IMAGING & THERAPEUTIC TECHNOLOGY

Transcription

1 IMAGING & THERAPEUTIC TECHNOLOGY Quality Assurance in Mammography: Artifact Analysis 1 Jacquelyn P. Hogge, MD Carolyn H. Palmer, RTR(M) Caroline C. Muller, RTR(M) Sherrill T. Little, MD Deborah C. Smith, RTR(M) Panos P. Fatouros, PhD Ellen Shaw de Paredes, MD Evaluation of mammograms for artifacts is essential for mammographic quality assurance. A variety of mammographic artifacts (ie, variations in mammographic density not caused by true attenuation differences) can occur and can create pseudolesions or mask true abnormalities. Many artifacts are readily identified, whereas others present a true diagnostic challenge. Factors that create artifacts may be related to the processor (eg, static, dirt or excessive developer buildup on the rollers, excessive roller pressure, damp film, scrapes and scratches, incomplete fixing, power failure, contaminated developer), the technologist (eg, improper film handling and loading, improper use of the mammography unit and related equipment, positioning and darkroom errors), the mammography unit (eg, failure of the collimation mirror to rotate, grid inhomogeneity, failure of the reciprocating grid to move, material in the tube housing, compression failure, improper alignment of the compression paddle with the Bucky tray, defective compression paddle), or the patient (eg, motion, superimposed objects or substances [jewelry, body parts, clothing, hair, implanted medical devices, foreign bodies, substances on the skin]). Familiarity with the broad range of artifacts and the measures required to eliminate them is vital. Careful attention to darkroom cleanliness, care in film handling, regularly scheduled processor maintenance and chemical replenishment, daily quality assurance activities, and careful attention to detail during patient positioning and mammography can reduce or eliminate most mammographic artifacts. Abbreviations: ACR = American College of Radiology, QC = quality control Index terms: Breast radiography, Breast radiography, quality assurance RadioGraphics 1999; 19: From the Department of Radiology, Medical College of Virginia of Virginia Commonwealth University, 1101 E Marshall St, Box , Richmond, VA (J.P.H., C.P., C.M., S.T.L., P.F., E.S.D.); and the Department of Radiology, University of Virginia Medical Center, Charlottesville, Va (D.S.). Recipient of a Certificate of Merit award for a scientific exhibit at the 1997 RSNA scientific assembly. Received February 17, 1998; revision requested June 16 and received July 10; accepted July 14. Address reprint requests to J.P.H. RSNA,

2 a. Figure 1. Static artifact. This artifact may have a ladybug (a) or dramatic lightning-like (b) appearance. INTRODUCTION Quality assurance refers to all systematic activities undertaken by the breast imaging staff to ensure high-quality mammography. Quality control (QC) more specifically refers to the technical aspects of mammography and is evaluated with the 11 QC procedures outlined in the Mammography Quality Standards Act and performed by the QC technologist (1,2). High-quality mammography demands a team approach involving assessment of mammograms on a case-by-case basis and immediate performance of repeat imaging when needed (3,4). Radiologists and mammographic technologists should be familiar with the numerous artifacts that may create pseudolesions or mask true abnormalities. An artifact is defined as any variation in mammographic density not caused by true attenuation differences in the breast (4). We have also included foreign bodies and implanted medical devices in this definition. Factors that create artifacts may be related to the processor, the performance of the examination by the technologist, the mammography unit, or the patient. In this article, we review a variety of mammographic artifacts, identify their causes, and discuss measures that can be taken to eliminate a given artifact. b. PROCESSOR-RELATED ARTIFACTS All facilities performing mammography are now required to implement daily processor QC to meet federal and American College of Radiology (ACR) criteria for accreditation (2). Processing conditions may vary widely from day to day, which can result in a number of unwanted artifacts if the processor components are not to specifications or are improperly maintained. Processor-related artifacts include static artifact, plus-density (black) or minus-density (white) linear artifacts, water marks, scratches and scrapes, film discoloration, and film mottling. Static artifact may appear as a static ladybug (Fig 1a) or as a dramatic lightning-like artifact (Fig 1b) (5) and is more likely to occur in the winter months when humidity is low. Other causes of static include underreplenishment of processor chemicals, improper film handling, and improper electrical grounding of the processor. The ACR Mammography Quality Control Manual recommends that darkroom humidity be maintained between 40% and 60% year-round 504 Imaging & Therapeutic Technology Volume 19 Number 2

3 a. b. Figure 2. Processor-related plus-density (dark) linear artifacts include entrance roller marks (a) and artifacts caused by inadequate film washing due to improper spring tension in the wash transporter rack (b). Arrows indicate direction of film travel. (2). Measures that can be taken to counteract static include humidity control; installation of static-reducing countertop materials in the darkroom; and use of static discharge systems, which provide a continuous flow of ionized air to reduce static charge (2). Processor-related linear artifacts are varied and can be caused by a number of mechanisms. An important first step is differentiation of such linear artifacts from those related to the mammography unit. The ACR Mammography Quality Control Manual describes a simple test that can be performed to determine the source of a linear artifact. Two films are exposed with the same technique and are then run perpendicular to each other in the processor. Processor-related linear artifacts will be parallel to each other on the two films, whereas artifacts attributable to the mammography unit or related equipment will be perpendicular to each other (2,5). Several processor-related linear artifacts have been identified. Entrance roller marks are wide plusdensity (dark) bands seen on the end of the film that first enters the processor (Fig 2a). Dirt on the rollers can produce minus-density linear artifacts parallel to the direction of film travel (6), whereas excess build-up of developer on the rollers may cause plus-density lines (1). Improper spring tension in the wash transporter rack may also cause plus-density linear artifacts due to inadequate washing of processor chemicals from the film (Fig 2b). Excessive roller pressure may produce linear artifacts perpendicular to the direction of film travel (6). Adequate processor cleaning and chemical replenishment along with proper mechanical adjustment of the processor should prevent or eliminate this type of artifact. Wet roller or surface drying marks (Fig 3) may be caused by worn and uneven or erratically rotating rollers, underreplenishment of processor chemicals, improperly low fixer temperature, improperly high dryer temperature, or March-April 1999 Hogge et al RadioGraphics 505

4 a. b. Figure 4. Water spots with clumping of silver halide granules (arrow). ineffective exit squeegee rollers. Poor dryer ventilation, excessively low dryer or fixer temperature, or depleted levels of fixer solution may cause damp films. Clumping of silver halide granules can occur if the dryer is too hot and water droplets form during the drying process (Fig 4). It is recommended that the dryer temperature be set as low as possible while still producing films that are sufficiently dry when they exit the processor (7). Scrapes and scratches create artifacts that are readily identified and can be caused by a number of factors including dirty or worn rollers (Fig 5), incorrect tension on the drive chains, improperly positioned crossover rollers, misaligned guide shoes (Fig 6) (5), improper alignment of the film on the feed tray, improperly mixed processor chemicals, or a deficient replenishment rate. Incomplete fixing causes fading and brownish discoloration of films (Fig 7). This problem may be caused by loss of circulation in the fixer tank, underreplenished or improperly mixed fixer, or low fixer temperature. A power failure that occurs during processing results in interruption of film travel through the processor. Artifact is produced by prolonged contact between the film and the developer chemicals, which causes overdevelopment and dark- Figure 3. Wet roller or surface drying marks. This wide, irregular and blotchy linear artifact (arrowheads) may be due to a variety of causes (arrow indicates direction of film travel). ening of the affected area of the film. Additional linear artifacts are caused by excessive pressure on the film from the transport rollers (Fig 8). Contaminated developer produces decreased contrast, film mottle, and plus-density linear artifacts (Fig 9). Such contamination most often occurs when the crossover rollers are removed, cleaned, and replaced without simultaneous replenishment of the processor chemicals. The developer chemicals should be replenished whenever the crossover rollers are cleaned (8). 506 Imaging & Therapeutic Technology Volume 19 Number 2

5 5. 6. Figures 5, 6. (5) Severe emulsion pick-off artifact caused by dirty rollers. (6) Scratches (black arrows) caused by misaligned guide shoes (white arrow indicates direction of film travel). Figures 7 9. (7) Fading and brownish discoloration caused by improper developer fixation. (8) Plus-density linear artifacts resulting from power failure. (9) Plus-density linear artifacts and film mottle caused by contaminated developer March-April 1999 Hogge et al RadioGraphics 507

6 a. b. Figure 10. Fingerprints left prior to exposure (arrows in a) and after exposure (b). TECHNOLOGIST-RELATED ARTI- FACTS Artifacts associated with performance of the examination by the technologist include improper film handling or loading, inadequate screen cleaning procedures, errors in use of the mammography unit and related equipment, and positioning and darkroom errors. Grease, food, lotions, and other substances on the hands will cause artifacts when the mammographic film is handled prior to processing. Fingerprints left before exposure will be minus density (white) (Fig 10a), whereas fingerprints left after exposure will be plus density (black) (Fig 10b) (5). Washing the hands prior to film handling and prohibiting food and beverages in the darkroom help prevent fingerprint artifacts. Pressure artifacts can be caused by touching the film with the fingernails or fingertips or by improperly storing film boxes lying flat rather than standing on end. Fingernail marks are seen as plusdensity or dark curvilinear artifacts. Fingertip pressure artifacts appear as minus-density oval artifacts that may be mistaken for a neodensity or a nodular density if seen on only one view (Fig 11). Neither artifact should be present on a repeat image. Improper storage of film boxes may cause a minus-density artifact in the same location on all films from the same box (Fig 12). In such a case, repeat images should be obtained with film from a different box and the box containing the films on which the artifact appears should be discarded. Poor screen-film contact can be due to a number of causes including air trapped between the film and the intensifying screen, improper seating of the film in the cassette, lint, hair, dust, a warped cassette frame, damage to the cassette latches, or a dented intensifying screen (1,4). Poor screen-film contact should be differentiated 508 Imaging & Therapeutic Technology Volume 19 Number 2

7 a. b. Figure 11. Fingertip pressure artifact. (a) Craniocaudal mammogram of the medial left breast shows an ill-defined, oval nodular density in the subcutaneous tissues (arrow). The artifact was seen only on this view; spot compression mammograms of this area (not shown) were negative. (b) Repeat craniocaudal mammogram of the left breast shows no abnormality. Figure 12. Pressure artifacts caused by improper film storage. Curvilinear minus-density artifacts (arrows) appeared in the same location on every film from the same box. from image blur due to motion or inadequate compression. Screen-film contact testing is performed initially for new screens, then semiannually and whenever persistent reduced image sharpness is detected. A copper mesh (40 wires per inch) is used to test screen-film contact as specified in the ACR Mammography Quality Control Manual. Five or less small (<1 cm) areas of poor contact are acceptable. Areas of poor screen-film contact larger than 1 cm that are not eliminated by repeat cleaning are unacceptable, and the cassette should be discarded (2). The most common cause of poor screen-film contact is air trapped between the film and the cassette when the film is loaded (6). The ACR Mammography Quality Control Manual recommends that cassettes be allowed to sit for at least 15 minutes after loading to allow trapped air to dissipate (2). If the cassette is used too quickly after loading, trapped air may cause a focal area of image blur (Fig 13a). This 15-minute waiting March-April 1999 Hogge et al RadioGraphics 509

8 a. c. Figure 13. Poor screen-film contact caused by air trapped between the film and the intensifying screen (arrows in a), improper seating of the film in the cassette (arrowheads in b), and lint (c). period necessitates that most facilities have at least eight cassettes of each size. If the film is not properly seated in the cassette along one edge, a rectangular area of poor screen-film contact will occur (Fig 13b). Lint (Fig 13c), dust, and hair usually cause readily identifiable artifacts with a surrounding area of poor screen-film contact. Occasionally, however, dust or fingernail polish will cause indistinct artifacts that simulate microcalcifications (Fig 14) (3). Because all cassettes are required to have a unique radiopaque identification number, an artifact caused by dust or other material on the intensifying screen may be recognized by noting its recurrence in the same location on films used in the same cassette. If an artifact is suspected, the mammographic view on which the finding is identified should be repeated with a different cassette. Daily darkroom cleaning and weekly cleaning of mammographic intensifying screens are both intended to reduce artifacts caused by debris or dust in the cassettes (2). In addition, b. many facilities restrict the use of fingernail polish and hand lotions by their mammographic technologists. Improper loading of films or the cassette into the mammography unit are common causes of unacceptable mammograms. Because mammographic film is single emulsion, the emulsion 510 Imaging & Therapeutic Technology Volume 19 Number 2

9 Figure 14. Artifacts caused by dirty screens. (a) Minus-density linear artifact caused by fingernail polish on the intensifying screen. (b) Dust artifact simulating microcalcifications (arrows). a. b. a. b. Figure 15. Underexposure caused by improper loading of film into the cassette. The films in a and b were exposed with the same technique (112 mas, 25 kvp); however, the film in a was correctly loaded with the emulsion side in contact with the intensifying screen, whereas the film in b was incorrectly loaded and is underexposed. on the film must be in contact with the intensifying screen for a properly exposed image. If the film is loaded into the cassette upside down so that the emulsion is not in contact with the March-April 1999 intensifying screen, the resultant image will be underexposed (Fig 15). Accidental loading of two films into the same cassette will result in a Hogge et al RadioGraphics 511

10 Figures (16) Grossly underexposed silhouette of the breast caused by accidental loading of two films into the cassette. (17) Artifact caused by improperly loaded film. The film became folded upon itself when it was loaded, resulting in a nonexposed area ( cutoff ), poor screen-film contact, and a linear crease. (18) Artifact caused by improperly loaded film. An irregular light leak (arrows) was produced when the edge of the film was closed in the cassette hinge during loading. grossly underexposed silhouette of the breast (Fig 16). A film that becomes folded upon itself inside the cassette during loading will have an underexposed area with a large linear crease artifact and an adjacent area of poor screen-film contact (Fig 17). Closure of the film bin door on a film can also produce crease marks. Films that have an edge caught in the hinge of the cassette may have irregularly spaced, ill-defined black marks at the edge due to light leak (Fig 18). Errors that occur during the weekly cleaning of intensifying screens and produce artifacts include loading films too soon after the cleaning solution is used and failing to use lint-free materials. Films that are loaded too soon after the screens are cleaned may be covered with a bizarre artifact caused by the removal or clumping of silver halide granules, which form clusters of black specks and white pick-off artifacts where they have been washed off (Fig 19) (5). The ACR Mammography Quality Control Manual Figure 19. Artifact caused by improper loading of film into a cassette that was still damp from cleaning. Bizarre clumping of silver halide granules and corresponding areas of white pick-off artifact are seen (arrows). 512 Imaging & Therapeutic Technology Volume 19 Number 2

11 20a. 20b. 21. Figures 20, 21. (20) Artifact resulting from failure to use a grid. The film in a was exposed with use of a proper grid, whereas the film in b was exposed without the grid in place. Note the degradation of contrast due to increased scatter in b. (21) Double exposure. This error requires that both images be repeated. Many of the new mammography units have built-in safeguard mechanisms to prevent such an error. recommends cassettes be allowed to air dry for 30 minutes after being cleaned with an acceptable screen cleaning solution (2). Failure to use lint-free materials may create more minus-density or pick-off artifacts than were present prior to cleaning (Fig 14). Improper operation of the mammography unit includes failure to use the grid, failure to change the cassette between exposures, and improper loading of the cassette into the Bucky tray. Failure to replace the grid may occur after a magnification view has been obtained because an air gap technique rather than a grid is used to reduce scatter in magnification imaging. Failure to use the grid will result in overall graying of the mammogram and degradation of image contrast secondary to increased scatter on the film (Fig 20) (3). This artifact can be differentiated from poor image contrast due to improper processor chemical maintenance in that it represents a sudden change from mammographic findings seen earlier the same day. Failure to change the cassette between exposures will result in double exposure (ie, superimposition of two images) and necessitates that both views be repeated (Fig 21). Some of the newer mammography units have built-in safeguard mechanisms to prevent double exposure. The cassette can be improperly loaded into the Bucky tray of the mammography unit either upside down, in which case the internal structure of the cassette will appear on the mammogram (Fig 22), or front-back reversed, in which case the identification flash will appear over a portion of the breast tissue (Fig 23). Both errors necessitate repeat imaging. March-April 1999 Hogge et al RadioGraphics 513

12 Figures 22, 23. (22) Artifact caused by improper (upside-down) loading of the cassette into the Bucky tray. The internal structure of the cassette is superimposed on the image. (23) Artifact caused by improper (front-back reversed) loading of the cassette into the Bucky tray. The identification flash is superimposed over breast tissue. Figure 24. Artifact caused by running films too close together through the processor. The two films were ruined when they became completely overlapped and stuck together. 514 Imaging & Therapeutic Technology Figure 25. Film fog artifact (arrows) caused when the technologist briefly turned on the darkroom overhead light before the film had entered the processor completely. Volume 19 Number 2

13 Figures (26) Light leak artifact (arrow) caused by a cracked cassette frame. (27) Unusual light leak artifact (arrows) caused by the cassette being incompletely latched before it was brought out of the darkroom. (28) Film fog artifact (arrows) on a test strip caused by fluorescent fingernail polish worn by the technologist. (29) Film fog artifact caused by a fluorescent bandage worn by the technologist. Note the word RAD exposed on the phantom image. Other technologist-related artifacts include those caused by running films too close together through the processor and by film fog. Running films too close together may cause them to overlap and stick together, potentially ruining both films (Fig 24). Film fog will occur if the film is exposed prematurely (ie, before being processed). It may also occur if there is a light leak in the darkroom, an improper safelight is used, the safelight filter is cracked, improper processor indicator lights are installed, the film is stored at too high a temperature, the film bin is accidentally exposed to light, or the darkroom overhead light is turned on before the film enters the processor completely (Fig 25). The safelight housing and filter should be inspected for any defects, and only the recommended wattage should be used. Use of a safelight with higher wattage than that recommended for the filter may cause more rapid fading of the safelight filter and subsequent film fog (5). The darkroom should be inspected for structural light leaks, which commonly occur at the door, in the ceiling, at fixtures, and around through-the-wall processors. The Mammography Quality Standards Act QC test for film fog is performed initially, semiannually, and whenever bulbs or filters are changed or excessive fog is suspected (2). Light leaks, which are a variation of film fog, may result if the edge of the film is caught in the hinge of the cassette (Fig 16), the cassette is cracked or broken (Fig 26), or the cassette is not securely latched before being brought out of the darkroom (Fig 27). Unusual causes of film fog include fluorescent T-shirts, fingernail polish (Fig 28), and bandages worn by the technologist during film handling (Fig 29). March-April 1999 Hogge et al RadioGraphics 515

14 ARTIFACTS RELATED TO THE MAM- MOGRAPHY UNIT Artifacts related to the mammography unit usually necessitate a service call to repair the malfunctioning component. Such artifacts may be caused by failure of the collimation mirror to rotate out of the field of view during exposure, grid inhomogeneity, failure of the reciprocating grid to move, material that has fallen down into the tube housing, compression failure, improper alignment of the compression paddle with the Bucky tray, and a defective compression paddle. Failure of the collimation or illumination mirror to rotate out of the x-ray beam during exposure produces a large, well-circumscribed oval artifact on the mammogram (Fig 30). The shape and intensity of this artifact vary with the mammography unit being used. Material that has fallen inside the tube housing will create a recurrent minus-density artifact that can be differentiated from pick-off artifact (caused by a dirty cassette screen) by noting that it occurs in the same location on multiple cassettes and has indistinct margins. Such an artifact will appear indistinct due to magnification and geometric blurring caused by the gap between the tube housing and the mammographic film. Grid lines were sometimes seen on the images produced with the stationary grids used on older mammography units. However, stationary grids are no longer recommended for use in mammography; current mammography units make use of reciprocating grids, so that grid lines are not seen unless the grid fails to move and produces inhomogeneous white and black lines (Fig 31). In such cases, repeat imaging is mandatory because calcifications may be either simulated or obscured. An unusually long or short exposure or an improper Bucky timing rate will produce faint grid lines (Fig 32). Recurrence of these lines necessitates a service call to adjust the rate at which the Bucky tray moves. Grid inhomogeneity may be detected on phantom images and may be due to warping of the grid. Improperly aligned compression devices may result in linear or rectangular cutoff due to superimposition of the posterior lip of the compression paddle on the mammogram (Fig 33). A Figure 30. Large, well-circumscribed oval artifact resulting from failure of the collimation mirror to rotate out of the path of the x-ray beam during exposure. similar artifact can be caused by a compression paddle that is too thin and bends under compression (4). The compression paddle and the image receptor should be the same size, and the compression paddle should remain parallel to the image receptor during compression, with the posterior lip of the compression paddle just behind the posterior edge of the image receptor. The posterior lip of the compression paddle should be straight rather than curved, and its vertical component should be high enough to prevent the overlapping of skin or other body parts on the image (2,4). Failure to maintain compression for the duration of the exposure results in image blur due to both motion and inadequate compression. PATIENT-RELATED ARTIFACTS Patient-related artifacts may be caused by motion or by superimposition of objects or substances such as body parts, jewelry, clothing, hair, implanted medical devices, foreign bodies, or substances on the skin. Many deodorants (Fig 34), powders (Fig 35), ointments (Fig 36), and lotions 516 Imaging & Therapeutic Technology Volume 19 Number 2

15 Figures 31, 32. (31) Inhomogeneous white and black lines owing to failure of the reciprocating grid to move. (32) Faint regular grid lines caused by an unusually short exposure. Figure 33. Artifact caused by an improperly aligned compression paddle. Improper alignment of the compression paddle with the Bucky tray causes the posterior lip of the paddle to be superimposed on the mammogram, creating a linear area of cutoff (arrows). This situation may also be caused by a compression paddle that is either broken or too flexible, and bends with compression. Figure 34. Deodorant artifact (arrows). March-April 1999 Hogge et al RadioGraphics 517

16 35. Figures 35, 36. (35) Powder artifacts. The patient presented with new bilateral calcification clusters for which she had been encouraged to undergo biopsy at another institution, although no magnification views had been obtained. On magnification views obtained at our institution, the calcifications proved to be powder artifacts (arrows). (36) Artifact caused by ointment (arrows). contain radiopaque components such as zinc, aluminum, and magnesium, which may simulate calcifications at mammography. Many facilities request that women not wear deodorant or powders on the day of their examination to reduce the occurrence of artifacts. Pharmaceutical skin patches such as nicotine or estrogen patches may create a pseudomass (3) consisting of a poorly defined area of opacity, usually on the mediolateral oblique projection. Such an artifact can be more easily identified when bubbles of air are trapped under the patch (Fig 37). The mammographic technologist is responsible for examining the patient prior to mammography and removing any pharmaceutical patches that will appear in the field of view. Motion artifact is usually due to patient motion during mammography and is more frequently seen on the mediolateral oblique projection because the breast is not supported by the Bucky tray as it is for craniocaudal views (Fig 38). Recurrent motion artifact in multiple patients may 36. indicate that the compression mechanism is not functioning properly. Motion artifact may also result if the cassette drops slightly during a mediolateral or mediolateral oblique exposure. Mammograms with motion artifact must be repeated because calcifications will be obscured (3,8). Motion artifact should be distinguished from poor screen-film contact and from image blur due to poor compression, which is most commonly seen anteriorly in patients with large or thick breasts. An additional anterior compression view should be obtained to complete the study in such cases (5). Most artifacts caused by superimposed objects are readily identified and are easily corrected by repeating the exposure after first making sure that the overlapping object is safely out of the path of the x-ray beam. Such objects may include the opposite breast (Fig 39), shoulder, chin, nose (Fig 40), earlobe, fingers (Fig 41), adipose tissue, 518 Imaging & Therapeutic Technology Volume 19 Number 2

17 Figures (39) Artifact caused by superimposition of the patient s opposite breast and clothing. (40) Artifact caused by superimposition of the patient s nose. (41) Artifact caused by superimposition of the patient s fingers. Figure 37. Artifact caused by a pharmaceutical patch (arrowheads). Note the bubble of air trapped under the patch. Figure 38. Severe motion artifact. Repeat mammography was performed because calcifications and masses would have been obscured. March-April 1999 Hogge et al RadioGraphics 519

18 a. Figure 42. Hair artifact. (a) Curvilinear white lines (arrowheads). (b) Tiny minus-density artifacts (arrowheads) simulating microcalcifications posteriorly on craniocaudal mammograms. The tip of a hair braid was responsible for this artifact. b Figures 43, 44. (43) Artifact caused by clothing (arrowheads) simulates a ruptured saline breast implant. (44) Artifact caused by a pacemaker. 520 Imaging & Therapeutic Technology Volume 19 Number 2

19 Figures 45, 46. (45) Artifact caused by a retained Dacron cuff from a Hickman catheter (marked with a metallic BB). The patient presented with a palpable lump that represented the cuff from a previously removed catheter. (46) Artifact caused by bilateral axillary-femoral grafts. Figure 47. Artifact caused by a fishhook. jewelry, or glasses. Extremely kyphotic patients, patients with paralysis, and uncooperative patients may present positioning challenges. Artifact from hair may be seen as large, curvilinear white lines (Fig 42a) or may simulate calcifications and be more difficult to recognize (Fig 42b). Hair artifact, which is usually seen on the craniocaudal projection, is largely prevented on current mammography units with use of extended face shields. Artifacts caused by clothing or examination gowns are also usually readily identified and corrected but may simulate a ruptured saline breast implant (Fig 43). Implanted pacemakers (Fig 44), catheters, automatic implantable cardioverter-defibrillators, retained Dacron cuffs from Hickman catheters (Fig 45) (9,10), and vascular grafts (Fig 46). Large implanted devices such as pacemakers prevent adequate visualization and compression of adjacent breast tissue, usually in the axillary tail on the mediolateral oblique view. Foreign bodies that may cause artifacts include calcified suture material, fishhooks (Fig 47), transected wires from previous needle localization procedures (11), metallic clips or coils from stereotactic biopsy or embolization procedures (12 14), surgical sponges (15), residual silicone from previous implant rupture (Fig 48), or bullets or metallic fragments from gunshot wounds (Fig 49). Tattoos are usually not visible mammographically, although lymphatic intravasation of the dye used for tattooing may occasionally be seen (Fig 50). CONCLUSIONS Many artifacts are easily identified and do not present a diagnostic challenge. Other artifacts may create pseudolesions or mask true abnormalities. Radiologists and technologists should be familiar with the broad range of artifacts and the measures required to eliminate them. Careful attention to darkroom cleanliness, care in film handling, regularly scheduled processor maintenance and chemical replenishment, daily quality assurance activities, and careful attention to detail during patient positioning and performance of mammography can reduce or eliminate most mammographic artifacts. March-April 1999 Hogge et al RadioGraphics 521

20 Figures (48) Artifact caused by residual silicone from previous implant rupture (arrow). (49) Artifact caused by metallic fragments from a gunshot wound (arrows). (50) Artifact caused by lymphatic intravasation of dye used in creating a tattoo (arrows). The tattoo itself was not visible mammographically. REFERENCES 1. Farria DM, Bassett LW, Kimme-Smith C, DeBruhl N. Mammographic quality assurance from A to Z. RadioGraphics 1994; 14: American College of Radiology. Mammography quality control manual. Reston, Va: American College of Radiology, Eklund GW, Cardenosa G, Parsons W. Assessing adequacy of mammographic image quality. Radiology 1994; 190: Bassett LW. Quality determinants of mammography: clinical image evaluation. In: Kopans DB, Mendelson EB, eds. Syllabus: a categorical course in breast imaging. Oak Brook, Ill: Radiological Society of North America, 1995; Cardenosa G. Breast imaging companion. Philadelphia, Pa: Lippincott-Raven, 1997; Langer TL. QC troubleshooting. In: de Paredes ES, ed. Syllabus: 11th annual update in breast imaging. Richmond, Va: Medical College of Virginia, 1996; Haus AG. Screen-film image receptors and film processing. In: Haus AG, Yaffe MJ, eds. Syllabus: a categorical course in physics. Oak Brook, Ill: Radiological Society of North America, 1993; Vyborny CJ, Schmidt RA. Technical image quality and the visibility of mammographic detail. In: Haus AG, Yaffe MJ, eds. Syllabus: a categorical course in physics. Oak Brook, Ill: Radiological Society of North America, 1993; Ellis RL, Dempsey PJ, Pile NS, Bernreuter WK. Mammography of breasts in which catheter cuffs have been retained. AJR 1997; 169: Beyer GA, Thorsen MK, Shaffer KA, Walker AP. Mammographic appearance of the retained Dacron cuff of a Hickman catheter. AJR 1990; 155: Korbin CD, Denison CM, Lester S. Metallic particles on mammography after wire localization. AJR 1997; 169: Liberman L, Dershaw DD, Morris EA, Abramson AF, Thornton CM, Rosen PP. Clip placement after stereotactic vacuum-assisted breast biopsy. Radiology 1997; 205: Beres RA, Harrington DG, Wenzel MS. Percutaneous repair of breast pseudoaneurysm: sonographically guided embolization. AJR 1997; 169: Fajardo LL, Bird RE, Herman CR, DeAngelis GA. Placement of endovascular embolization microcoils to localize the site of breast lesions removed at stereotactic core biopsy. Radiology 1998; 206: Fornage BD. Sonographic diagnosis of a retained surgical sponge in the breast. J Clin Ultrasound 1987; 15: Imaging & Therapeutic Technology Volume 19 Number 2