Chapter 42: ADVANCED BREAST ULTRASOUND

Ultrasound detects breast lesions by depicting their structural, dynamic, and vascular differences from the surrounding normal tissue.^{1,2,3,4,5, and 6} The structural features are depicted on gray scale; the dynamic, traditionally, by probe palpation and attempts to move the lesion with a free finger, or better, using elastography (see Figure 42-1B); and the vascular with Doppler, perhaps enhanced by contrast agents. Few lesions larger than 1 cm in diameter lack all of these distinctions (these are the undetectable "non-mass" lesions), but for small lesions (<0.5 cm), the changes are often too subtle to be detectable reliably or are completely absent so that the differential diagnostic features may disappear.⁷ These form the limits of sensitivity and specificity of ultrasound, which are being improved continuously as better scanners and radically new techniques become available.^8,9

High-frequency (at least 7 MHz and preferably 12 MHz) linear array transducers are the most suitable for breast scanning, whereas hybrid mechanical and electronic systems are used for three- and four-dimensional (3D/4D) scanning.

Deformability, which is related to the elastic properties of tissues, is part of the algorithm used to determine the likelihood of malignancy in routine breast ultrasound. Visual comparison of tissue images before and after deformation lacks sensitivity except at extreme differences in elasticity. Analysis of the raw ultrasound image data, which contains very much more information than can be displayed for visual perception, can detect very small deformations. Measurement of tissue deformation (strain) with applied force (stress) by analysis of the returning ultrasound echoes (elasticity imaging) was described over 25 years ago^10,11 and has been developing rapidly over the last 2 decades.^{12,13, and 14}

When tumors grow in or infiltrate through tissues, they increase the tissues' stiffness or hardness. The very high sensitivity of ultrasound to changes in stiffness allows elasticity imaging to show changes that are not visible on conventional ultrasound imaging. Pathology that has the same B-mode appearance as the surrounding tissues can be identified because it is very slightly stiffer, even though it cannot be felt with the examining hand. The elasticity of different tissues is determined by comparing whole-frame raw ultrasound data (data before demodulation) while a force is applied to the tissue. Changes in radiation force (RF) data from frame to frame are caused by distortions that depend on differences of stiffness in the tissues. Stiff tissues distort less than soft tissues for a given force, and the variation in the amounts of distortion detected in the RF data provides information on the elastic properties of the tissues under investigation.

In freehand elastography, the tissues are compressed with the transducer while the raw data are analyzed and compared with previous image frames. The same region of interest is kept within the frame while deforming the tissues with the probe. A variety of methods has been developed to measure the displacement.¹⁵ One tracks the small displacements (0.5% to 1%) between frames, another uses autocorrelation with phase domain processing (similar to pulsed Doppler),¹⁶ and another uses tissue Doppler shift changes at low frequencies to measure stiffness during active compression.¹⁷

A different approach is used in acoustic radiation force imaging (transient elastography): focused ultrasound creates an impulsive force to produce a transverse shear wave.^18,19 Ultra-fast ultrasound imaging detects the minute tissue movements produced, allowing calculation of the velocity of the shear wave which is proportional to the coefficient of elasticity of the tissue.

Regardless of the method used, the amount of movement is related to the stiffness of the tissues. Differences in movement enable a strain map of comparative elasticity to be constructed so that the difference in stiffness between fat, Cooper's ligaments, and normal parenchyma can be appreciated. In general, benign lesions have low stiffness values while malignant lesions are stiffer than the surrounding breast.

The breast is a layered structure that lies between the superficial and deep layers of the superficial fascia.²⁰ The skin and the fibroglandular plate (the breast parenchyma) are relatively echogenic, whereas the subcutaneous and retromammary fat layers are echo poor. The glandular layer is thicker in the upper outer quadrant (Figure 42-1). The chest wall muscles are also echo poor in younger women, but often become more echogenic with the fatty infiltration of age. Cooper's suspensory ligaments form echogenic bands that pass from the deep layer of the superficial fascia to the underside of the skin. The lactiferous ducts are widest at their termination under the areola and may be traced out toward the periphery of the gland as fine echo-poor lines. Scans in radial projections, aligned with the ductal system, reveal the lobar anatomy of the breast and any lesions it contains.

The hormone-induced changes in the breast produce wide variations in the relative sizes of the layers. Fat may be absent in adolescents and young women, and also during pregnancy and lactation when the glandular layer hypertrophies. With postmenopausal atrophy, the glandular layer becomes thinner and is partly or almost completely replaced by fatty tissue.

On elastography, normal parenchyma is stiffer than fat, and Cooper's ligaments show up as linear regions of higher stiffness.

On color Doppler, normal vessels are seen, typically running along normal structures such as the Cooper's ligaments, with a linear or smoothly curved profile; they taper smoothly and branch regularly. Three sets may be recognized: branches from the lateral thoracic artery running in the anterior part of the breast, branches from the internal mammary artery passing laterally into the breast from the parasternal region, and branches of the intercostal arteries penetrating the posterior surface of the gland.

Benign lesions are generally simple in structure, with smooth borders and uniform internal echoes; they do not disrupt the surrounding structure of the breast.⁴ They distort or move freely through the tissues on probe compression or palpation, and are usually poorly vascularized with simple vessel architecture. They show low elastography scores using the Ueno/Tsukuba scoring method,^{21,22, and 23} and appear smaller or the same size on B-mode.

Malignancies, by contrast, are generally complex or chaotic in structure, with irregular, poorly defined borders and an infiltrating edge.^24,25 Their internal echoes are heterogeneous and, if fibrous, they cast an acoustic shadow that is also irregular in its intensity. They interrupt the surrounding breast architecture and, because they infiltrate, they drag the surrounding tissues with them as they are moved. They are hard and appear larger on elastograms than on B-mode.²⁶ Usually they are vascular and the vessels are tortuous and irregular, penetrating radially into the lesion.^27,28 Sometimes intervascular shunts can be demonstrated.

These appearances are summarized in more detail in the BIRADS Lexicon of breast ultrasound descriptors which is used in the differentiation of benign and malignant lesions. The BIRADS system is used to grade lesions as Category 1, normal; 2, benign; 3, probably benign finding; 4, suspicious abnormality; 5, highly suggestive of malignancy. The classical ultrasound features are described and important exceptions are discussed below for each type of lesion.

Typical cysts are well-defined, echo-free spaces with smooth walls and increased through-transmission of sound ("acoustic enhancement") (Figure 42-2).³⁰ They are usually rounded but may be flattened when lax, and they often form in bunches. Because they arise from the ductal and glandular structures, they are partly attached to the surrounding breast tissue and thus are not very mobile. They are always avascular. The elastographic appearances of cysts depend on the technique used. With compression elastography, the variation of viscosity and particulate content in cysts results in differing behavior of the returning signals and thus various elastographic patterns. Movement of reflectors in cyst fluid is caused by probe motion or acoustic streaming forces rather than stiffness. Elasticity algorithms may interpret small movements of cyst contents as if they are strain deformations, producing artifacts (see Figure 42-1B). True fluids do not transmit shear waves like solid tissue, and so appear as signal voids on transient elastography; however, many cysts contain viscid material and so may show shear wave stiffness at low levels.

High-resolution ultrasound can detect cysts as small as 1 mm in diameter with confidence, and it is often a surprise how many more cysts are detected than can be palpated or are visible on mammography. Provided all the ultrasonic criteria are fulfilled, the diagnosis is completely reliable and no further investigation is required, although the fluid may be aspirated to relieve pain or to reassure the patient or the physician in charge.

Any cyst that departs from these strict criteria must be needled and the specimen sent for cytology because it may represent an intracystic tumor or, more rarely, a necrotic carcinoma or one that, by chance, has arisen adjacent to a cyst. Thickening or irregularity of the wall is particularly worrying, as is vascularization. These appearances can be mimicked by bleeding into a cyst after needle aspiration: ultrasound (and mammography) should precede needling.³¹ Some simple cysts do return low-level echoes because they contain particulate matter such as cellular debris, cholesterol crystals, oil/water emulsions, or blood; typically, these echoes are very uniform. Sometimes the echogenic material within cysts can be made to move when Doppler is turned on in their region. Such moving particles may produce color-flow signals, even when they are not visible on the B-mode scan.

The pathology of benign breast change (fibrocystic dysplasia, anomalies of normal development and involution [ANDI]) is very varied and this leads to a spectrum of ultrasonic appearances.³² Often the scan is normal or simply shows regional thickening of the gland, most commonly in the upper outer quadrants, that corresponds with the thickening or tenderness. This in itself can be helpful, especially if the patient, who is often understandably anxious, can be shown that the symptomatic region has the same appearance as the rest of the breast. There may be generalized or segmental ductal changes in which the ducts become prominent, sometimes with a fine, echo-poor cuff surrounding them. This produces a sponge-like or mottled appearance, which is not pathologic and is frequently encountered in asymptomatic young women. Benign breast change usually has a very similar elasticity pattern to the surrounding breast tissue, which is reassuring. Fibrosis or scarring within an area of benign breast change may, however, cause some increase in stiffness, but often this is not very great.

None of these appearances is worrying, but localized ANDI changes, particularly when there is fibrosis, can produce a poorly defined, incompressible, echo-poor region with distal shadowing that suggests malignancy (Figure 42-3). Differentiation using ultrasound may be impossible but, because the lesion is not a true mass, it is traversed by ductal structures, whereas they are deviated or interrupted by a cancer. If such ducts can be demonstrated, the appearance is reassuring. A feature that may also be helpful is the fact that such "benign shadowing" reduces or disappears when the lesion is compressed by the probe, in contradistinction to shadowing deep to malignancies, which persists.³³ Absence of vascularity on color Doppler is another typical feature, although when benign breast change is accompanied by inflammation, an increase in vascularity may occur.

Most fibroadenomas have typical appearances as well-defined lesions with an oval or occasionally macrolobulated shape and uniform internal echoes of about the same intensity as the subcutaneous fat (Figure 42-4). Because fibroadenomas only attenuate if they are calcified or have undergone hyaline necrosis, there is usually increased through-transmission of sound, which is obvious as intensely reflective bands deep to the lesion. There is no reaction from the surrounding tissues, with the mass seeming to float within the breast tissue. Their mobility is an obvious ultrasound feature elicited by palpation with a free finger alongside the transducer. Fibroadenomas tend to lie with their long axis parallel to the skin as they accommodate themselves to the local anatomy defined by the breast lobules and Cooper's ligaments.

Typically, they have a smaller area of stiffness (strain footprint) than the B-mode lesion. They may be as soft as the surrounding tissues or be slightly stiffer. Absence of stiffness or a small strain footprint is supportive of a lesion being a fibroadenoma when it is surrounded by breast parenchyma. However, fibroadenomas detected on screening mammograms are usually surrounded by fat, so they may be difficult to identify on B-mode because they are often isoechoic with fat. Elasticity imaging can detect such lesions because they are stiffer than fat lobules.³⁴ When there is clinical concern, they can be biopsied using ultrasound instead of stereotactic guidance.

Although small fibroadenomas are hypovascular, larger and fast-growing fibroadenomas may be moderately vascular, with 1 or 2 vessels of simple arrangement often running parallel to the margins of the lesion.^6,35

The distinctive features of classical fibroadenomas usually allow a confident diagnosis, but with very small lesions, where the appearances of carcinomas may converge, there is often some doubt. Occasionally, the uniform echoes of an inspissated cyst may falsely suggest a fibroadenoma, but its spherical shape and immobility are distinguishing features.

In some centers, ultrasonically typical fibroadenomas are managed conservatively with fine-needle aspiration (FNA) or core biopsy (to evidence their benign nature), or with serial ultrasound surveillance, or sometimes both, provided the patient is comfortable with retaining the lesion.³⁶

Phyllodes tumors, which can have a very similar B-mode appearance to fibroadenomas, are usually stiffer than the surrounding tissues and have a strain footprint that is as large as or slightly larger than the B-mode abnormality. Their size and vascularity may be suggestive, especially when considered in relation to the patient's age (giant fibroadenomas mainly occur in younger women), and the otherwise uniform internal echoes are often broken up by fluid spaces.

Carcinomas have a range of ultrasound appearances, but the most typical is a mass with an echo-poor central nidus (less reflective than fat), often somewhat heterogeneous, with distal shadowing that is often intense and characteristically variable across the lesion.^24,37 The nidus tends to lie with its long axis anteroposteriorly, reflecting its invasive behavior as it grows through the layers of the gland and, for the same reason, it tends to disrupt the adjacent fibroglandular structures, though these features are less obvious with subcentimeter lesions (Figure 42-5). They have poorly defined outer margins, which have a microlobulated contour, perhaps the single most reliable sign of malignancy.

Around the nidus, an echogenic "halo" representing the infiltrating edge of the malignancy may be seen. Typically, it is more obvious at the sides of the lesion than anteriorly, probably because of the different orientation of the invading columns with respect to the direction of the ultrasound beam. The halo is more difficult to detect when the surrounding tissue is itself echogenic, as occurs in premenopausal women. It is often easier to appreciate with real-time scanning because it moves with the nidus on probe or finger palpation. Measurements that include the halo are closer to histological measurements of tumor size but are more difficult to make than those of the nidus because the halo merges gradually into surrounding tissue.³⁸

Cancers are generally hard and so do not distort under probe pressure and, when invasive, they may be fixed to the adjacent breast tissue, which is dragged or rotated when the tumor is pushed by the finger. Invasion into the pectoralis muscle can be demonstrated in this way. The stiffness (coefficient of elasticity) of cancers is higher than normal tissues and benign lesions. Tissue stiffness increases with increasing probe pressure so the biggest difference of stiffness between benign and malignant tissue is seen with minimal precompression free-hand elasticity imaging.³⁹ Typically, cancers have a larger strain (elasticity) image than the gray-scale image (Figure 42-6), while benign lesions often have a smaller strain image than the gray-scale image.^40,41 Using the Ueno/Tsukuba scoring method, cancers are stiff with high scores.^21,22,42 The extent of the strain image of cancers seems to be a more accurate measurement of tumor size than the gray-scale image.⁴³ Even in situ and mucinous cancers cause increased stiffness compared with the surrounding normal tissues. Mucinous cancers and cancers with internal necrosis may exhibit a more heterogeneous strain pattern than solid tumors.

The neovascularization that is always present in carcinomas larger than a few millimeters in diameter may be demonstrable on color Doppler. Often the increased vascularity with numerous, tortuous vessels is very obvious and forms a useful additional sign of malignancy, especially where the typical anatomical features are not seen, for example, in inflammatory or non–mass-forming carcinomas. The degree of vascularization of a breast cancer correlates with its invasive potential.⁴⁴ The vascularity surrounding the lesion may be more visible than that within the lesion because of necrosis or loss of signal caused by the complex, chaotic nature of the tissues. The external and surface vessels frequently have a radial alignment, penetrating into the mas.⁶ Benign masses may also have Doppler signals, but the vessels usually track around the lesion rather than penetrating into it and are usually sparse and lack the chaotic pattern of malignant neovascularization.⁴⁵

There are often characteristic changes in the breast around a cancer. Thickening of Cooper's ligaments, caused by invasion or desmoplastic reaction, gives a typical stellate pattern when they retract. This is often much more clearly seen on three-dimensional scans that have been reformatted to show the coronal plane (C-scan).⁴³ Skin thickening results from invasion or edema and is most obvious in inflammatory cancers. Dilated ducts may result from obstruction or invasion, so that both the upstream and downstream portions may be dilated. Involved nodes are often obvious in the axilla and parasternal regions, and can be distinguished from reactive nodes by their spherical shape and effacement of the normal echogenic hilum.⁴⁵ Like the primary tumor, involved nodes may be vascular with peripheral vessels penetrating the capsule—a sign of neovascularization.

In addition to such typical lesions, a variety of other patterns may be encountered, some of which correspond to particular histological types. Where the cancer contains little fibrous tissue, shadowing is less apparent, and very cellular carcinomas may even show increased through-transmission with a confusingly benign appearance (eg, medullary, tubular, and mucinous carcinomas).⁴⁷ The halo is absent if the tumor does not invade locally. The nidus can be echogenic and blend with the fibroglandular layer or match that of the subcutaneous fat—a feature of colloid carcinomas.⁴⁸ In these instances, the tumor may be invisible on the static scan, although it can be picked out by its harder consistency than the surrounding tissues on probe palpation or on elastography. Although patchy necrosis may occur with larger lesions, extensive necrosis is rare except after chemotherapy; it produces a partly cystic lesion whose walls are irregular. Coexistence of a cyst or fibroadenoma and a carcinoma is well recognized; the ultrasound features of the malignant component are usually typical.⁴⁹

Some important types of carcinoma are particular diagnostic problems for ultrasound. Small cancers (<5 mm) sometimes have characteristic appearances that allow their diagnosis but may mimic the appearances of fibroadenomas; therefore, biopsy is required. Elastography is promising in their characterization. Widely infiltrating carcinomas that do not form a discrete mass are difficult to detect on ultrasound; even when extensive (eg, lobular carcinomas), they may not be apparent (ultrasonically occult lesions). Inflammatory cancers may mimic an infective process or developing abscess with increased vascularity, although the hardness of the affected portion of the breast and the associated skin thickening (peau d'orange) may be suspicious (Figure 42-7).⁵⁰ Smaller infiltrating carcinomas without a mass may be undetectable or may cause only subtle architectural disturbance or shadowing, which should be considered as suspicious if it persists on probe compression. They may be more rigid than the surrounding tissue and show as stiff regions on elastography.

Lobular, low-grade, and small high-grade ductal in situ carcinomas (DCIS) cannot be detected reliably by ultrasound; often no abnormality is seen, although occasionally microcalcifications in a portion of dilated duct may be detectable. The higher the grade of DCIS and the greater the extent of disease, the greater the likelihood of their being identified on ultrasound. Classically, DCIS is detected mammographically because of the microcalcifications associated with the comedo necrosis. Prior mammography increases the detection rates of small lesions by indicating the site of abnormality. In women under 35 who present with an area of vague thickening, for whom mammography may be contraindicated, medium- to high-grade DCIS may be detected on ultrasound, particularly if it causes clinical thickening felt by the patient. This change in consistency often shows up as increased stiffness on elasticity imaging. Absence of stiffness in clinical areas of thickening is nearly always indicative of normal tissue. The ducts are usually dilated with echo-poor contents and punctuate bright echoes that are caused by microcalcifications. Diagnosis is often aided by hypervascularity, which is probably caused by increased flow in the adjacent normal tissue induced by cytokines secreted by the tumor.⁵¹ Mammography can often help guide the experienced sonographer to recognize smaller areas of DCIS, which otherwise may not have been appreciated on routine ultrasound, and ultrasound can then be used to target core biopsy.

Metastases and lymphoma in the breast tend to be better defined than infiltrating carcinomas and lack an echogenic halo.⁵² Otherwise, they have similar appearances, with low-level heterogeneous echoes. They are also incompressible and often are vascular.

Intraduct tumors can be diagnosed if they are surrounded by fluid (Figure 42-8).⁵³ However, when there is no surrounding fluid, they appear as well-defined, solid masses and can be confused with small carcinomas or fibroadenomas. They are usually stiff on elasticity imaging. They can be diagnosed with needle core biopsy, but distinction between benign and malignant forms requires excision biopsy.

Benign clumps of calcification are readily demonstrated on ultrasound as intensely echogenic foci with distal shadowing. Microcalcifications are seen as echogenic puncta but without shadowing (because they are smaller than the beam width, some sound passes alongside them). Although these small calcifications are obvious against a dark background, such as within the nidus of a cancer, they are difficult to detect in tissue that is itself echogenic, such as the glandular part of the breast, and this means that microcalcifications are often missed. This reduces the sensitivity of ultrasound for detecting smaller in situ carcinomas—one of the main limitations of ultrasound for screening.

Silicone and other filling materials used for breast implants have a homogeneous structure and thus are echo-free.^54,55 Silicone has a sound speed of about half that of soft tissue, so the image is distorted, appearing to be twice as deep as it actually is—lateral geometry does not depend on the speed of sound and so is unaffected. The envelope forms an echogenic line around the liquid, and it may have infoldings that form echogenic planes within the echo-free prosthesis. Folds and creases sometimes present as palpable masses and can be identified with simultaneous palpation and ultrasound imaging of the clinical "mass." Because it can be difficult to distinguish internal folds from the separating layers of a torn envelope, ultrasound is not as reliable as magnetic resonance imaging (MRI) in detecting rupture, especially when it is confined by the fibrous capsule that usually forms around the implant ("intracapsular rupture").^56,57 The presence of silicone granulomas around the prosthesis and within axillary lymph nodes is indicative of rupture or significant leak of liquid silicone through the envelope. Ultrasound is also good at distinguishing palpable silicone granulomas from sinister masses (see the next section).

Just as elsewhere in the body, hematomas in the breast are seen as shaggy-walled cavities containing echogenic fluid in which debris and blood products may be seen to move with changing posture or changing probe pressure. Abscesses have the same general appearance; the distinction is usually obvious clinically, but when there is doubt, specific features, such as the presence of echogenic gas bubbles and the hyperemic periphery seen on Doppler, are useful in their differentiation.⁵⁸ The size of a cavity is often helpful in deciding whether aspiration is needed, and this will be performed under ultrasound guidance.⁵⁹

Trauma to the fat of the breast may cause various appearances. The most common is increased echogenicity within the fat (Figure 42-9).⁶⁰ This cannot always reliably be differentiated from an echogenic carcinoma (though these are rare), and diagnosis requires either needle biopsy or sequential follow-up ultrasound if there is no history of trauma. Significant trauma may result in fat necrosis and the formation of oil cysts, which may contain pure oil or a mixture of water and oil. A characteristic pattern results: the oil, which is echogenic if it is an emulsion, floats over the watery component, producing a fluid level that is upside-down by comparison with the usual layering due to dependent debris.⁶¹

Fibrous scars are typically seen as fine, linear, echo-poor structures with obvious shadowing (Figure 42-10).⁶² Any widening is suspicious of tumor recurrence, although innocent widening occurs where there is excessive fibrosis, as in a keloid or around sutures. Mature scars are avascular on color Doppler, and so any such signals in a scar more than about 3 months old means that a biopsy should be performed. Scars have a smaller strain footprint on elastography than the extent of B-mode abnormality and shadowing.

Radial scars, when visible on ultrasound, produce poorly defined, echo-poor regions, with shadowing and distortion of the surrounding breast consisting of indrawing of the Coopers ligaments to which they are fixed.⁶³ Thus, they mimic carcinomas and the distinction is impossible on ultrasound imaging. In addition, some are vascular on color Doppler.

Silicone granulomas have a pathognomonic appearance as masses with high-level echoes together with an echogenic flare deep to the lesion that is reminiscent of the ring-down artifact that occurs behind a gas bubble.⁶⁴ Apart from calcium and gas, no other lesion in the breast has echoes more intense than the gland itself. The most common site is at the periphery of silicone prosthesis, where they produce a striking "snow storm" appearance in comparison with the echo-free lake of the silicone (Figure 42-11). Because they may present as palpable nodules, this clear distinction from carcinomas, which have the usual echo-poor appearance, is clinically useful.

Ultrasound is preferable to mammography for assessing male breast problems, of which gynecomastia is by far the most common.^65,66 Gynecomastia, which can present as unilateral or bilateral development of breast tissue, is most commonly seen as an irregular area of low echogenicity in the retroareolar region, with a ductal pattern extending out to the periphery (Figure 42-12). They often exhibit marked vascularity on color Doppler, particularly if growing rapidly. The vessels tend to run in a centripetal pattern from the periphery of the developing breast towards the nipple. The tissue is soft and the darker areas of developing parenchyma are easily flattened and deformed with the probe, particularly in the retroareolar region—a reassuring appearance. Less commonly, the parenchyma can be mixed with fat and resemble the female pattern.

Male breast cancers can have the same appearances as typical breast cancers in the female, with an irregular echo-poor area with posterior shadowing, or they can be relatively benign-looking with well-defined margins like a fibroadenoma, which are even rarer than male breast cancers with this appearance.

Ultrasound is an adjunctive technique that is essential in the management of breast problems by improving the specificity of noninvasive diagnosis.^67,68 It is clearly the best method for guiding interventional procedures, most commonly for biopsy, but also to drain collections and to place marker wires before surgical excision of impalpable lesions. ^68,69 A suspicious finding on ultrasound, such as the demonstration of vascularity in what was thought to be a nodule in a scar or in the wall of a presumed cyst, should prompt biopsy or re-biopsy. Similarly, the demonstration of a region of stiffness on elastography is a signal that further investigation, or at least surveillance, is needed.⁹ When there are multiple focal lesions, ultrasound is often helpful in deciding which should be biopsied.

Ultrasound often also provides reassurance that surveillance can be continued and surgery avoided. Examples include typical fibroadenomas, echogenic masses associated with prostheses, and the demonstration that ductal structures traverse an apparent mass in focal benign breast change.

In general, ultrasound should not be used alone for screening because it misses the microcalcifications of DCIS that are so well demonstrated on mammography.⁷⁰ As an adjunct to mammography, ultrasound screening has been shown to pick up more small cancers in dense breasts, but with a significantly increased benign biopsy rate.⁷¹ In young women with a high risk of developing breast cancer, a strong family history or genetic risk, breast MR is the screening modality of choice.⁷² It is more sensitive than ultrasound and mammography, which also carries an increased risk from ionizing radiation. The high cost and high false-positive rate of breast MR screening means that new developments in breast ultrasound, such as elasticity imaging, need to be assessed for screening to improve detection and reduce the false-positive rate.

The rapid advances in ultrasound technology over the past few years have benefited breast ultrasound along with more general applications.^73,74 The marked improvements in the capabilities of digital scanners with sophisticated composite transducers have led to better imaging and Doppler, with lower noise floors and improved spatial resolution. The extended field of view available on many systems facilitates comparison of textures across the whole breast and is useful for follow-up studies (Figure 42-13).⁵⁵ Three-dimensional data sets achieve similar benefits and allow re-slicing that permits the identical view to be repeated for accurate follow-up comparisons. The introduction of microbubble contrast agents not only improves the display of small vessels, but also makes three-dimensional Doppler studies more informative.^75,76 They also provide the opportunity to perform functional studies by timing the wash-in and wash-out of a bolus injection; whether or not this has clinical value remains to be explored.⁷⁷

Entirely new methods have been applied to the breast, including whole-breast ultrasound systems,⁷⁸ ultrasound computed tomography with reconstruction both for attenuation and for sound speed,⁷⁹ and elastography,¹⁵ which has been incorporated into the main text of this chapter since it has begun to be used as a clinical tool.