Even if the location of a lesion is accurately identified by the mammogram, the relative position of this lesion may change when the patient changes position for the ultrasound. Usually, the mammogram is performed in the upright position and the ultrasound in the supine position. Because the breast is a flexible structure, the breast changes its shape from one position to the other. In the upright position the position of the breast drops due to gravity. Whereas in the supine position, the breast flattens against the chest wall. Compared with other external landmarks, the nipple may be lower in the upright position compared with the supine position. Therefore, a lesion above the nipple in the upright position may shift to the same level as the nipple in the supine position. For example a 10:00 lesion in the upright position may become a 9:00 lesion in the supine position. The larger the breast, the greater the movement of the breast.
Finally, the external position of the handheld transducer does not necessarily correlate with the internal position of the lesion. Even if a linear transducer is used, the examiner commonly uses a variety of angles and hand pressures to optimally visualize the abnormality. Even slight angulation will produce a discrepancy between the position of the sonographic transducer and the actual position of the lesion. Furthermore, transducer pressure may cause the lesion to shift position relative to the nipple. Both of these factors may produce a discrepancy between the mammographic position and the sonographic position.
Because external landmarks are not reliable in cross correlating mammographic/sonographic abnormalities, one should use internal landmarks. However, one must use a technique that addresses the problems listed earlier: (1) limited sonographic field of view compared with mammography, (2) differences between mammographic and sonographic patient position and technical orientation, and (3) nonuniformity of breast anatomy both between individuals and within the same individual.
By using internal breast anatomic landmarks, one immediately addresses the first two problems. If one identifies location by the internal anatomy of an organ, then relating the position of a focal abnormality from the limited sonographic field of view to a wide field of view modality is not a problem. Furthermore, unlike external landmarks, internal landmarks do not shift with body position. For example, in the abdomen, one is commonly challenged with the problem of sonographically deciding whether a small liver lesion previously identified on a computed tomography (CT) scan is cystic or solid. Even though the CT scan is a wide field of view modality, experienced examiners easily sonographically localize the position of the CT lesion. These examiners are able to confidently localize the lesion because they relate the lesion to the internal anatomy of the liver as displayed on CT and then sonographically find the same hepatic location using the same internal hepatic landmarks. Furthermore, this process of anatomic cross correlation would not be different even if the patient were lying prone because unlike external landmarks, internal anatomic landmarks do not change relative to each other.
Even though internal landmarks solve the problems of limited sonographic field of view and positional changes, many examiners are inhibited from using internal breast landmarks as there is great anatomic variation between different breasts. However, whenever one performs a breast sonogram to identify a mammographic abnormality, one should always have the corresponding mammogram. The mammogram provides an anatomic map to the patient's breast. If one is able to cross correlate mammographic structures with sonographic structures, then one may use the mammogram as a sonographic guide to locating the lesion. Therefore, ideally the sonographic examiner should be able to sonographically interpret the mammographic image.
To systematically cross correlate sonography with mammography, one should be familiar with normal breast anatomy. There are seven main sonographically different structures in the breast and chest wall of the average normal 45-year-old woman. These structures from superficial to deep are the following (Fig. 2–4):
1. Skin: The skin is a hyperechoic 3 mm layer on the surface of the breast.
2. Subcutaneous fatty layer: This structure lies under the skin and appears as an anterior hypoechoic layer of tissue, which tends to thicken at the periphery of the breast and is more prominent in the medial portion of the breast.
3. Superficial fascia: The superficial and deep layers of the fascia envelop the breast. The superficial layer is an undulating hyperechoic line within the subcutaneous fat that parallels the skin. The deep layer is an hyperechoic line within the retromammary fat that parallels the anterior chest wall muscles.
4. Cooper's ligaments: The ligaments are curved hyperechoic lines within the subcutaneous fat, which extend from the superficial fascia to the deeper adjacent tissue.
5. Glandular tissue: Normal glandular tissue consists of hyperechoic glandular lobes that are in a radial arrangement around the nipple. Each lobe is in the shape of a prolate ellipse and is surrounded anteriorly by the subcutaneous fat and posteriorly by the retromammary fat. Within the lobes, main ducts originate from the nipple and end in a series of terminal duct lobular units.
6. Anterior chest wall muscles: This layer is formed by the pectoralis minor and major muscles and appears as a hypoechoic solid layer posterior to the retromammary fat.
7. Chest wall: This structure consists of ribs connected by intercostal muscles and covered on the deep surface by pleura.
Figure 2–4. (A). A schematic diagram showing the normal structures of the breast: skin (1), subcutaneous fatty layer (2), superficial fascia (3), Cooper's ligaments (4), glandular tissue (5), anterior chest wall muscles (6), chest wall (7). (B). Radial breast sonogram (10:00 position): Sonographic image demonstrates the normal anatomic structures of the breast: skin (1), subcutaneous fatty layer (2), superficial fascia (3), Cooper's ligaments (4), glandular tissue (5), anterior chest wall muscles (6), chest wall (7). (C). Right MLO mammogram: This is the mammogram of the patient whose sonogram is seen in Figure 2–2B. In the upper outer quadrant, the sonographic image corresponds to the outer edge of the patient's fibroglandular density. Box denotes location of transducer.
Most of these anatomic structures are also generally identifiable mammographically
Rules of Cross Correlation
Once one is familiar with normal sonographic and mammographic breast anatomy, one will be able to anatomically cross correlate the two modalities. If one is not familiar with cross correlation of these modalities, one should not be intimidated by the task of cross correlating sonographic and mammographic structures. A simple way to start is to remember two basic imaging rules. The first rule is that the background breast tissue appearance is similar on the two modalities. This means that normal fibroglandular parenchyma is white (or dense) on mammography and white (or hyperechoic) on sonography. Furthermore, fat appears dark (or lucent) on mammography and dark (or hypoechoic) on sonography (Fig. 2–5). The main exception to this rule is the presence of dilated ducts that are sonographically dark (hypoechoic) and mammographically white (dense). When the breast tissue is filled with dilated ducts such as in ductal ectasias, the breast tissue appears white (dense) on mammography but has numerous linear dark structures (dilated ducts) on sonography.
When one is aware of the appearance of breast tissue with the two modalities, then when one reviews a mammogram one should be able to predict the sonographic appearance of the breast. For example, a mammographically fatty, lucent breast will be sonographically hypoechoic. Conversely, a mammographically dense breast usually sonographically exhibits diffusely hyperechoic fibroglandular parenchyma.
Generally, the breasts of most women are not completely dense or lucent. This system of pattern recognition is even more valuable in these breasts. When a breast has a mixture of tissues, these tissues form unique mammographic structural patterns that may be sonographically reproduced. For example, commonly the mammogram