The quality of mammographic images can be assessed based on eight criteria: positioning, compression, exposure, contrast, sharpness, noise, artifacts, and labeling. 1 Among women 40 or more years old, periodic screening mammography may reduce deaths due to breast cancer by 50%. Therefore, the level of image quality attained for screening mammography is clearly of lifesaving importance. 2
THE 8 CRITERIA
1. Positioning . Adequate positioning of the mediolateral oblique (MLO) view may be assessed by means of the posterior nipple line (PNL), a line drawn from the nipple at an upward angle perpendicular to the edge of the pectoral muscle (Figure 1). If the breast has been adequately positioned for the MLO view, the posterior nipple line should reach the tip of the pectoral muscle in 80% or more of cases. Depending on the anatomy of the individual patient, the angle at which the MLO view should be obtained will vary from 30° to 60° to the vertical axis. The breast compression plate should be aligned perpendicular to the long axis of the pectoral muscle; otherwise, the muscle may resist the compressive force, and the posterior breast tissue will pull back from the image receptor.
|Figure 1. (left) Adequate positioning of the mediolateral oblique view can be assessed by means of the posterior nipple line. (Click the image for a larger version.)|
To provide a reference for future studies, the technologist should record the angle of compression. This practice will also encourage the technologist to adjust the angle to suit different patients. The pectoral muscle should not have either a straight or a convex anterior border and should be wider at the top than at the bottom. Pulling the breast upward and outward as far as possible before compression is applied (Figure 2) will open the inframammary fold and prevent the overlapping and poorly compressed tissue that occurs when the breast is allowed to sag.
Even with proper positioning, the pectoral muscle will be seen in only 30% to 40% of cases in the craniocaudal (CC) view. When the muscle is not visible, the posterior nipple line can be measured from the posterior edge of the nipple to the margin of the film. With adequate positioning, the length of the PNL on the CC view should be no more than 1 cm shorter than the length of the PNL on the MLO view.
|Figure 2. Proper positioning will open the inframammary fold and prevent the overlapping and poorly compressed tissue that occurs when breast is allowed to sag.|
The posterior medial breast cannot be seen at the MLO view, so it is vital to include this tissue in the CC view. Proper CC view positioning will allow inclusion of this area without rotating the breast medially or laterally. For both MLO and CC views, skin folds should be minimal or absent. To include as much upper posterior breast tissue as possible on the CC view, the breast must be elevated as far as the inframammary chest-wall attachment allows before compression is applied.
2. Compression . Compression allows sharper images by bringing the breast closer to the film, immobilizing the breast to prevent motion, and reducing breast thickness to permit shorter exposure times. By bringing the breast closer to the film, compression serves to decrease geometric lack of sharpness. Compression enhances contrast by reducing scattered radiation and permitting use of a lower kilovolt peak (kVp) beam. By reducing breast thickness, compression also allows a lower radiation dose. If the breast were not adequately compressed, film density would not be uniform, since thinner anterior tissues would be overpenetrated (compared with thicker posterior tissues). Cysts and normal glandular tissue are more easily compressed than carcinoma, so compression helps the radiologist assess the density of masses more accurately. Lesions are also more readily seen because compression separates superimposed breast tissues.
If compression is adequate, CC and MLO views should exhibit the same degree of separation for breast structures. Inadequate compression can be detected as motion, underexposure on a single view, or nonuniform exposure levels. For the MLO view, compression will be impeded if the breast has not been lifted sufficiently or if shoulder or arm tissue has been included. The breasts of patients with well-developed pectoral muscles will be more difficult to compress.
3. Exposure . Three prerequisites must be in place before the adequacy of exposure can be evaluated. They are low ambient room light, sufficient view-box luminance, and the masking of the lighted area of the view box that is outside the margins of the film. Especially in dense breasts, underexposure is far more common than overexposure. In breasts of all types, there must be sufficient penetration of the pectoral muscle for the underlying breast tissue to be seen. High illumination can be employed to compensate for overexposure, but repeat imaging will be necessary in cases of underexposure.
Low exposure levels may be caused by inadequate compression, processing problems (such as stale, contaminated, or diluted developer; inadequate replenishment rates; or developer temperatures that are too low), placement of the photocell over fatty (rather than fibroglandular) tissue, a phototimer density setting that is too low, poor phototimer function, a low mA output, or a kVp that is too low.
4. Contrast . In mammography, high contrast is of extreme importance. Breast tissues have a limited range of x-ray attenuation, so contrast must be high to permit their differentiation. The first steps in achieving high-contrast mammograms include selection of a high contrast film screen system, use of a low kVp beam and a mammographic grid, dedicated mammography film processor, and a suitable x-ray tube target/filter combination such as molybdenum/molybdenum. Excessive contrast is not helpful, however, because it can prevent the acquisition of a single image on which both thinner (fatty) and thicker (dense) breast tissues can be visualized adequately.
Poor contrast levels may be caused by kVp that is too high, insufficient compression, lack of a grid, processing deficiencies (stale, contaminated, or diluted developer; inadequate replenishment rates; or low developer temperature), low film gradient, inappropriate target material and/or filtration, or underexposure.
5. Sharpness . Unless images are sharp, it may not be possible to visualize microcalcifications, the margins of masses, or fine spiculations. Localized areas of lack of sharpness due to the spreading of light can be caused by poor contact between film and screen (possibly due to damaged cassettes, improper placement of film in the cassette, or air trapping). Between the loading of a cassette and its exposure, at least 15 minutes should be allowed to elapse in order to eliminate air trapping.
An image may blur because the focal spot is too large or is damaged, the target-to-film distance is too short, the film-screen system resolution is inadequate, or the patient has moved. The patient’s motion could indicate too long an exposure, poor compression, or simply failure to hold her breath.
6. Noise . Radiographic mottling, or noise, is largely due to the statistical fluctuation of photons absorbed in different areas of the intensifying screen (quantum mottling). Increased noise can be due to underexposure and/or to the use of excessively fast, low-dose film-screen systems. Noise can interfere with the radiologist’s ability to see tiny structures such as microcalcifications.
7. Artifacts . There are several types of errors that can produce artifacts in completed images. Dust may be the culprit, whether because the screens have not been cleaned thoroughly or because the darkroom needs to have dust removed. Processor problems can create scratches or roller marks, and grid malfunctions can produce visible grid lines. Misalignment of the cassette and the x-ray beam or the lip of the compression plate can produce band-like artifacts. Light that is transmitted through clear areas at the edges of the film will have an adverse effect on visual perception. Therefore, the x-ray beam must be collimated to the edges of the film.
8. Labeling . The Mammography Quality Standards Act (MQSA) includes several film-labeling requirements. A permanent identification label is required for each film; it must contain the imaging facility’s name, address, and zip code, as well as the patient’s first and last names. A unique patient-identification number must be used. This can be a medical record number, Social Security number, or date of birth. The examination date must also appear on the label.
Near the aspect of the breast closer to the axilla, radiopaque markers must be placed to indicate whether the view is MLO or CC, and whether the right or left breast is shown. Unique initials must be used to identify the technologist who performed the study; these can be part of the patient-identification label or can be placed on the cassette holder using radiopaque letters. An Arabic numeral written or pressed onto the screen is used to identify the screen (so that defective or artifact-producing screens can be identified). A Roman numeral is used to identify the mammography unit. Obviously, this step is unnecessary if the facility has only one unit.
Patient-identification systems using flash cards are strongly recommended, since these are the most permanent form of identification. Adhesive labels will not be reproduced when films are copied. Date stickers, color-coded by year to make examination sorting easier, are helpful because they permit the study date to be read in overhead light. While they are not required, technical factors appearing on the film can be helpful; these include kVp, milliamperes (mA), exposure time, compression force, compressed breast thickness, and degree of obliquity.
Stephen A. Feig, MD, is professor of radiology, Mount Sinai School of Medicine of New York University, and director, breast imaging, Mount Sinai Hospital, New York. This article has been excerpted from Assess and Improve Your Mammographic Image Quality—Every Film, Every Day, which he presented at the 90th scientific assembly and annual meeting of the Radiological Society of North America on November 29, 2004, in Chicago.
- American College of Radiology Committee on Quality Assurance in Mammography. Mammography Quality Control Manual. Reston, Va: ACR; 1999.
- Feig SA. Screening mammography: effect of image quality on clinical outcome. AJR Am J Roentgenol. 2002;178:805-807.