1. Quality assurance and quality control requirements under Mammography Quality Standards Act (MQSA)
• Tests done by the technologists and the required frequency
Actions required if tests fail
• Tests done annually by the physicist
• Phantom images
• Films labeling
Required under MQSA
2. Indicated imaging algorithms in screening
• Standard views
• Additional images done in women with implants
• Exaggerated craniocaudal views laterally
• Anterior compression views
• Imaging women with small or large breasts
3. Positioning basics
• Mediolateral oblique (MLO) views
• Craniocaudal (CC) views
• Exaggerated craniocaudal views
4. Assessing adequacy of positioning
• MLO views
• CC view
Posterior nipple line measurement
• Exaggerated craniocaudal views
5. Assessing images for motion
6. Artifacts related to
• Software processing
• Standard views
• Viewing conditions
• Approach to lesion detection
8. Evaluation of images
• Technical adequacy of images
• Review images in totality for global changes
Diffuse trabecular or parenchymal changes
Technical factors used for exposure
Focal and global areas of parenchymal asymmetry
• Review images specifically looking for
Masses (correlating size and distance from nipple on CC and MLO views)
• Review images in thirds
CC: lateral, retroareolar, medial
MLO: upper, mid and lower
• Review specific locations
CC: medial quadrants, subareolar, retroglandular fat and tissue interface
MLO: fatty strip between pectoral muscle and tissue, subareolar, upper cone of tissue, tissue immediately superior to inframammary fold (IMF)
CC and MLO: all fat-glandular tissue interfaces
Under the Mammography Quality Standards Act (MQSA) (1), mammographic facilities are required to have a quality assurance (QA) and quality control (QC) program. For digital systems, facilities need to implement the QA/QC program as specified by the manufacturer of the unit(s) being used (1,2). Weekly quality control tests typically include DICOM printer quality control, detector flat-field calibration to assure the system is calibrated properly, artifact evaluation (for the detector), signal-to-noise and contrast-to-noise measurements to assure consistency of the digital image receptor, phantom image to assure overall image quality, compression thickness indicator to assure indicated compression thickness is accurate to ±0.5 cm from actual thickness, diagnostic review workstation QC, and viewboxes and viewing conditions. The visual checklist, repeat analysis, and compression QC test are done monthly, quarterly, and semiannually, respectively. The annual tests done by physicists typically include mammographic unit assembly evaluation, collimation assessment, artifact evaluation, kVp accuracy and reproducibility, beam quality assessment (half-value layer [HVL] measurement), evaluation of system resolution, automatic exposure control (AEC) function performance, radiation output rate, phantom image quality evaluation, signal-to-noise and contrast-to-noise measurements, diagnostic review workstation QC, breast entrance exposure, AEC reproducibility, and average glandular dose. Evaluation of detector ghosting is optional. The reader is advised to consult the equipment manufacturers’ QC manual for details on the specific tests required as well as how each test should be done and interpreted.
For facilities using film-screen systems, QC tests that need to be done by the technologist daily, monthly, quarterly, and semiannually and the annual tests required of the physicist are detailed under the MQSA (1) and the American College of Radiology’s (ACR) Mammography Quality Control Manual (3). The reader is advised to consult this manual for details on how each test is done and interpreted.
A basic understanding of the phantom image in mammography is important since this is used routinely to ensure the entire imaging system (and the processor if film-screen mammography is in use) is operating optimally with respect to uniformity, lack of artifacts, and overall image quality. As part of the QA/QC program, phantom images are done weekly. These are also done after the equipment is serviced, when film or screen type is changed (for facilities doing film-screen mammography) or as needed when troubleshooting problems with the imaging chain. The reader is encouraged to review and understand how phantom images are obtained, evaluated, and scored (ACR’s Mammography Quality Control Manual) (3). The phantom (Radiation Measurement, Inc. RMI 156 or Nuclear Associates 18-220) simulates a 4.2-cm-thick compressed breast composed of 50% glandular and 50% fatty tissue. It contains six fibers (range, 1.56 to 0.4 mm), five specks (1.56- to 0.4-mm diameter range), and five masses (2.00- to 0.25-mm diameter range). An acrylic disc (4-mm thick, 1-cm diameter) is placed, or permanently attached to the phantom.
At a minimum, the four largest fibers, three largest speck groups, and three largest masses should be seen (Fig. 2.1). The image is also reviewed for artifacts and these are factored into the scoring. Regions of interest (ROI) are placed one over the disc and the other in an area adjacent to disc so that signal-to-noise ratios (SNR) and contrast-to-noise ratios (CNR) can be calculated. The measured SNR must be equal to or greater than 40 and the CNR must be within ±15% of the value determined by the medical physicists when the image receptor was installed or after any major upgrade. If the criteria specified by the equipment manufacturer are not met, the service engineer needs to be contacted before any further clinical imaging is done.
Mammography films are legal documents and must be labeled appropriately as required under MQSA (1). The information that must be included on the images is listed in Table 2.1. When printing images from a digital study, please make sure that the required data is not printed over any part of the breasts or pectoral muscles.
Technical factors (kVp, mAs, cm of compression) used for exposure should be available readily so that they can be reviewed in conjunction with the images. As will be discussed and illustrated repeatedly, the technical factors used for exposure (right breast compared with left or from 1 year to another) in some patients provide additional information that can be used in establishing the presence of underlying pathology and formulating differential considerations.
By definition, screening mammography is done in asymptomatic women (4). We recommend annual screening mammography starting at age 40. In women with a strong family history of breast cancer, we start screening annually at age 30. Additionally, if a patient has two first-degree relatives with breast or ovarian cancer, particularly if the diagnosis of breast cancer in the relative occurred premenopausally, she is referred to a genetic counselor for risk assessment. If the risk assessment yields a 20% or higher lifetime risk for breast cancer, we recommend annual screening with mammography and breast magnetic resonance imaging (5).
Unique patient identifying number
Date of study
Radiopaque laterality and projection markers placed closest to axilla
Facility location (minimum: city, state, and zip code)
Mammography unit identification number/room number (if more than one unit/facility)
Craniocaudal (CC) and mediolateral oblique (MLO) views of each breast are done for screening (4). At the discretion of the technologist, anterior compression and exaggerated craniocaudal (XCCL) views are obtained in a small number of women to adequately compress anterior tissue and image posterolateral tissue in the CC projection, respectively. A metallic BB is used to mark any prominent skin lesions that may be mistaken for a breast mass. Our technologists are required to document the location and reason for using skin markers on the patient’s history form. Likewise, they document surgical scars on a diagram of the breast provided on the history form. We do not routinely use radiopaque markers on biopsy scars or the nipples (see excisional biopsy section in Chapter 11 for additional discussion).
In women with implants, we obtain four views of each breast: CC and MLO views with the implants in the field of view and CC and MLO views with displacement of the implants (1,3,4,6). Implant-displaced (ID) views may be difficult (or not possible) to obtain in women with implants that are encapsulated (see Fig. 11.49). If the ID views cannot be done, the technologist documents this in the patient’s history form and the radiologist states this in the breast imaging consultation report. We do not obtain any special consent from patients with implants prior to imaging them.
IMAGING WOMEN WITH SMALL BREASTS (OR MEN)
In women with small breasts, for ID views in some women, or in male patients, it can be hard for the technologist to maintain breast positioning, particularly for MLO views, without potentially hurting her hand and scraping the skin over her knuckles as compression is applied. After positioning, the technologist holds the breast in position, and as the paddle comes down on the breast, she needs to pull her hand out from under the paddle. This maneuver can be a challenge when using the standard compression paddle in these groups of patients. To overcome these potential limitations, a compression paddle that is half the width of the standard compression paddle is available (Fig. 2.2A); this paddle is also sometimes helpful in obtaining optimal axillary views.
IMAGING WOMEN WITH LARGE BREASTS
For most digital systems, compression paddles (Fig. 2.2B, C) with resulting collimation are available in two sizes (18 × 24 cm and 24 × 30 cm). The selection of which size to use is made based on the woman’s breast size. The large paddle and the resulting larger field of collimation is not used for women with small breasts. Likewise, the small paddle and small field of collimation is not used on a woman with large breasts. Some women with large breasts may need more than two views of each breast to image all the breast tissue adequately in the two standard projections (Fig. 2.3).
MEDIOLATERAL OBLIQUE VIEWS
Factors to be considered in obtaining optimal positioning of the breasts on MLO views are listed in Table 2.2. If these simple concepts are followed routinely, a maximal amount of tissue is included on the images while compression-related discomfort is minimized.
In positioning patients for MLO views, four separate concepts need to be considered, understood, and applied consistently: selection of a patient-specific angle of obliquity, relaxation of the pectoral muscle, medial mobilization of the breast and underlying pectoral muscle, and the need for an out and upward pull of the breast tissue and underlying muscle. The breasts are skin appendages. In trying to maximize the outward pull of a skin appendage so it can be imaged, it is best to pull the appendage out parallel to the underlying muscle; the pectoral muscle is the muscle underlying the breast. Consequently, the technologist needs to assess the orientation of the pectoral muscle for the individual patient (Fig. 2.4) and use this to determine the angle of obliquity for the MLO view. By necessity, the angle will be different for different patients. Tall women will have a more vertically oriented pectoral muscle compared to short women. In an effort to include as much tissue as possible on MLO views, the breasts should be pulled away from the body parallel to an underlying relaxed pectoral muscle (7).
Technologist works from behind and the medial side of patient
Angle of obliquity
Specific for each patient
Determined by technologist based on the obliquity of the pectoral muscle
Relaxation of the pectoral muscle
Inward rotation of humeral head
Ipsilateral arm down
Medial mobilization of breast tissue and pectoral muscle
Need to maintain medial mobilization
Minimizes skin stretching in the upper inner quadrant
Patient needs to stay in unit (tendency is to pull back out)
Breast pulled up and out
Include a small amount of abdomen
In addition to selecting an appropriate angle of obliquity, the technologist needs to make sure that the pectoral muscle is relaxed; otherwise, she may be fighting the resistance added by a tense muscle. The pectoralis major muscle inserts on the upper third of the humerus; so the muscle tenses if the arm is elevated and the humeral head is externally rotated. During positioning, the patient’s arm should be kept down (behind or on the bucky) while making sure that she inwardly rotates her humeral head. The amount of tissue and muscle included on the images is increased if the pectoral muscle is relaxed.
With respect to the importance of breast mobility, consider the basics of breast anatomy. The lateral and inferior aspects of the breasts are the most mobile (Fig. 2.5A). Tissue in the upper and particularly inner quadrants has little inherent mobility. Unfortunately, most of our equipment is designed so that, for MLO views, the compression paddle travels from the upper inner quadrant toward the lower outer quadrant (Fig. 2.5B). As the compression paddle is engaged, we are attempting to mobilize tissue with little inherent mobility in the upper inner quadrants into the field of view. Inevitably, as the compression paddle moves, tissue and potentially lesions roll out from under the paddle and are excluded from the field of view. The movement of the compression paddle over fixed tissue in the upper inner quadrant of the breast also results in skin stretching that almost certainly accounts for much of the discomfort patients associate with mammography and attribute to breast compression. If we use natural breast mobility effectively, we can minimize the amount of tissue that is excluded as well as the amount of skin stretching (7). Every 1 mm that we are able to mobilize tissue and underlying muscle medially represents 1 mm less of the compression paddle travelling over fixed tissue. In addition to mobilizing the breast and muscle medially, the breast tissue needs to be actively pulled out away from the body to include a maximal amount of posterior tissue, and upwardly, so that the inframammary fold (IMF) is opened. Ideally, a small amount of upper abdomen is included on the image.
The tip of the digital detector is positioned at the apex of the axilla so that it is snug against the body along the mid-axillary line. As the patient is rotated toward the digital detector and the breast is mobilized, the technologist needs to smooth the tissue going up against the detector; otherwise, skinfolds can develop (Figs. 2.6 and 2.7A). Also, there should be no air gap (Fig. 2.7A) between the pectoral muscle and upper portion of the breast and the digital detector. If the air gap is not cleared, uneven exposure (overexposure) of the pectoral muscle and axillary tissue will be apparent (Fig. 2.7A, B). Last, there should be no space between the digital detector and the breast posteriorly (Fig. 2.7C); otherwise, posterior tissue is excluded. The technologist should not be able to advance her index finger into the apex of the axilla.
If the correct angle of obliquity for the individual patient is selected, the pectoral muscle is relaxed, the breast is medially mobilized and maintained there as the breast is pulled up and out, the amount of tissue included on the MLO view is maximized, and exclusion of tissue, particularly upper inner quadrant tissue, is minimized. In assessing the adequacy of positioning on MLO views, the interpreting radiologist should consider the factors listed in Table 2.3. Consider the length and shape of the pectoral muscle on the MLO views (8). The pectoral muscle should be thick (wide) at the axilla, have an anterior convex margin, and extend to the level of the nipple (Fig. 2.8). Selection of an inappropriate angle of obliquity, inadequate relaxation of the muscle, failure to medially mobilize or maintain medial mobilization of breast tissue, or allowing the patient to lean back slightly can result in a concave pectoral muscle edge, a triangular pectoral muscle, or a muscle that is parallel to the edge of the film (Fig. 2.9). The IMF should be open with a small amount of abdomen included on the image. The technologist should exercise care to not include too much abdomen, or allow a skinfold (Fig. 2.6) to develop up against the bucky laterally, since this may limit compression. A small triangular density superimposed on the pectoralis major muscle in a small number of women is the pectoralis minor muscle (Fig. 2.10).
Wide (thick) pectoral muscle (PM) at the axilla
PM to level of nipple
Convex anterior margin of PM
Breast pulled up and out (no sagging)
Small amount of upper abdomen
In some women, positioning may be limited secondary to a physical disability such as kyphosis, paraplegia, a frozen shoulder, Parkinson disease, or absence of the pectoral muscle (Poland syndrome) (9) (Fig. 2.11A, B). In these situations, we work closely with the patient to obtain the best images possible and document the limitations and our efforts to obtain adequate images on the patient’s history form. If the limitations are significant, we describe this in our breast imaging consultation report. In contrast to patients with Poland syndrome in whom there is absence of a pectoral muscle, rarely, in patients with a history of prolonged steroid (Fig. 2.11C) exposure, or those with disuse (e.g., paralysis of the ipsilateral arm, polio) of the chest wall musculature (Fig. 2.11D), it may appear as though there is no pectoral muscle, yet on close review of the images, the “ghost” of a pectoral muscle is apparent as the striations of the pectoral muscle are seen in an otherwise fatty replaced muscle.
Factors to be considered in obtaining optimal positioning of the breasts on CC views are listed in Table 2.4. If these concepts are followed routinely, a maximal amount of tissue is included on the images and compression-related discomfort is minimized.
In positioning patients for CC views, several concepts need to be considered, understood, and applied consistently: upward mobilization of the breast to the extent permitted by the natural mobility of the IMF, placement of the imaging receptor up against the chest wall, positioning of the contralateral breast on the image receptor, outward pull of the breast, and the lateral tug. In considering the CC view, keep in mind that the breast is mobile inferiorly while upper tissue, particularly close to the chest wall, is relatively fixed in position. Yet, when the compression paddle is engaged, it moves over upper inherently fixed tissue posteriorly in an attempt to mobilize it inferiorly. The technologist should identify the natural position of the IMF as she holds and lifts the breast as far as the natural mobility of the IMF allows. If the technologist is able to mobilize the breast upward to meet the paddle, exclusion of upper posterior tissue can be minimized. Every millimeter the breast is moved upward to meet the paddle is 1 mm less that the paddle will need to move down over fixed tissue (7); this results in inclusion of a maximal amount of tissue and minimizes skin stretching and the associated discomfort. The image receptor should not be placed higher than the elevated IMF position; otherwise, posterior tissue is excluded. As the breast is lifted, the technologist needs to be careful so that a skinfold doesn’t develop between the inferior aspect of the breast and the IMF or that a portion of abdomen is not included on the image. When present, this skinfold (or abdomen) can simulate the pectoral muscle. A sharp lucency (air) outlines the edge of a skinfold but is not seen associated with the pectoral muscle (Fig. 2.12). Also, as the breast is lifted upward it also needs to be actively pulled out away from the body.
The position of the contralateral breast needs to be considered. If the contralateral breast is left pendulous, it can become an impediment to the digital detector being placed against the chest wall. The technologist needs to place the contralateral breast on the edge of the imaging receptor so that the detector can go up against the chest wall (Fig. 2.13A). The technologist should check to make sure there is no space (air gap) between the chest wall and digital detector; she should not be able to put her index finger through the cleavage (Fig. 2.13B).
The inclusion of medial tissue on CC views needs to be emphasized; however, as much as 2 cm of lateral tissue can be pulled into the image without giving up medial tissue (Fig. 2.14). The lateral aspect of the breast is mobile and has to be pulled actively into the image by the technologist as compression is applied. As the lateral aspect of the breast is tugged, a skinfold often develops laterally. This skinfold can sometimes be rolled out if the technologist puts her finger under the paddle and rolls it out with the skinfold.
Technologist works from medial side of patient
Neutral position of IMF is identified
Breast tissue is lifted to the extent of natural mobility of IMF
Pull breast tissue out
Lateral tug: active pull of lateral tissue into field of view
Contralateral breast up on digital detector (not left pendulous)
Detector to chest wall
In women with no history of breast surgery the nipple should point straight out
Suspect medial exaggeration if nipple is pointing laterally
Suspect lateral exaggeration if nipple points medially
In women who have had no breast surgery, the nipple can be used as an indicator of unwanted exaggeration on CC views. In an optimally positioned CC view, the nipple should point straight out. If the nipple points medially, there may be lateral exaggeration and similarly, if the nipple points laterally, medial exaggeration may be present. Caution should be exercised, however, because in women with significant tissue mobility laterally, a routine CC may seem exaggerated laterally when it is not; the apparent exaggeration laterally is a reflection of an optimal lateral tug.
In assessing the adequacy of positioning on CC views (Table 2.5), pectoral muscle should be seen in 30% to 40% of patients (7,8) (Fig. 2.15A); in younger women, the pectoral muscle may be imaged more commonly than in older patients. If the pectoral muscle is not seen, look for cleavage (Fig. 2.15B). If pectoral muscle or cleavage is seen, you are assured that medial tissue has not been excluded from the image; if neither is seen, consider measuring the posterior nipple line (PNL) (see below). Laterally, retroglandular fat should be seen; if tissue is seen to the edge of the film, an XCCL view may be indicated.
In some women, the pectoral muscle is atrophied and only the sternal insertion is seen medially on the CC view. The rounded or triangular appearance of the muscle in these women can simulate the anterior portion of a mass (Fig 2.16) (10). Alternatively, a round density partially seen on the CC view medially superimposed on the pectoral muscle (with no correlate on the oblique view) can represent the sternalis muscle (Fig. 2.17; also see Fig. 10.1). Even with spot compression views, it is usually only partially seen. The sternalis muscle is an uncommon normal variant in chest wall musculature; it is the remnant of a muscle that would extend from the infraclavicular region to the caudal aspect of the sternum (11). It is commonly a unilateral finding. When the MLO view is reviewed in these patients, no abnormality is seen.