Ultrasound

Ultrasonography employs mechanical energy (sound) rather than electromagnetic radiation to produce a pictorial representation of the internal structure of the breast. The image is produced by transmission of sound pulses into the breast and measurement of the returning echoes at later times, depending upon the depth of interfaces between different tissue types. The transducer functions as both transmitter and receiver. An attractive feature of sono-graphic imaging is that there are no known carcinogenic effects of ultrasound at the power levels employed for diagnostic purposes.

In addition to cyst-solid differentiation, other indications as listed in the ACR's Standard for the Performance of Breast Ultrasound Examination (ACR, 2000b) are: (1) identification and characterization of palpable and nonpalpable abnormalities and further evaluation of clinical and mammographic findings, (2) guidance of interventional procedures, and (3) evaluation of problems associated with breast implants.

8.1.1 Ultrasound for Cyst-Solid Differentiation

One of the major applications of breast ultrasonography is in distinguishing cysts from solid masses. Ultrasound is the least costly, most rapidly performed, and most readily available additional imaging method for evaluating mammographic masses that may represent cysts. Ultrasound is more accurate than either mammography or physical examination for identifying cysts. If a mass demonstrates the four sonographic criteria of round or oval shape, circumscribed margins, posterior acoustic enhancement, and anechogenicity, a benign cyst can be diagnosed with nearly 100 percent accuracy (Bassett et al., 1987b; Feig, 1992; Hilton et al., 1986; Kopans et al., 1984; Rubin et al., 1985; Sickles et al., 1984). It is unnecessary to aspirate simple cysts unless they cause symptoms, such as pain (Hilton et al., 1986; Mendelson, 1998). In addition, with high resolution ultrasound, cysts that contain homogeneous low-level internal echoes can be commonly encountered. These complicated cysts, which mimic solid lesions, require aspiration if they are symptomatic or if the diagnosis is uncertain, but they may otherwise be placed in a follow-up category (Kolb et al., 1998; Venta et al., 1999). Furthermore, ultrasound can be used to characterize complex cysts containing solid components that require biopsy.

In the past and currently in some situations, needle aspiration guided by palpation was employed as a rapid and possibly less expensive method to achieve cyst-solid differentiation, while simultaneously providing therapy (Bassett et al., 1987b; Kopans, 1986). As utilization of ultrasound in breast imaging practices has become more frequent, it has been observed that when the palpable mass is a cyst, other simple cysts may also be present, none of them requiring intervention (Kolb et al., 1998; Mendelson, 1998). If ultrasound imaging is planned, it should be performed prior to any intervention. The introduction of blood into a cyst during an incomplete or unsuccessful aspiration may change the ultrasound appearance from that of a simple cyst to a complicated or complex cyst, or to an indeterminate solid-appearing lesion.

8.1.2 Ultrasound for Benign-Malignant Differentiation

Although ultrasound has been established as reliable in cyst-solid differentiation, the characterization of solid masses as benign or malignant has been more difficult. Previously, technical inadequacies, interobserver variability in performance, and interpretation, and inexperience resulted in a mistrust of diagnoses and subsequent management of solid breast masses based on their ultrasound appearances (Jackson, 1990; 1995). Also noted were difficulties encountered in locating hypoechoic masses, particularly if small or in fatty breasts, where the tissue surrounding the lesion might be isoechoic with the mass (Bassett and Kimme-Smith, 1991; Sickles et al., 1983). The lack of contrast between lesion and surrounding breast tissue is similar to that in mammography of a dense mass hidden in dense breast tissue (i.e., the sensitivity of the imaging technique diminishes). For these reasons, it has been believed that solid masses were not well enough characterized sonographically to forego tissue sampling on the basis of the ultrasound appearance alone.

Although there is overlap of features of benign and malignant masses, and operator dependence remains a problem (Mendelson et al., 2001; Merritt, 1999; Stavros et al., 1995), progress has been made in the characterization of solid breast masses. A combination of sonographic features such as shape, orientation, margin, echo pattern, and posterior acoustic characteristics has greater predictive value for malignancy than any single, stand-alone sono-graphic feature (ATL, 1997; Cole-Beuglet et al., 1980; Stavros et al., 1995). In 1995, Stavros and colleagues published the constellation of features that enabled him to characterize a solid lesion as probably benign, with less than a two percent likelihood of malignancy [e.g., uniform hyperechogenicity relative to fat; oval or gently lobular shape; and thin, circumscribed margin (Sickles, 1991; Stavros et al., 1995)]. However, the Stavros et al. findings have not been validated in a multicenter study or other peer-reviewed single-institution publication. Therefore, the use of ultrasound to characterize solid breast lesion as probably benign remains controversial and it is generally accepted that ultrasound cannot reliably avert tissue sampling and histologic diagnosis unless characteristically benign sonographic features (e.g., inflammatory lymph nodes) are demonstrated.

As with other imaging techniques, irregularity of margin and shape are the dominant features that predict malignancy at ultrasound with PPVs of 80 to 93 percent (ATL, 1997; Mendelson, 1999; Mendelson et al., 2001; Stavros et al., 1995). Other features, including orientation and acoustic attenuation characteristics, are less specific.

Considerable research continues in the management of masses seen with ultrasound. For consistency in interpretation, which has long required a solution, a lexicon of descriptors similar to that used in the Breast Imaging Reporting and Data System (BI-RADSĀ®) for mammography is being developed by ACR (1998). The ultrasound findings, together with the mammographic interpretation, clinical examination, and patient's history should result in more specific assessments and management plans. Indeed, the use of ultrasound adjunctively with mammography has significantly reduced the numbers of benign biopsies (ATL, 1997; Zonderland et al, 1999).

Although Doppler has been studied for its contribution to the characterization of masses, the initial enthusiastic endorsement of Schoenberger and colleagues, who found 100 percent sensitivity and specificity in distinguishing benign from malignant masses, has never been corroborated (Schoenberger et al., 1988). Subsequently, several researchers reported their disappointing results (Adler et al., 1990; Dock et al., 1991; Jackson et al., 1993). Unless refinements of Doppler technique, such as the use of intravenous contrast agents, are shown to be of value, blood-flow characteristics of masses may be regarded as an additional and optional feature to apply in the categorization of masses as benign or malignant, largely accomplished through the application of morphologic criteria.

8.1.3 Ultrasound for Breast Cancer Screening

Breast ultrasound is regarded as a targeted examination at the current time, although the use of ultrasound for breast cancer screening has been proposed for at least 20 y. In the early 1980s, automated water-path scanners were available. These scanners had transducer frequencies as high as 7.5 and 10 MHz (Jackson et al., 1986), but the equipment was costly, cumbersome and time consuming in clinical application. In preliminary clinical studies, automated water-path scanners failed to detect many small non-palpable cancers that were detected by mammography (Kopans, 1984; Sickles et al., 1983).

A further negative note was sounded by the European Group for Breast Cancer Screening as a result of a literature review and consensus conference held in 1996. This group acknowledged the benefits of ultrasound as an adjunct to mammography, but cited the published low sensitivities and specificities for ultrasound in the screening setting with attendant risk of high FP and FN rates (Teh and Wilson, 1998). In North America, the results of several studies have supported the institution of a screening trial. Gordon and Goldenberg (1995) reported on 1,575 solid masses, 44 (0.3 percent) of which were cancers seen only with survey ultrasound in the 12,706 women being evaluated for palpable or mammographic masses. Kolb et al. (1998) studied 3,626 women with normal mam-mographic and clinical examinations. Women with fatty breasts were excluded. Ultrasound depicted 215 solid masses not seen mammographically, of which 11 (five percent) were malignant. Nine hundred seventy-four (974) women (27 percent) had cysts, and 132 (3.6 percent) had complicated cysts with no malignancy found in follow-up of that group. Six thousand, one hundred thirteen (6,113) asymptomatic women were screened with ultrasound by Buchberger et al. (1999). Twenty-three (23) cancers were identified in 21 women, and 353 masses found incidentally were biopsied or aspirated. The average size of the cancers seen with ultrasound was 0.9 cm, no larger than that found by screening mammography. Studies by Crystal et al. (2003), Kaplan (2001), and Leconte et al. (2003) have also reported that ultrasound can often find small breast cancers that were not detectable in dense breasts by screening mammography.

Kolb et al. (2002) has recently extended his 1998 study to include evaluation of 11,130 asymptomatic women in 27,825 screening sessions. This study found that mammography alone had a sensitivity of 98 percent in women with fatty breasts, with sensitivity decreasing as breast density increased, to a sensitivity of 48 percent in women classified in the highest ACR BI-RADSĀ® density category. Excluding fatty breasts, Kolb et al. (2002) found that breast ultrasound had a sensitivity of 75 percent while physical examination alone had a sensitivity of 32 percent. Adding screening breast ultrasound for women with nonfatty breasts to screening mammography for all women in the study yielded an overall 97 percent sensitivity, compared with a sensitivity of 74 percent when physical examination was added to screening mammography. The difference between these two combined screening strategies was highly significant (p < 0.001).

These studies suggest that certain groups ultimately may benefit from ultrasonographic screening. These include those with mammographically dense breasts; women with known carcinomas for multifocal or multicentric disease (Berg and Gilbreath, 2000), and other high-risk women, by virtue of family history or biopsy histologies of lobular neoplasia or atypical ductal hyperplasia. However, because efficacy has not yet been demonstrated, the use of ultrasound for breast cancer screening is currently considered experimental.

Possible impediments to screening breast ultrasound are cost, additional physician time, and likelihood of increased numbers of benign biopsies. Importantly, although screening breast ultrasound will undoubtedly uncover occult carcinomas presenting as masses, it is unlikely to be effective in screening breast tissue for microcalcifications, the major form of presentation for DCIS, unless the microcalcifications are embedded in a mass or contained within a dilated duct. Axial and lateral resolutions of the transducers used for breast ultrasound are approximately the size of larger microcalcifications (0.2 to 0.5 mm), but specular reflectors representing acoustic noise or transverse views of connective tissue elements may simulate microcalcifications, making them difficult to see and characterize.

A multicenter trial of breast ultrasound is in development, based on protocols written with support from the Office on Women's Health, DHHS (ACR, 2000b). Such a trial will be performed with high resolution, linear array transducers with compound scanning capability and a foot print of 50 mm or greater to scan larger areas of breast tissue expeditiously. This type of trial is needed to determine if the effectiveness of screening ultrasound found by Kolb et al. (2002) can be reproduced by a cross-section of experienced breast ultrasound users.

8.1.4 Ultrasound to Evaluate Complications of Breast Implants

Ultrasound has been used for evaluation of complications of silicone implants, such as rupture or leakage. Extracapsular rupture, extrusion of silicone into the breast parenchyma and surrounding tissues, has a distinctive sonographic appearance: "echogenic noise" or a "snowstorm pattern," obliterating sono-graphic information located posterior to the silicone (Caskey et al., 1999; DeBruhl et al., 1993; Harris et al., 1993; Mendelson, 1992). Other findings in extracapsular rupture, usually of long duration, are silicone granulomas, angular, hypoechoic masses that can be palpable. One of the major roles for ultrasound in patients who have had breast augmentation is distinguishing between a paren-chymal lesion and an implant-related finding, such as a wrinkle or fold. Extravasated silicone may also be seen on mammograms, depending on its location. Intracapsular rupture, signifying the degradation of the silicone polymeric shell of the implant, has some sonographic signs including the 'stepladder' effect of the shreds of implant envelope suspended in the silicone gel. Intracapsular rupture is also suggested by low level internal echoes with or without cystic areas within the implant.

Ultrasound is more accurate than mammography in identifying implant rupture, but MRI and CT are more sensitive and specific (Berg et al, 1995; Gorczyca et al, 1994; Ikeda et al, 1999).

If a lesion can be visualized sonographically, no matter what its size, any of the percutaneous procedures can be performed with ultrasound guidance. These procedures include presurgical needle-hookwire localization, spring-activated or vacuum-assisted core needle biopsy, fine-needle aspiration, abscess drainage, and retrograde ductography. Sonographically-guided procedures can be performed rapidly in real time, providing an opportunity for instant adjustment of the path of the needle towards its target. Core biopsies performed with ultrasound guidance are less costly than ster-eotactically-guided procedures (Liberman et al., 1998; Parker and Klaus, 1997; Parker et al., 1993). Ultrasound guidance, for most radiologists, is the preferred method (Jackson, 1995; Parker and Klaus, 1997), particularly for masses. For microcalcifications not present within a mass, stereotactic guidance is used most often.

The most common method of performance of these procedures is that of freehand placement of the biopsy or localization needle under direct sonographic visualization. For optimal viewing of the entire needle shaft, the needle entry should be nearly horizontal, perpendicular to the acoustic beam, and entirely within the plane of the beam (ACR, 2000b).

0 0

Post a comment