Applied Radiology

Dr. Shaw deParedes is Professor of Radiology and Director of Breast Imaging at the Medical College of Virginia, in Richmond; she is also on the editorial board of this journal. Dr. Dunn is an Assistant Professor of Radiology and Ms. Cousins is a mammographer, both at MCV.

________________________________________________________________________

Augmentation mammoplasty for cosmesis has been performed on greater than 2 million American women. The procedures have included augmentation by direct injection of various substances and augmentation by placement of a prosthesis or implant. Reconstruction of the breast after mastectomy is performed by creation of a breast mound with a myocutaneous flap, or by insertion of an implant.

Early procedures for augmentation included the direct injection of various substances, including talc, paraffin and silicone. This method is no longer performed in the United States, but is occasionally still used in Asia. The first silicone gel implant was placed in 1962.1 In 1992, the FDA placed a moratorium on silicone breast implants after the question of a possible association with connective tissue disease was raised. Currently in the United States, only saline implants are permitted for augmentation purposes.

Direct injection for augmentation

In the early 1900s, liquid paraffin was used commonly for augmentation mammoplasty by direct injection into the breast. Other substances utilized for injection-type breast augmentation have included wax, oils, fat, and talc. These injections often caused hard masses with tissue necrosis. In Japan in the 1950s, augmentation was performed by direct injection of silicone into the breast; this practice was subsequently adopted in the United States. Problems similar to those encountered with paraffin injection occurred with silicone, including the formation of dense, hard, tender breast masses caused by the response of the tissue to the foreign material. Silicone injections were banned in the United States in 1965; however, some women continued to have the procedure performed elsewhere.2

Mammography in the patient who has undergone augmentation bv direct injection is challenged by the extreme density of the tissue, the hardness and retraction of tissue, which compromise positioning, and the extensive masses and calcifications that develop, limiting accurate diagnosis. The use of higher kVp ranges (30 to 32) and alternate target/filter combinations (MO/Rh or W/Rh) to shorten the exposures and better penetrate the tissue are often of some aid in optimizing the image quality. Hyperdense parenchyma associated with diffuse spherical or eggshell calcifications of the fat necrosis type are typical mammographic findings (figure 1).3 Sonography is of limited value, often showing diffuse shadowing from the fibrotic reaction.

Implant types

Though only saline implants are currently permitted for augmentation, three basic types of implants may be seen on mammography: saline, silicone gel, and double lumen, which contains silicone in one chamber and saline in the other. The envelopes of the implants are made of either a smooth silicone elastomer, which appears on mammography with very well-defined margins, or a textured elastomer, which, on imaging, has fuzzy, irregular margins. Some types of saline implants, particularly those used as tissue expanders for breast reconstruction, are equipped with a port, through which saline is injected for gradual filling.

Silicone implants are much more dense on mammography than are saline prostheses. Silicone implants may be associated with gel bleed4 (migration of silicone through the implant envelope) or rupture with extravasation of silicone into the tissues. Saline implants have a greater tendency to deflate if ruptured (figure 2), but are not associated with the tissue reaction that occurs with silicone.

Implants are placed in either a retroglandular (prepectoral) location or in the subpectoral space. After insertion, the normal implant incites a fibroblastic response of the contiguous tissue; this will ultimately develop into a thin capsule which encloses the implant. One of the complications associated with implants, particularly those in a retroglandular location, is capsular thickening and contraction. It is believed that periprosthetic infection is associated with capsular contraction.5 Implantation in a retropectoral or submuscular location and the use of textured implants are both thought to be associated with fewer cases of capsular contraction.6

Mammographic imaging of the augmented breast

The presence of implants makes imaging the breasts more challenging and, therefore, correctly positioning the breast for mammography is vital. The degree to which implants do obscure breast tissue depends on their position, density, and level of encapsulation. Silverstein et al7 found a 49% decrease in measurable breast tissue on mammography after augmentation mammoplasty; Guenther et al8 report that only 7 of 20 cancers in women with implants were detected on mammography.

Standard mammographic imaging of the patient with implants includes

the following: standard mediolateral, oblique, and craniocaudal views, and a modified positioning, described by Eklund et al,9 known as the implant displacement (ID) views. Standard views are utilized to visualize the posterior tissues and low axillary region of the implant, whereas implant displacement views allow for better compression and visualization of the anterior parenchyma.

To position the patient for the ID views, the technologist must palpate the anterior edge of the implant, pulling the tissue that is anterior to it forward as the compression device is brought down. As compression is applied, the implant should be displaced posterosuperiorly (figure 3). These implant displacement views may not be satisfactory, however, for patients with encapsulated implants: In this situation, a standard ML or 90° lateral view may be helpful to visualize more of the tissue obscured by the implants.

Complications of implantation

In addition to the complications already mentioned (gel-bleed and capsular contracture), other complications related to implants include frank rupture, infection, hematoma formation, and capsular calcification. Implants also have been found to have a questionable association with connective tissue disease.

Rupture of silicone can occur in two basic sites, each affecting the imaging appearances of the breast differently. Rupture of the implant with extravasation of its contents into the space surrounding it (intracapsular location) is differentiated from rupture through the fibrous capsule and into the tissue (extracapsular location). Silicone can track into the axilla and be visualized in lymph nodes. It also can extend into the duct system and present as a silicone nipple discharge.

Clinical findings, including change in contour or location of the implant, flattening of the implant, and breast pain are associated with both intracapsular and extracapsular rupture. Additionally, extracapsular rupture may be associated with palpable masses (silicone granulomas) and silicone nipple discharge.

Mammographic and sonographic findings of implant rupture

A normal implant is seen on mammography as having a smooth contour and an oval shape. The density of each type of implant will vary: silicone is of high density; saline is less dense. A double-lumen implant with a saline outer shell has a double-density or bag-within-a-bag appearance.When the inner lumen of a double-lumen implant ruptures, there is admixture of the components. Mammographically, loss of the interface between silicone and saline occurs, and an admixture of the densities will result. Rupture of only the outer saline lumen will produce a mammographic appearance similar to that of a single-lumen silicone implant.

The mammographic finding of a change in implant contour (bulging or peaking) suggests either herniation or intracapsular rupture. Presence of such findings should prompt further studies to evaluate the integrity of the implant. With extracapsular rupture, mammography may demonstrate the appearance of free silicone droplets in the breast or axilla. These are hyperdense and may be calcified.

Sonography is helpful for the evaluation of a possible implant rupture, particularly in single-lumen silicone implants. The normal implant has an echogenic line or interface surrounding an anechoic oval space;10 textured implants appear as multilayered echogenic lines anteriorly at the surface.11 Wrinkles or folds in the surface can be seen on sonography as extending from the periphery of the implant inward and changing in shape between the transverse and sagittal planes. Because of the visualization of the internal lumen within the anechoic space, sonographic imaging of double-lumen implants can be confused with the pattern of intracapsular rupture.

Descriptions of the sonographic findings of intracapsular-space implant rupture have included diffuse internal echoes, paralleling horizontal lines,l2 and the stepladder (figure 4) sign.13 Berg et all4 found the combination of the stepladder sign and moderate-to-marked low level echoes (figure 5) to be present in 26 of 40 ruptured implants, with a sensitivity of 65% and a specificity of 57%; Debruhl et all3 found these signs to be associated with a 70% sensitivity and a 92% specificity for intracapsular rupture.

With extracapsular rupture, sonography demonstrates a classic "snowstorm appearance" (figure 6), caused by free silicone mixing with the breast tissue. This area of hyperechogenicity has a defined anterior margin, but a diffuse, indistinct posterior margin. Additionally, hypoechoic "masses" (figure 7), representing larger collections of free silicone, may be visualized in the breast tissue and are usually associated with the snowstorm pattern (table 1).15

MRI

Magnetic resonance imaging (MRI) has proven to be the single most effective tool for the diagnosis of silicone implant rupture. When two orthogonal sequences are used to evaluate the implant,l2,16 a sensitivity of 94% and a specificity of 97% has been reported. MRI allows multi-planar imaging of the entire implant and the surrounding tissue, addressing the limitations of both mammography and ultrasound. In addition, it is not operator dependent. MR can distinguish between silicone and normal breast tissue because of its specifically-designed pulse sequences. However, cost is a consideration, as MR is significantly more expensive than either mammography or ultrasound. For optimal resolution, MR is performed with the use of a dedicated breast coil, with the patient lying in a prone position and the breast resting dependently within the coil. Axial and sagittal or oblique sagittal scan planes allow straight forward comparison with mammographic images. Many different pulse sequences can be used that distinguish between silicone, water, and fat.

Technical factors

Most medical-grade silicones are composed of methyl polysiloxane with varying degrees of polymerization. The protons of the methyl groups are the source of the MR signal. Though the implant envelope is composed of silicone, the many additional cross-linkages between the methyl groups renders it an elastic solid, with only minimal MR signal.l7

There are a variety of suitable pulse sequences for imaging silicone implants, the use of which is determined by the relative Larmor precessional frequencies and the Tl and T2 properties of the fat, muscle, and silicone content in the breast. At 1.5T,l2 the relative resonance frequency of silicone is approximately 100 Hz less than fat and 300 Hz less than water. The T1 value of fat is less than that of muscle, which is less than that of silicone. Therefore, on a Tl-weighted image, fat has the highest signal intensity and silicone the lowest. Because the Tl of fat is so short, its high signal intensity suppresses the intensities of the other tissues; therefore, a fat-suppression sequence will facilitate better visualization of the other tissues. Using clinical suppression techniques, extracapsular silicone is seen as a hypointense region within a hyperintense region. The fat saturation pulse will also suppress silicone to some degree because of the similarities in the resonance frequencies of fat and silicone.

In terms of T2 values, the T2 of silicone is greater than fat, which is greater than muscle. Silicone, therefore, has a high signal intensity on a T2-weighted image, with fat intermediate in signal intensity and muscle the lowest. With the addition of water suppression on a T2-weighted image, cyst fluid and muscle will have a very low signal intensity, with silicone's remaining high. This facilitates the easier detection of extracapsular leaks.

Other sequences which take advantage of the different relative relaxation times of fat, silicone and water can be used. An inversion recovery sequence with a short TI (STIR) will maintain a high signal from silicone while suppressing the signal from fat. A silicone-selective sequence can be accomplished with the use of a chemical water-suppression pulse in conjunction with STIR.18

Intact implants

A normal silicone implant has a smooth, well-defined margin, and should have a homogeneous signal intensity. A potential space exists between the inner surface of the fibrous capsule and the prosthesis shell. MR cannot normally differentiate between the hypointense thin shell of an intact prosthesis and the fibrous capsule. As the fibrous capsule thickens, hardens, and contracts, it can squeeze and deform the intact prosthesis, causing infoldings or creases in the envelope. This can be seen in both single-lumen and double-lumen implants. An outer saline lumen often deflates over time, and the collapse of the outer shell around the silicone lumen results in extensive infolding. Rupture of a double-lumen implant with a saline-filled outer lumen usually only involves the outer saline-filled component; as this component is of a relatively small volume (100 ml to 200 ml), there usually is no noticeable change in the contour of the implant.

Normal radial folds are one of the most commonly encountered findings in intact implants, and must be distinguished from intracapsular rupture. Radial folds are hypointense lines that emanate from the periphery of an implant at the fibrous capsule-shell junction, and correspond to the infoldings of an intact shell. These folds are often thicker than the lines associated with the "linguine" sign because they consist of two adjacent shell layers. They may be short and straight (simple), or long and more curved (complex), and are frequently seen on multiple contiguous images. Some implants have a polyurethane coating covering the envelope surface. Because of this coating, a moderate to large amount of peri-implant reactive fluid is seen commonly.

Intracapsular rupture

Rupture of an implanted prosthesis is the most common complication associated with implants. Intracapsular rupture is well demonstrated on MR. The "linguine sign,"-multiple, low-signal intensity, wavy lines found within an implant-first described by Gorczyca et al,16 is specific for intracapsular rupture, and is the most reliable sign of this complication on MR imaging. These low-signal-intensity lines represent pieces of free-floating collapsed envelope surrounded by silicone gel, and are the equivalent to the ultrasound "stepladder" sign (figure 8).l6,13 An "inverted teardrop," "noose," or "keyhole" sign may be the only indication of rupture. This represents a loop-shaped, hypo-intense structure that is contiguous with the implant envelope, representing a small focal invagination of the implant shell with silicone on either side. This is caused by an extensive gel bleed or by gel that has escaped through a focal rupture.l9,20

A subtle sign of early intracapsular rupture and shell collapse is hypointense subcapsular lines which closely parallel the fibrous capsule on contiguous images, corresponding to a minimally displaced ruptured shell (figures 9 and 10). Phase-encoding artifact, caused by patient motion during the imaging process, may cause a hypointense curvilinear line which could be confused with subcapsular lines if located in a subcapsular position. Repeating the sequence with less patient motion or with a change in phase-encoding direction should eliminate the artifact. Hyperintense water droplets within the silicone are frequently seen in conjunction with teardrops, subcapsular lines, or the linguine sign. When seen alone, however, they are not indicative of rupture.

Extracapsular rupture

A diagnosis of extracapsular rupture is made when there is silicone outside of the fibrous capsule, either within the breast parenchyma or along the axilla or chest wall (figure 11). The localization of free-silicone collections can be helpful in preoperative planning. However, previous silicone injections or residual silicone from prior rupture or implant explantation cannot be distinguished from that of a new rupture.

A number of studies have examined the sensitivities and specificities of mammography, ultrasound, and MRI for the detection of implant rupture. However, only a limited number have compared results of all three modalities in a single study population. Gorczyca et al21 surgically placed forty single-lumen implants (20 ruptured/20 intact) within the abdominal walls of rabbits and imaged these implants using mammography, ultrasound, MR, and CT. The ruptures were all intracapsular, which would limit the usefulness of mammography. MR was found to have the best performance, with a sensitivity of 81% and a specificity of 93%. MR was followed by sonography (sensitivity 70%; specificity 92%) and mammography (sensitivity 11%; specificity 89%).2l

Everson et al,22 using mammography, sonography, MR, and CT evaluated 32 women with 63 implants. All patients subsequently underwent surgical explantation and 22 ruptures were found. MR again had the best performance (sensitivity 95%; specificity 93%), followed by sonography (sensitivity 59%, specificity 70%) and mammography (sensitivity 23%, specificity 98%).22 Berg et al14 evaluated 144 implants (122 single-lumen; 22 double-lumen) with surface-coil MR and ultrasound. These implants were subsequently surgically removed. Overall, MR imaging for rupture was found to have a sensitivity of 98% and a specificity of 9l% for patients with intact implants. Ultrasound was found to have a sensitivity of 65% for rupture, based on the stepladder sign and/or low-level echoes, and a specificity of 57%. An overall accuracy rate of 84% was determined for MR, and 49% for ultrasound.14 AR

Acknowledgement

The authors wish to thank Ms. Louise Logan for her excellent assistance in manuscript preparation.

References

1. Cronin TD, Gerow F: Augmentation mammoplasty: A new "natural feel" prosthesis. In: Transactions Third International Congress Plastic Surgery, pp 41-49. Amsterdam, Excerpta Medica, 1964.

2. Brenner RJ, Gorczyca DP: "The Silicone Controversy: A Brief History of the Problem." In: The Augmented Breast, Gorczyca DP, Brenner RJ (eds), pp 1-5. New York, Thieme, 1997.

3. Shaw de Paredes E: "The Postsurgical Breast. In: Atlas of Film Screen Mammography." Shaw de Paredes E. (ed), chapter 11. Baltimore, Williams & Wilkins, 1992.

4. Brogs CS: Fact and fiction about breast implant "bleed." Plast Reconst Surg 60:615-616, 1977.

5. Burkhardt BR: Capsular contracture: Hard breasts, soft data. Clin Plast Surg 15:521, 1988.

6. Hakelius L, Ohlsen L: A clinical comparison of the tendency to capsular contracture between smooth and textured gel-filled silicone mammary implants. Plast Reconst Surg 90:247, 1992.

7. Silverstein MJ, Handel N, Gamagami P, et al: Mammographic measurements before and after augmentation mammoplasty. Plast Reconst Surg 86:1126-1130, 1990.

8. Guether JM, Tokita KM, Giuliano AE: Breast conserving surgery and radiation after augmentation mammoplasty. Cancer 73:2613-2618, 1994.

9. Eklund GW, Busby RC, Miller SH, Job JS: Improved imaging of the augmented breast. AJR 151:469-473, 1988.

10. Cole-Beuglet C, Schwartz G, Kurtz A, et al: Ultrasound mammography for the augmented breast. Radiology 146:737-742, 1983.

11. Harris KM: Ultrasonographic findings of the failing breast implant. In: The Augmented Breast, Gorczyca DP, Brenner RJ (eds), chapter 11. New York, Thieme, 1997.

12. Berg WA, Caskey CI, Hamper UM, et al: Diagnosing breast implant rupture with MRI, ultrasound and mammography. Radiographics 13:1323-1336, 1993.

13. Debruhl ND, Gorczyca DP, Ahn CY, et al: Silicone breast implants: Ultrasound evaluation. Radiology 189:95-98, 1993.

14. Berg WA, Caskey CI, Hamper UM, et al: Single and double lumen silicone breast implant integrity: Prospection evaluation of MRI and US criteria. Radiology 197:45-52, 1995.

15. Rosculet KA, Ikeda DM, Forrest ME et al: Ruptured gel-filled silicone breast implants: Sonographic findings in 19 cases. AJR 159:711-716, 1992.

16. Gorczyca DP, Sinha S, Ahn C, et al: Silicone breast implants in vivo: MR imaging. Radiology 185:407-410, 1992.

17. Habal MB: The biologic basis for the clinical application of the silicones. Arch Surg 119:843-848, 1984.

18. Mukundan S, Dixon WT, Kruse BD, et al: MR imaging of silicone gel-filled breast implants in vivo with a method that visualized silicone selectively. J Magn Reson Imag 3:713-717, 1993.

19. Soo MS, Kornguth PJ, Walsh R, Elenberger CD: Intracapsular implant rupture: MR imaging findings of early shell collapse. Radiology 197(P):371, 1995.

20. Soo MS, Kornguth PJ, Walsh R, et al:

Complex radial folds versus subtle signs of intracapsular rupture of breast implants: MR findings with surgical correlation. AJR 166:1421-1427, 1996.

21. Gorczyca DP, DeBruhl ND, Ahn CY, et

al: Silicone breast implant ruptures in an animal model: Comparison of mammography MR

imaging, US, and CT. Radiology 190:227-232, 1994.

22. Everson LI, Parantasnen H, Detlie T, et al: Diagnosis of breast implant rupture: Imaging findings and relative efficacies of imaging techniques. AJR 163:57-60, 1994.