Downloaded from www.ajronline.org by 104.240.207.132 on 10/20/15 from IP address 104.240.207.132. Copyright ARRS. For personal use only; all rights reserved

711

Ruptured Gel-Filled Breast Implants: Sonographic

Katrina A. Rosculet1 Debra M. Ikeda1’2 Maureen E. Forrest3 Robert

M. Oneal4

Jonathan M. Rubin1 Deborah 0. Jeff ries3 Mark A. Helvie1

159:711-716,

Since

after re-

Presented at the annual meeting of the American Roentgen Ray Society, Orlando, FL, May 1992. 1 Department of Radiology, University of Michigan Hospitals, Box 0326, 1500 E. Medical Center Dr., Ann Arbor, MI 48109-0326. 2Present address: Division of Breast Imaging, Department of Radiology, Stanford University Medical Center, 300 Pasteur Dr., Am. S-072, Stanford, CA 94304. Address reprint requests toO. M. Ikeda.

Findings in 19 Cases

OBJECTIVE. The purpose of this study was to describe and illustrate the sonographic appearances of 19 ruptured silicone gel breast implants. MATERIALS AND METHODS. We retrospectively reviewed the sonograms of 16 patients with 19 ruptured silicone gel implants from two institutions. The ruptured implant was confirmed at surgery in 17 cases and by mammographic and clinical findings of a ruptured implant combined with biopsy findings of a silicone granuloma in two cases. Breast sonograms were available for review in all patients. The clinical presentation of each patient was recorded. The sonograms and mammograms were reviewed, and the findings were correlated with the surgical findings. In 16 of the 19 ruptured implants, mammographic findings suggested rupture, including lobulation of the contour of the implant and/or silicone extrusion into the breast parenchyma or axilla. In two ruptured implants, mammographic findings were normal, and in one case, no mammogram was available. In those three patients, palpable masses and clinical findings were suggestive of rupture. RESULTS. Sonography showed a unique echogenic appearance called echodense noise, in 17 of the 19 ruptured implants; in 10 of the 17, sonograms showed hypoechoic masses of extruded silicone also. In two ruptured implants, sonograms showed only the hypoechoic masses of extruded silicone gel. CONCLUSION. Our experience suggests that echogenic noise is a unique sonographic sign of ruptured silicone gel breast implants and may be caused by phase aberration related to the speed of sound being slower in silicone than in soft tissue. AJR

Received February 3, 1992; accepted vision April 14, 1992

Silicone

October

the first

breast

1992

implant

women

in the United States

silicone

gel breast

implants

was

placed

in 1 962,

have had silicone

approximately

gel breast

is one of the occasional

implants

complications

1 to 2 million

[1 ]. Rupture of their

of

use that

has received a great deal of attention recently. Rupture of the breast implant is often evident clinically, usually because of sudden breast deformity, change in breast consistency, or breast size reduction. Occasionally the diagnosis is difficult such as when patients have a breast mass, pain, discomfort, or a perception that the implanted breast has changed. In these cases, mammography may be useful, but is not always conclusive. The role of sonography in the evaluation of silicone gel breast implants is not clear, although some have advocated its use [2].

We reviewed

the sonographic

findings

in 19 ruptured

implants

in 16 patients.

of Radiology, St. Joseph Mercy Hospital,

5301 E. Huron River Dr., Ann Arbor,

MI

Materials

and Methods

48106. 4Department

of Plastic and Reconstructive

Sur-

gery, University of Michigan Hospitals, Ann Arbor, MI 481 09-0326, and Department of Plastic Surgery, St. Joseph Mercy Hospital, Ann Arbor, MI 48106. 0361 -803x/92/1 594-0711 © American Roentgen Ray Society

We retrospectively reviewed our log sheets of women with silicone gel breast implants who were seen at the University of Michigan Hospital between January 1 989 and January 1992 or at St. Joseph’s Mercy Hospital between October 1988 and January 1992. During the periods specified, 30,748 mammograms and 2330 breast sonograms were obtained at the University of Michigan Hospital, and 7377 breast sonograms and 80,536 mammograms were obtained at St. Joseph’s Mercy Hospital. Requirements for inclusion in the study were

712

ROSCULET

Downloaded from www.ajronline.org by 104.240.207.132 on 10/20/15 from IP address 104.240.207.132. Copyright ARRS. For personal use only; all rights reserved

(1) a ruptured implant found at surgery (1 7 cases) or mammographic and clinical findings of a ruptured implant with a silicone granuloma present on biopsy (two cases) and (2) breast sonograms available for review. Sixteen women with 1 9 ruptured implants fulfilled the study criteria.

Mammography

was performed

in 15 women

with state-of-the-art

equipment (GE/CGR 600T or GE/CGR SOOT unit, Columbia, MD, or Mammex Mag DC unit, Technomed USA, Bayshore, NY), and sonography was performed with either a Hitachi EUB400 (Tokyo, Japan;

1 3 women) or an Acoustic

Imaging (Phoenix,

AZ; three women)

unit

with 7.5-MHz transducers. Sonograms and mammograms were reviewed independently by three radiologists. The following data were recorded for each case from the clinical records and from review of the mammograms and sonograms: (1 ) reason for implant, (2) location of implant (subpectoral or retromammary), (3) clinical presentation, (4) contour of implant seen on mammograms, (5) presence or absence of silicone gel extrusion outside the implant shown on mammograms,

and (6) sonographic

findings

(echodense

noise, hypoechoic

masses,

or irregularity of the contour of the implant). The 1 6 women were between 33 and 54 years old (average, 45 years old). The implants had been in place for 6-20 years (average,

13 years). In 1 4 women (1 7 breasts) with suspected ruptured implants, the implant was removed. In two women (two breasts), biopsy showed a foreign body (silicone) granuloma; these two women also had mammographic and clinical findings suggestive of rupture. Fourteen women had implants for breast augmentation and two for reconstruction after mastectomy, for a total of 16 women in whom

I 9 implants were found to be ruptured implants

that ruptured

were

retromammary

at surgery.

Sixteen of the 19

and three

were

subpec-

toral. The patients underwent imaging for various reasons: palpable masses (seven breasts), mass and pain (one patient, two breasts), mass and breast deformity (one breast), change in breast shape (three patients, four breasts), pain (one breast), or screening mammography (three patients, four breasts). In 16 of the 19 ruptured implants, mammographic findings sug-

gested rupture, including a lobulated contour of the implant

and

ET AL.

silicone

AJR:159, October1992

extrusion

between

the implant

and fibrous

capsule

or be-

the axillary tissue planes (1 2 cases), extrusion into breast parenchyma (two cases), and a markedly lobulated contour (two tween

cases). Mammographic other did not undergo

findings were normal in two patients, and one mammography; palpable masses in these three

patients were examined

sonographically.

Results

At sonography a unique homogeneous low-level echogenic appearance, labeled echodense noise, was noted in 17(90%) of the 19 ruptured implants. This echodense noise, which corresponded to silicone gel leakage, did not allow imaging of structures beyond it (Figs. 1 and 2). Seven ruptured implants

masses

showed

this finding

(representing

alone,

and 1 0 showed

junction with the echogenic In two ruptured implants,

of silicone gel) in connoise (Figs. 3 and 4). sonography

dense noise, but did show hypoechoic to extruded conjunction

hypoechoic

conglomerates

did not show

masses

echo-

corresponding

silicone gel masses either alone (one case) or in with an irregular contour of the implant (one case).

The hypoechoic masses were separate from the implant, except in one patient who had a palpable mass corresponding to an irregular hypoechoic mass adjacent to the implant on sonography.

Surgery

confirmed

cone gel within a fibrous with a communication the implant (Fig. 5). In the two patients

that the mass

wall adherent between with

the mass

normal

contained

sili-

to the implant capsule, of silicone

mammographic

gel and findings

and palpable breast masses, echodense noise and hypoechoic masses were seen on sonograms, suggesting ruptured implants not seen on mammograms. Each patient had a surgically confirmed implant rupture at the site of the palpable mass, as detailed by the surgical report.

Fig. 1-49-year-old woman with a ruptured 1 l-year.old implant in left breast noticed softening of the breast after a fall from a horse 10 days caner. A, Lateromedial mammogram (cropped) of left breast reveals rupture of implant (Im), with streaks of silicone gel in breast parenchyma. B, Sonogram of left breast shows echogenic noise corresponding to extruded silicone gel (solid arrow), breast tissue (B), and silicone gel implant (Im). Note abnormal breast-implant boundary on either side of echogenic noise (open arrows). C, Sonogram of another region in left breast shows distinct echogenic noise corresponding to extruded silicone gel (arrow). Posterior to noise, all sonographic Information Islost.Hypoechoic siliconegel implant is seen adjacent to noise (Im).

RUPTURED

Downloaded from www.ajronline.org by 104.240.207.132 on 10/20/15 from IP address 104.240.207.132. Copyright ARRS. For personal use only; all rights reserved

AJR:159, October1992

Fig. 2-45-year-old A, from B, C,

SILICONE

BREAST

IMPLANTS

713

woman with a ruptured 7.yearold implant in left breast (cropped) of left breast shows extruded sIlicone gel (straight arrow). A small mass of silicone gel has migrated Implant to left upper extremity, adjacent to humerus (curved arrow). Sonegram of silicone gel mass In left upper extremity shows focal echodense noise (silicone gel mass) that completely obscures underiylng humerus. Gross surgical specimen shows ruptured silicone gel Implant contained In a fibrous capsule. Medloisteral

asymptomatic

oblique

Fig. 3.-36.year.old

mammogram

woman with ruptured

10-

year-old Implants noted masses and pain bilaterally. A, Medlolateral oblique mammogram (cropped) of left breast shows focal Iobulatlon of superior aspect of Implant (arrow), corresponding to a Icm mass In upper part of left breast B, Sonogram of left breast shows a hypoecholc mass (black arrow) with surrounding echodense noIse at site of Implant rupture (white arrows).

A

B

In eight of the nine women with a palpable mass, a sonographic abnormality (either echodense noise or hypoechoic

Discussion

masses

Sonography has been used to evaluate complications of breast implants including rupture, with past reports describing sonographic evidence of rupture as linear echoes through the implants representing invaginations of the implant envelope [3, 4] and a decreased anteroposterior diameter of the ruptured implants [3]. An irregular contour of the implant was seen in only one of our cases, perhaps because our group of patients included women who had subtle or confusing clinical findings, often referred for imaging to “rule out rupture,” and those in whom unsuspected silicone gel leaks were found at screening mammography. The unique echogenic noise that is such a striking feature

of silicone)

the remaining dense

noise

sonograms, region

corresponded

woman indicating

rupture

but no discrete

of the palpable

to the palpable

with a palpable

mass.

finding.

In

mass, areas of echo-

of the implant

were

seen

on

lesions could be identified

in the

The cause

mass

of the palpable

could not be ascertained from the report or the medical chart, but at surgery the implant was ruptured, as suggested by echodense noise seen over other (impalpable) regions. In one case, mammography showed a 1-cm conglomerate of extravasated silicone gel corresponding to a hypoechoic mass on sonography; aspiration under sonographic guidance yielded

silicone

gel (Fig. 6).

Downloaded from www.ajronline.org by 104.240.207.132 on 10/20/15 from IP address 104.240.207.132. Copyright ARRS. For personal use only; all rights reserved

714

ROSCULET

ET AL.

AJR:159, October1992

Fig. 4.-46-year.old woman with a ruptured 18. year-old implant noted nodularity in upper portion of right breast. A, Mediolateral oblique mammogram (cropped) of right breast shows silicone gel (arrow) extruded from ruptured implant. Note capsular calcification. B, Sonogram of right breast shows extruded silicone gel as a hypoechoic mass (arrows). Echogenic noise was noted in other regions (not shown).

Fig. 5.-33-year-old woman with a ruptured 9year-old implant noted a 2-cm mass in lower outer quadrant of right breast Sonogram of right breast shows a hypoecholc mass (arrow) thought to be extruded silicone gel adjacent to Implant (Im). At capsulectomy, a mass with macroscopic amounts of silicone gel communicated with implant and was adherent to fibrous capsule, representing a walledoff leak from the implant.

Fig. 6.-39-year-old of left breast A, Lateromedial implant and a I-cm B, Sonogram of sponding to both structures posterior confirmed rupture.

of many of our cases rarely has been described in the radiologic literature. Of note, however, ours is a retrospective study of a biased population consisting only of women who had ruptured implants at surgery. Herzog [5] described two women who had small nodules with “echodense shadows” behind them that corresponded histologically to silicone granulomata and silicone lymphadenopathy [5]. Levine and Collins [4] alluded to “echogenic confusion” on sonography in a single case of a grossly ruptured implant. Harris et al. [2] have presented a preliminary report describing this unique appearance associated with ruptured breast implants. We found this echogenic noise to be unique and different from the typical “dirty shadow” formerly ascribed to gas [6].

woman mammogram

with 15- to 20-year-old (cropped)

implants

noted a 1-cm mass in upper part

of left breast shows lobulation of superior contour of

silicone globule (arrow), corresponding to mass marked by a skin marker. left breast shows a hypoechoic mass (arrow), separate from implant, correpalpable mass and mass seen on mammography. Note sonographic breast to mass. Fine-needle aspiration of mass yielded silicone. Removal of implant

It is not analogous

to the shadow

seen behind

gas or stones,

because large masses of silicone gel or the intact implant clearly transmit sound (Figs. 4B and 6B). We hypothesize that the appearance is caused by aggregates of silicone with diameters on the order of a wavelength within breast paren-

chyma or axillary tissues. The speed of sound in silicone is significantly slower (990 m/sec) than in soft tissue (1540 m/ sec) [7]. This creates large differences in speeds of sound across the ultrasound beam when it propagates across tissue impregnated with silicone. These velocity discontinuities produce focal areas in the wave that advance ahead or lag behind and are out of phase with neighboring areas, producing a loss of coherence

of the beam,

destroying

beam

focus

RUPTURED

AJR:159, October1992

SILICONE

and, ultimately, the image [8]. This is termed phase aberration [8]. We considered the possibility that this noise is due to reverberations

phenomena Downloaded from www.ajronline.org by 104.240.207.132 on 10/20/15 from IP address 104.240.207.132. Copyright ARRS. For personal use only; all rights reserved

is, large

or “multipathing”

require

surfaces

the presence with

fluctuations

in the tissue.

of specular that

However,

reflectors,

are small

such

that

compared

with a wavelength. The small droplets of silicone produced by a ruptured implant do not meet these criteria. In fact, because of the difference in the speed of sound in silicone vs soft tissue, these small droplets would produce large fluctuations in phase, canceling any strong reverberations. We must then explain why some ruptured implants with obvious silicone gel extrusion on mammography do not show this sonographic appearance (noise), but manifest as irregular hypoechoic masses, which were proved at surgery to represent conglomerates of silicone gel outside the implant. When the beam encounters a large mass of silicone gel, that is, large compared with the wavelength, the sound wave reflects and refracts when the wave front is not perpendicular to the interface. The beam is bent by the boundary and propagates through the silicone gel in the wrong direction, but is still coherent. Therefore, one sees the boundary between the silicone and the breast parenchyma and structures behind the boundary, which appear normal but are actually in the wrong position. This is what happens when sound is focused behind cysts in routine clinical sonograms, but only to a greater degree [9, 10]. This explanation also accounts for bulges in implants not producing the noisy appearance, as in the walled-off leak shown in Figure 5. If the wave front is perpendicular to the silicone-breast parenchyma boundary, the sound beam is not bent but propagates much more slowly through the silicone gel and still retains coherence. Rupture was suggested with film-screen mammography in all but two cases, in agreement with a prior study suggesting a high accuracy of mammography in showing ruptures of silicone gel implants [1 1]. It should be noted, however, that our population was biased as it contained only patients with ruptured implants. We speculate that the area of rupture in our two cases with normal mammographic findings and abnormal sonographic findings was not imaged on the mammogram because of the location of the rupture (upper inner quadrant in one and upper outer quadrant in the other), which may have been hidden by the dense silicone gel implant. As the woman with the implant that ruptured in the upper inner quadrant also had severe capsular formation, the region may not have been optimally imaged because of the hardness of the breast and difficulty in positioning. Many of our other cases showed extrusion of opaque silicone gel into the surrounding tissues on mammography, a finding previously described in case reports of ruptured implants [1 2-1 5]. On the other hand, sonography showed ruptured implants (echodense noise) as the cause of palpable breast masses despite normal mammographic findings in two of our cases. Lobulation of the contour of the implant also was seen in some of our cases. However, when a lobulated contour on mammography is the only radiologic sign suggesting an implant has ruptured, the diagnosis is not certain because falsepositive diagnoses of ruptured implants can be due to hernia-

BREAST

IMPLANTS

715

tion of the intact implant through a fibrous capsule that can form around breast implants [1 6]. Clinically, it may be difficult to identify ruptured implants in women with fibrous capsules forming

around

breast

implants.

The Baker

classification

ca-

tegorizes encapsulation as follows: grade I, normal; grade II, palpable firmness with a natural appearance; grade Ill, firm with visible distortion; and grade IV, obvious spherical distortion [1 7]. In some cases, particularly those with Baker Ill-lV encapsulation of implants, sonography can be technically most difficult at the inframammary crease because of the inflexibility of the transducer combined with a stiff and immobile implant. We have found this can usually be overcome by turning the patient onto her side or having her raise her arm higher above her head. Implant lobulations detected with mammography may raise the clinical question of whether the implant is lobulated, herniated,

or ruptured.

We speculate

that sonography

would

be particularly valuable in these patients to show that simple implant lobulations caused by capsular formation are continuous with the rest of the intact implant, and that herniations of intact implants through the fibrous capsule are not ruptures; in either

of these

two

scenarios

involving

implant

contour

lobulations without rupture, we speculate sonography would not show the echodense noise we described. On the other hand, it would be difficult to distinguish herniation from walledoff leaks from implants (Fig. 5). The diagnosis of rupture of a silicone gel breast implant may be missed on mammography, as the rupture site may be hidden behind the radiodense implant or not included in the field of view. However, mammography remains critically important in the evaluation of breasts containing implants to evaluate the breast parenchyma for cancer [1 8]. Sonography provided valuable additional information when a ruptured implant was clinically suspected and mammographic findings were normal or when mammographic findings were equivocal, as in lobulated contours. The echodense noise produced by silicone gel dispersed in soft tissue is a valuable sonographic sign seen in most of our cases. Of note, however, all of our patients who were identified retrospectively for this investigation had either abnormal mammographic findings or clinical signs of ruptured implants and had surgical confirmation of rupture, producing sample bias into the population studied. The value of this sign in women with mammographic findings not suggestive of ruptured implants and normal findings on physical examination of the breast is yet to be determined. As the sensitivity and specificity of this sonographic sign are not known, further study of this sign is needed.

ACKNOWLEDGMENTS We thank Cynthia E. Sims and Melanie L. Shaw for secretarial support in the preparation of this manuscript and J. Brian Fowlkes for technical consultation.

REFERENCES 1. Department bulletin.

of Health and Human Services. Background

information

Food and Drug Administration on the possible health risks of silicone

ROSCULET

716

breast implants (rev. February 8, 1991). Rockville, MD: Department of Health and Human Services, 1991 2. Harris KM, Ganott MA, Shestak K, Loskin W, Tobon H. Detection of silicone implant leaks: a new sonograpic sign (abstr). Radiology 1991;181(P): 134-135

Downloaded from www.ajronline.org by 104.240.207.132 on 10/20/15 from IP address 104.240.207.132. Copyright ARRS. For personal use only; all rights reserved

3. Van Wingerden JJ, van Staden MM. Litrasound mammography

in prosthe-

co-related breast augmentation complications. Mn Plast Surg I989;22 :32-35 4. LevIne RA, Collins TL Definitive diagnosis of breast implant rupture by ultrasonography. Pleat Reconstr Surg I99I;87 :1126-1128 5. Herzog P. SIlicone granulomas: detection by uftrasonography (letter). Plast Reconstr Surg 1989;84: 856-857 6. Rubln JM, Adler RS, Bude RO, Fowlkes JB, Carson PL. Clean and dirty shadowing at US: a reappraisal. Radiology I991;181 :231-236 7. Table of ultrasonic properties. Richmond, WA: Metro Tek Inc., 1982 8. Trahey GE, Frelburger PD, Ng G, Sullivan DC. The impact of acoustic velocity variations on target detectability in ultrasonic images of the breast. Invest Radlol 1991;9:782-791 9. Robinson DE, Wilson LS, Kossoff G. Shadowing and enhancement in ultrasonic echograms by reflection and refraction. JCU J Clln Ultrasound

ET AL.

AJR:159,

1819:181-188

10. LaFollette PS Jr, Ziskin MC. Geometric and intensity distortion in echography. Ultrasound Med Blol 1986;12 :953-963 11. Andersen B, Hawtof D, Alani H, Kapetansky 0. The diagnosis of ruptured breast implants. Plast Reconstr Surg 1989;84 : 903-907 12. Scott lR, Muller NL, Fitzpatrick DG, Warren Burhenne U. Ruptured breast implant: computed tomographic RadiolJ 1988;39: 152-154

13. Theophelis rupture.

LG, Stevenson

Plast Reconstr

Surg

and mammographic

TR. Radiographic 1986;78

findings.

evidence

Medical

Publishing,

Jolnt statement

on mammography

Manuscripts

Beginning with the January 1 993 issue, accepted papers to be published in the AJR will be edited and printed by using computer diskettes rather than hard copies ofthe manuscripts. the editing

and publication

process

and result

in less delay

between

submission of manuscripts and their appearance in the Journal. Also, with the new system, authors will receive page proofs rather than galleys before publication. This will enable them to see exactly what their final article will look like. Effective immediately, upon acceptance of manuscripts for publication, authors will be asked to send a copy of their revised manuscript on a 31/2-in. or 5#{188}-in.DOS-compatible diskette

in Wordperfect

or ASCII

format

along

with

the hard copy

of the revision.

Articles

will

be edited directly on the diskettes. Changes will be highlighted electronically; as before, a hard copy of the edited manuscript and figure proofs will be mailed to authors for approval before publication. Note that hard copies of manuscripts will still be required for the peerreview

and revision

of breast implant

: 673-675

1990;181

18. American College of Radiology. lines. ACR Bull 1989;45:1, 4

Instead of Hard-Copy for Editing and Printing

will expedite

Can Assoc

14. Apesos J, Pope TL. Silicone granuloma following closed capsulotomy of mammary prosthesis. Ann Plast Surg 1985;14 : 403-406 15. Ho WC. Radiograhpic evidence of breast implant rupture (letter). Plast Reconstr Surg 1987;79:1009-1010 16. Smith OS. False-positive radiographic diagnosis of breast implant rupture: report of two cases. Ann Plast Surg I985;14:266 17. Bostwick J. Plastic and reconstructive breast surgery. St Louis: Quality

AJR to Use Diskettes

This change

October1992

processes.

The AJR already takes great pride in the speed with which the peer-review process is accomplished (mean time, 3.4 weeks) and in the amount of time between submission of manuscripts and their final appearance in print (6 months). This is nearly twice as fast as many other journals. With the new system, these intervals will be even shorter. Authors will see their work in print faster, and the timeliness of the Journal will be greater than ever. Robert N. Berk Editor-in-Chief

guide-