Automated breast ultrasound (ABUS): A pictorial essay of common artifacts and benign and malignant pathology

The curved surface of the wide transducer used in ABUS creates a good contact with the contour of the breast, in combination with the use of skin lotion and the disposable membrane on the transducer. This ensures a uniform compression across the entire breast (Fig. 1).

Fig. 1.

The transducer is automatically adjusted in the range of 6–15 MHz, but the applied bandwidth depends on breast thickness, varying from 2.5 to 6 cm. ABUS uses wide beams of 25 mm with coverage of a large field of view (FOV), as opposed to a focused beam in handheld ultrasound (HHUS) (Fig. 1).

Incoherent compounding and crossbeam technology with steered beams in three separate angle groups improves contrast resolution by smoothing out the speckle pattern to optimize resolution at deep structures, such as prepectoral tissue, pleura, ribs, and even behind implants.

ABUS also makes use of high-end software to increase the resolution throughout the images, such as nipple shadow compensation, speckle reduction imaging, and breast border detection.

High-quality ABUS examination requires highly skilled technicians. Therefore, practical training and instruction of technicians is crucial to obtain good quality images.

This goal can only be achieved with a relaxed and well-informed patient who feels comfortable. Prior to scanning, it is important to instruct the patient not to speak or move during the scan, and to breathe smoothly to avoid motion artifacts.

The examination is performed in the supine position (Fig. 2). The arm is elevated, and usually a pillow is placed under the shoulder to optimize the patient’s comfort (Fig. 2).

Fig. 2.

The most important aspect before starting the ABUS is to spread out the breast tissue evenly, so that the nipple is pointing upwards. The nipple must be positioned just under the center because the upper quadrant contains more tissue than the inferior one.

A hypoallergenic water-soluble lotion is spread on the breast and axilla, and an additional lotion is applied around the nipple to obtain an adequate contact between the probe and the skin to avoid the air bubble artifact.

Three levels of compression may be used to optimize the contact between the probe and the skin. Automated and continuous volume acquisitions are obtained in the axial plane when the transducer moves from the inferior part of the breast to the superior section (Fig. 2). A single movement of the transducer allows the coverage of breast volume of 15.3 × 17 × 5.0 cm, with variable slice thickness of 0.5–10 mm at slice intervals of 0.5 to 2 mm.

For each breast, three volumes are obtained: the anteroposterior (AP) with the nipple in the center, the lateral view including the upper part and axillary tail of the breast, and the medial volume covering the inner part of the breast with the inframammary fold (Fig. 3). Additional views are often taken in patients with large breasts.

Fig. 3.

On the first scan, the technician evaluates the depth to ensure that the deep and peripheral breast tissues are included in the FOV. Depending on the breast size, there are five sizes to choose: extra small, small, medium, large, and extra-large, with a depth from 3.5 to 6 cm. The overall examination takes 15–20 minutes⁽⁴⁾.

When the examination is completed, the volume data are processed automatically in multi-planar reconstruction with coronal and sagittal views, after which they are transferred to the workstation^(5–7).

Wide axial views (Fig. 4A) provide detailed information on the nipple and regions behind the areola, the subcutaneous adipose tissue, and the hyperechogenic fibroglandular tissue. Due to the exquisite resolution for deeply located structures, and the large FOV, the evaluation of margins, shape, orientation, and echogenicity of lesions is optimal.^(2,6)

Fig. 4.

The coronal view (Fig. 4B), with thin-sectional frontal views of the breast at different depths of the breast, is the most valuable view. It allows visualizing the entire breast anatomy from the skin to the pectoral muscles and underlying ribs^(3,4).

The coronal views are also very useful for evaluating irregular stellate borders, distortion, and posterior shadowing even in small malignant lesions.

The most reliable sign for the characterization of a suspicious lesion on the coronal view is the retraction phenomenon sign, a stellate pattern in which white spicules are caused by the desmoplastic reaction of the tissue surrounding a malignant lesion (Fig. 5)^(2,8,9).

Fig. 5.

Another major advantage of ABUS is that measurements can easily be done in three directions, and the distance to the skin, the nipple, and other lesions is easily made in case of multifocality or multicentricity.

Due to the similar supine position of the breast during ABUS and the positioning during breast surgery, the coronal views are particularly valuable for surgeons to localize a lesion in the preoperative setting of breast conserving surgery.

ABUS is also useful for monitoring the effects of neoadjuvant chemotherapy and for the evaluation of fibrosis^(3,4). The reproducibility of the ABUS technique makes it a valuable tool in the follow-up of benign lesions as well.

Compression on the breast is sometimes a challenge, especially on the sternum and axilla, in women with small breasts.

Consequently, ABUS may be a rather painful experience in some women. Although there is less compression in ABUS compared to mammography, it is still more pronounced than in handheld US (HHUS), with three grades of compression in ABUS.

Although some axillary lymph nodes at level 1 can be visualized, sometimes the axilla is not completely evaluated, depending on the constitution of the patient. A fatty axilla makes it easier to see the lymph nodes.

In contrast to ABUS, HHUS offers the availability of color Doppler and elastography. Another drawback is that ABUS is more time-consuming than full-field digital screening mammography (FFDSM) both in terms of performing the examination and interpreting the findings^(10,11).

Although breast implants are not a contraindication for ABUS, and potential leaks and tears can easily be identified, the firm structure can be a challenge for positioning.

Recognition and classification of artifact patterns minimizes false-positive interpretations⁽²⁾ The origin of artifacts can be variable.

Technique- and software-related artifacts

Inadequate transducer positioning: drop-out artifact

A well-performed technique, with a uniform breast compression and smoothing of the superficial tissue before the transducer is placed, can solve the problem of skinfold and drop-out artifacts at the border of the breast (Fig. 6).

Fig. 6.

Firm breasts and large breast implants may be difficult to compress. Consequently, it can be a challenge to position the transducer correctly and achieve an optimal contact area.

Air contact artifacts

Sound waves can be entrapped between the transducer and the patient’s skin, creating a reverberation pattern from the skin, with a total loss of signal underneath an air bubble. Air bubbles cause typical “black holes” on the image. They can be prevented by using enough water-based lotion and applying an appropriate compression (Fig. 7)⁽³⁾.

Fig. 7.

Nipple shadow

Disturbing acoustic shadowing behind the nipple, due to a column of altered echogenicity and high acoustic impedance of the dense tissue behind the nipple, has a hypoechogenic mass-like appearance. This artifact is often software-related. It is mostly seen on the lateral and medial views due to the tangential ultrasound reflection at curved structures behind the nipple. Therefore, the anteroposterior view is the best view to evaluate the retroareolar region (Fig. 8)^(2,3).

Fig. 8.

Patient- or lesion-related artifacts

Wandering shadows

Adjacent to the edges of the curved surface of Cooper’s ligaments, scattering of ultrasonic waves may lead to a loss of signal and acoustic shadowing. The shadowing may “wander”, especially while scrolling through the transverse planes, which is why it is referred to as “wandering shadows” (Fig. 9)⁽²⁾. Unfortunately, the current ABUS technology does not allow the correction of the transducer angle. The same effect of shadowing can occur in HHUS, but can be corrected by adjusting the angle of the transducer.

Fig. 9.

Sinusoidal wave pattern

Heavy breathing or talking may result in sinusoidal waves and image distortion. The waves have the same frequency as breathing. This wave pattern can distort the evaluation of the deeper regions close to the chest wall, especially on the sagittal and coronal reconstructions. Even tachycardia can cause artifacts during scanning (Fig. 10)⁽²⁾.

Fig. 10.

Skip artifact

A horizontal or sagittal line artifact on coronal and sagittal views can be created when the transducer moves over the bumpy surface of a firm mass e.g., a fibroadenoma, a region of nodular dense breast tissue or even an implant (Fig. 11)⁽²⁾.

Fig. 11.

White wall artifact

The posterior enhancement of a cyst, seen on HHUS as well, can create a white wall artifact on the coronal plane at the level of the posterior enhancement of the cyst. This can also influence the image interpretation and is not solely a sign of a benign lesion. Using additional planes is often very useful for the correct analysis of artifacts and to prevent misinterpretation (Fig. 12).

Fig. 12.

Shadowing due to surgical scars and clips

In patients with surgical scars and clips, the sound attenuation artifact extends posteriorly to the scar at the skin or behind this clip respectively. This may mimic tumor recurrence (Fig. 13).

Fig. 13.

Simple breast cysts are benign round or oval anechoic lesions with retroacoustic enhancement and thin wall^(12,13). The lesions are usually well-defined, with the longest diameter parallel to the skin. In women with multiple cysts, ABUS provides an overview of fibrocystic breast tissue and helps to differentiate cysts from a possible underlying malignancy or a complex or complicated cyst (Fig. 14).

Fig. 14.

Fibroadenomas are encountered in 15–25% of women undergoing breast cancer screening and are more common in younger women⁽¹⁴⁾. They usually manifest as well-circumscribed hypoechoic masses surrounded by a thin hyperechoic capsule, without architectural distortion. Old lesions may be sclerotic, and may contain calcifications and septa. In cases of multifocal or multicentric bilateral fibroadenomas, ABUS is not only useful for complete mapping of all lesions, but also for rendering reproducible images on follow-up examinations (Fig. 15)⁽¹⁵⁾.

Fig. 15.

Fat necrosis is a benign breast condition that can arise from trauma, breast surgery or radiation treatment in fatty breast tissue.

The ultrasound features of fat necrosis vary depending on the stage of the lesion, ranging from multi-cystic appearance surrounded by hyperechogenic edema in the acute phase, hyperechogenic tissue due to edema in the subacute phase, or a peripheral oil cyst in the chronic stage.

Fat necrosis can also present as an angular, spiculated lesion due to fibrosis, and may mimic breast carcinoma. In conjunction with clinical history, ABUS can be helpful for distinguishing it from a true malignancy by meticulous analysis of the follow-up images along with different stages of fat necrosis (Fig. 16)⁽¹⁵⁾.

Fig. 16.

Intramammary lymph nodes (IMLN) are incidental and common benign findingson US and mammography. They are surrounded by breast tissue in all directions. They are typically circumscribed, smaller than 10 mm, with an oval or reniform shape, and hilar fat. IMLNs are predominantly seen in the upper outer quadrant in the breast and often adjacent to a vein. However, they can be located anywhere in the breast (Fig. 17)⁽¹⁶⁾.

Fig. 17.

Intraductal papilloma is a benign neoplasm of the intraductal epithelium with a fibrovascular core. On physical examination, it may present with nipple discharge or a palpable, painless, and mobile mass. It may also be associated with gynecomastia.

The differentiation between intraductal papilloma and papillary carcinoma is not possible with imaging alone.

Therefore, a histopathological examination is necessary to make the final diagnosis (Fig. 18).

Fig. 18.

In case of papillomatosis or multiple peripheral localized papillomas, ABUS can be helpful for the precise evaluation of the location and extent of different lesions⁽¹⁷⁾.

Radial scars or CSLs result from an idiopathic process with sclerosing ductal hyperplasia, and are unrelated to surgical scarring. They are considered as B3 lesions, i.e. benign lesions of uncertain malignant potential. On US images, they typically present as a spiculated lesion or hypoechogenic distortion with a central fibrous core, radiating ducts or small cysts. Sometimes, only a spiculated lesion is seen on mammography, with no specific sign on HHUS. ABUS may be useful for detecting architectural distortions with a retraction pattern, especially on coronal images (Fig. 19)^(10,18).

Fig. 19.

Lipoma is one of the most common benign breast lesions composed of mature fat cells. Clinically, lipoma presents as a painless asymptomatic lesion or a soft, non-tender palpable mass. On ABUS, it is oval-shaped and longitudinally oriented to the chest wall. The echogenicity of lipomas varies from a mild hyperechoic to iso- or hypoechoic lesion surrounded by a thin echogenic capsule (Fig. 20). Lipomas can be large, and the wide FOV of ABUS allows the coverage of the entire lesion⁽¹⁹⁾.

Fig. 20.

IDC is the most frequent type of breast carcinoma, accounting for 80–90% of all cases. The ultrasound appearance varies depending on the proliferation rate and hormone receptor status. More indolent subtypes or hormone-receptor positive types of IDC are often hypoechoic, irregular, or indistinct, have spiculated margins, and show posterior acoustic shadowing. The retraction phenomenon is present in the majority of IDCs on coronal views (Fig. 21, Fig. 22)^(22–24).

Fig. 21.

Fig. 22.

On the other hand, more aggressive invasive cancers, such as triple-negative IDCs and other high-grade IDCs have mostly circumscribed or micro-lobulated borders, are often round, and have posterior acoustic enhancement⁽⁹⁾. Therefore, these lesions seem to mimic benign tumors, which can lead to false negative results on ABUS⁽²⁾.

About 11% of invasive cancers are of the lobular type⁽²⁵⁾. In the majority of ILCs, the US features consist of a hypoechoic irregular mass with indistinct angulated margins, an echogenic halo, and posterior acoustic shadowing⁽²⁶⁾. ILCs may also manifest without a distinct mass or with hyper- and isoechoic patterns, either with or without focal acoustic shadowing, commonly with subtle distortion. This is more frequently observed in ILCs than in IDCs⁽²⁶⁾.

Consequently, the extent of ILCs is often underestimated with US, but nevertheless US has a higher sensitivity for depicting ILCs than mammography⁽³⁾. In addition, ABUS may be very useful for ILC detection by demonstrating a retraction pattern on coronal images even in the absence of a distinct mass (Fig. 23).

Fig. 23.

ITC accounts for 2% of all breast cancers, and it is a low-grade cancer⁽²⁵⁾. Its US appearance typically consists of a small hypoechoic irregular mass with an echogenic halo. The margins are often spiculated, with the presence of posterior acoustic shadowing. It commonly causes a severe retraction pattern on coronal ABUS images.

This histological type accounts for 2% of all breast cancers⁽²⁵⁾. It is a mucin-producing ductal cancer presenting on US as a complex solid/cystic mass, commonly oval or lobulated, with posterior acoustic enhancement. A small mucinous tumor can be falsely interpreted as a tiny cyst. A retraction pattern on coronal ABUS images is often demonstrated and even subtle architectural distortions on multiplanar reformatted ABUS images can often be used as clues to suggest the diagnosis.

This cancer type presents as tiny intraductal microcalcifications on mammography or as non-mass enhancement or ductal enhancing pattern on MRI. Most cases of DCIS are difficult to detect by US. Subtle US changes include duct dilatation with discrete intraductal microcalcifications⁽²⁷⁾. They are also difficult to depict on ABUS, unless extensive.