiencies: Inadequate contrast was responsible for 862 (89%) citations; excessive contrast, for 96 (10%) citations; and other reasons, for six (1%) citations. In 288 sets of images, there were 458 noise-related deficiencies, including 339 (74%) for a visually striking mottle pattern and 119 (26%) for limited visualization of detail because of noise.

The data showed significant correlation among compression, sharpness, contrast, and exposure deficiencies. The greatest correlations were noted between compression and sharpness (r = 0.4046, P < .001) and between exposure and contrast (r = 0.4766, P < .001). Significant correlations were also noted between compression and exposure (r = 0.1537, P < .001), sharpness and exposure (r = 0.1083, P < .001), and contrast and sharpness (r = 0.1413, P < .001). Table 2 compares the deficiencies in the submitted images of the primarily fatty versus the primarily dense breasts.

There were 703 sets of images with artifact problems that resulted in failure in the clinical image–evaluation process. In the total sample of submitted images, there were 1,241 serious deficiencies in the artifact category, including 465 (37%) for lint, 362 (29%) for scratches or pickoff (small defects in film emulsion due to processing), 148 (12%) for grid-related artifacts, 115 (9%) for roller marks, and 151 (12%) for other artifacts.

There were 465 sets of images with labeling problems among cases that resulted in failure in the clinical image–review process. In the total sample of submitted images, there were 831 serious deficiencies in the labeling category, including 419 (50%) for inadequate identification of the facility, 101 (12%) for failure to identify the radiologic technologist, 94 (11%) for failure to place an identifying number on the intensifying screen for each cassette, 87 (10%) for improperly labeled mammographic view, 78 (9%) for inadequate identification of the patient, and 52 (6%) for other labeling problems.

We calculated the frequency of serious artifact and labeling problems in the total sample of submitted images. Of the 831 serious labeling problems, 542 (65%) were identified in the sample of images that passed the clinical image review.


   DISCUSSION
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Familiarity with criteria for quality images and maintenance of that level of quality is the foundation of a mammography quality assurance program. Daily clinical image evaluation by the radiologic technologist and the interpreting physician is an important component of that quality assurance program. Analysis of the ACR MAP clinical image–evaluation data provided an opportunity to identify the most common problems on clinical images submitted by mammography facilities nationwide. We expect the results of this analysis to be useful in the design of educational curricula in breast imaging.

Findings of our review demonstrated that common problems were positioning and compression. Problems in the positioning category were most frequently due to an inadequate amount of pectoralis major muscle on the mediolateral oblique view, poor visualization of posterior tissue on the craniocaudal or mediolateral oblique view, sagging breasts, or skin folds. Inadequate compression was manifested by overlapping breast structures, nonuniform tissue exposure, and motion unsharpness. Instructions for achieving proper positioning and compression are explained and illustrated in a number of publications (8–10). Thus, it appears that many practicing radiologic technologists still have a need for actual hands-on training for the standard views, the mediolateral oblique and the craniocaudal views.

Image clarity, or the ability of the mammogram to portray diagnostic information clearly, depends on a number of other factors, which include exposure, sharpness, contrast, and noise (11–18). In the image quality category of exposure, generalized underexposure was a far more frequent deficiency than overexposure. This is of concern because lesions can be obscured in dense underexposed breast tissue, which leads to false-negative mammograms. Generalized underexposure can result from inadequate compression, technical factors, equipment malfunction, or a combination of these factors. Nonuniform exposure of the breast parenchyma with inadequate penetration of denser tissues was another commonly cited deficiency.

Sharpness is the ability to define the edges of breast structures. Unsharpness, also referred to as blur, is manifested by poor delineation of the edges of linear structures, tissue margins, and microcalcifications. Common causes of blur include patient motion and poor screen-film contact. Contrast, which is the degree of variation in optical density between different areas of the film, is affected by film processing, exposure, film type, and scatter reduction (with grids, compression). Inadequate contrast was a more frequent deficiency than excessive contrast.

Noise, manifested as a visually striking mottle pattern, was the least common category that led to failure of the image-review process. However, when noise was substantial enough to limit the visualization of detail, this deficiency sometimes led to failure of the clinical image–review process. Quantum mottle is the major source of noise in mammography. The fewer photons used to make the image, the greater the observed quantum mottle. Other causes of noise include vigorous or prolonged processing, fast image-recording systems, and underexposure.

Although failure of the clinical image–review process was not based solely on the presence of artifacts, egregious artifacts sometimes led to failure. The most common artifacts cited by reviewers were dirt or lint, scratches or pickoff, grid lines, and processor roller marks. Careful film handling and diligent quality control measures should eliminate many of these artifacts.

While deficiencies in labeling alone usually do not lead to a failed clinical image evaluation, inappropriate labeling was a common problem on submitted images (19). The most common deficiency in this category was inadequate identification of the mammography facility. The final regulations of the MQSA, which went into effect on April 28, 1999, list specific requirements for proper image identification (Figure). According to the final regulations, "each image shall have the following information indicated on it in a permanent, legible, and unambiguous manner and placed so as not to obscure anatomic structures": (a) name of the patient and additional patient identifier, (b) date of examination, (c) view and laterality, (d) facility name and location (at a minimum, the location should include city, state, and zip code of the facility), (e) technologist identification, (f) cassette or screen identification, and (g) mammography unit identification if there is more than one unit in the facility (2).



View larger version (43K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Diagram demonstrates proper image identification and labeling. (Reprinted, with permission, from reference 9.)

 
This data analysis documents the importance of submitting two sets of images to represent different breast parenchymal tissue compositions, one of fatty breast tissue composition and the other of dense breast tissue composition (Table 2). A facility may produce acceptable mammograms in women with fatty breasts but unsatisfactory mammograms in women with dense breasts, or vice versa. Fatty breasts may present greater challenges for positioning when they are larger than dense breasts. On the other hand, compression and exposure pose a greater problem in dense breasts than in fatty breasts. This latter finding may be attributed to the challenge of separating and penetrating dense fibroglandular tissue. Likewise, we expected more problems with sharpness in dense breasts, because there are compression challenges and longer exposures associated with dense breasts. However, the lack of difference could be attributed to a higher tolerance among reviewers for blur in dense breasts.

According to our data, deficiencies in exposure, contrast, sharpness, and compression are frequently interrelated. For example, there is a correlation between compression and exposure deficiencies, because adequate exposure depends on sufficient compression of dense fibroglandular tissue. Compression and sharpness deficiencies also often coexist on the same images. This correlation is likely due to the motion unsharpness that results from inadequate compression.

In summary, the high proportion of mammography units in facilities that fail the clinical image–review process is not surprising because of the stringent criteria used for image evaluation. The high expectations are in keeping with the intent of the MQSA, the educational emphasis of the MAP, and the fact that facilities select their best representative images for review. Although many facilities have deficiencies in their clinical image evaluation when applying for accreditation, a majority of facilities successfully complete the clinical image–evaluation process upon reapplication. In fact, it is generally acknowledged that the quality of mammography has markedly improved since the inception of ACR MAP (20).


   Acknowledgments
 
We acknowledge the important contributions to the American ACR MAP clinical image–evaluation process made by Carl J. Vyborny, MD, PhD, Harold Lasky, MD, and Robert Schmidt, MD, who developed the original clinical image–evaluation criteria and grading mechanisms. We also acknowledge the expertise and contributions of the more than 110 radiologists who participate as clinical image reviewers for the ACR MAP. We are especially indebted to Marie Zinninger, MSN, Associate Executive Director of the ACR, for her commitment, support, and leadership in the development of quality mammography.


   Footnotes
 
Abbreviations: ACR = American College of Radiology, MAP = Mammography Accreditation Program, MQSA = Mammography Quality Standards Act

Author contributions: Guarantor of integrity of entire study, L.W.B.; study concepts, D.M.F., L.W.B.; study design, M.A.F., S.B., L.W.B., D.M.F.; definition of intellectual content, D.M.F., L.W.B.; literature research, L.W.B.; data acquisition, P.A.W., M.A.F.; data analysis, L.W.B., S.B., D.M.F.; statistical analysis, S.B.; manuscript preparation, D.M.F., L.W.B.; manuscript editing, D.M.F., L.W.B., S.A.F.; manuscript review, S.A.F., D.M.F., L.W.B., P.A.W.


   References
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Public Law 102-539. The Mammography Quality Standards Act of 1992. 1992; §354:.
  2. Food and Drug Administration. Mammography facilities: requirements for accrediting bodies and quality standards and certification requirements—interim rules. Federal Register December 21, 1993; 208:55980-55981.
  3. Bassett LW, Hirbawi IA, DeBruhl N, Hayes MK. Mammographic positioning: evaluation from the view box. Radiology 1993; 188:803-806.[Abstract/Free Full Text]
  4. McLelland R, Hendrick RE, Zinninger MD, Wilcox PA. The American College of Radiology Mammography Accreditation Program. AJR Am J Roentgenol 1991; 157:473-479.[Abstract/Free Full Text]
  5. Center for Devices and Radiological Health Certification/Accreditation Support System. Facility counts report, 1998 Rockville, MD: Food and Drug Administration, 1998.
  6. American College of Radiology. Breast Imaging Reporting and Data System (BiRADS) 3rd ed. Reston, Va: American College of Radiology, 1998.
  7. Snecdor GW, Cochran WG. Statistical methods 7th ed. Ames, Iowa: Iowa State University Press, 1980.
  8. Bassett LW/MIDDLE>, Hendrick RE/MIDDLE>, Bassford TL, et al. Quality determinants of mammography. Clinical practice guideline, no. 13. Agency for Health Care Policy and Research publication no. 95-0632 Rockville, Md: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services, October 1994.
  9. American College of Radiology (ACR). Mammography quality control manual Reston, Va: American College of Radiology, 1999; 23-77.
  10. Eklund GW, Cardenosa G. The art of mammographic positioning. Radiol Clin North Am 1992; 30:21-53.[Medline]
  11. Alter AJ, Kargas GA, Kargas SA, Cameron JR, McDermott JC. The influence of ambient and viewbox light upon visual detection of low-contrast targets in a radiograph. Invest Radiol 1982; 17:403-406.
  12. American College of Radiology. Recommended specifications for new mammography equipment Reston, Va: American College of Radiology, 1995.
  13. Curry TS, Dowdey JE, Murry RC. The radiographic image. In: Curry TS, Dowdey JE, Murry RC, eds. Christensen's physics of diagnostic radiology. 4th ed. Philadelphia, Pa: Lea & Febiger, 1990; 196-218.
  14. Eklund GW, Cardenosa G, Parsons W. Assessing adequacy of mammographic image quality. Radiology 1994; 190:227-307.[Abstract/Free Full Text]
  15. Helvie MA, Chan HP, Adler DD, Boyd PG. Breast thickness in routine mammograms: effect on image quality and radiation dose. AJR Am J Roentgenol 1994; 163:1371-1374.[Abstract/Free Full Text]
  16. Kimme-Smith C, Rothschild PA, Bassett LW, Gold RH, Moler C. Mammographic film-processor temperature, development time, and chemistry: effect on dose, contrast and noise. AJR Am J Roentgenol 1989; 152:35-40.[Abstract/Free Full Text]
  17. Vyborny CJ, Schmidt RA. Mammography as a radiographic examination: an overview. RadioGraphics 1989; 9:723-764.[Abstract]
  18. American College of Radiology (ACR). Mammography quality control manual Reston, Va: American College of Radiology, 1992; 79-114.
  19. Bassett LW, Jessop NW, Wilcox PA. Mammography film-labeling practices. Radiology 1993; 187:773-775.[Abstract/Free Full Text]
  20. General Accounting Office. Mammography services: initial impact of new federal law has been positive. Publication no. GAO/HEHS-96-17 Washington, DC: General Accounting Office, October 1995.



This article has been cited by other articles:


Home page
 radtechHome page
J. MINIGH
Mammographic Film Artifacts.
Radiol. Technol., May 1, 2006; 77(5): 389M - 402M.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bassett, L. W.
Right arrow Articles by Feig, S. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bassett, L. W.
Right arrow Articles by Feig, S. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE