| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Office of Surveillance and Biometrics, Center for Devices and Radiological
Health, Food and Drug Administration, 1350 Piccard Dr., HFZ-541, Rockville, MD
20850.
2
Department of Radiology, University of California at San Diego School of
Medicine, 410 W. Dickinson St., San Diego, CA 92103.
3
Division of Breast Imaging, Department of Radiology and Greenebaum Cancer
Center, University of Maryland School of Medicine, 22 S. Greene St.,
Baltimore, MD 21201.
4
Department of Radiology, Duke South Hospital, Duke University Medical Center,
Box 3808, Second Floor, Red Zone, Durham, NC 27710.
Received February 14, 2000;
accepted after revision March 24, 2000.
The opinions or assertions presented herein are the private views of the
authors and are not to be construed as conveying either an official
endorsement or criticism by the United States Department of Health and Human
Services, the U. S. Public Health Service, or the Food and Drug
Administration.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Women identified as part of a National Cancer Institute cohort study on breast implants, living in the Birmingham, AL, area were invited to undergo MR imaging of their current silicone gel breast implants at the Kirklin Clinic at the University of Alabama at Birmingham. Three radiologists independently examined and rated all MR images for signs of implant rupture and extracapsular silicone.
RESULTS. A total of 344 women with silicone gel breast implants underwent MR imaging. Breast implant rupture was reported by at least two of three radiologists for 378 (55.0%) of the 687 implants in this study. Another 50 implants (7.2%) were rated as indeterminate (suspicious) for rupture. A majority of women in this study, 265 (77.0%) of 344, had at least one breast implant that was rated as ruptured or indeterminate. Radiologists also agreed that silicone gel could be seen outside the fibrous capsule that forms around the implant in 85 (12.4%) of the 687 implants affecting 73 women (21.2%). Factors that affected implant rupture were implant age and location (submuscular or subglandular). The median implant age at rupture was estimated to be 10.8 years with a 95% confidence interval of 8.4-13.9 years.
CONCLUSION. The prevalence of silent or occult silicone gel breast implant rupture is higher than was previously suspected. Most women in this study had MR imaging evidence of at least one ruptured silicone gel breast implant.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
One study of 749 women with breast implants in place for a mean of 7.8 years indicated that 5.7% of those women underwent breast surgery because of a ruptured implant [3]. In a study of 317 Scottish women with implants, 10 (3.1%) had implants replaced after implant rupture [4]. These studies enumerated women who went to surgery with an indication of implant rupture but not women who had implants explanted for other reasons and were incidentally found to have ruptured implants.
Studies describing the status of implants after explantation have reported a much higher prevalence of implant rupture. The prevalence of implants that were not intact was reported to be between 23% and 65% of explanted implants [2, 5,6,7].
The discrepancy between studies of women with an indication of implant rupture before surgery and the higher prevalence of ruptured implants found in explantation studies could be attributed to referral bias in the population having their implants removed. However, this discrepancy could be caused by a mammographically or clinically silent (asymptomatic) rupture. A retrospective examination of screening mammograms from 350 asymptomatic women with breast implants indicated that 16 women (4.6%) had ruptured implants [8]. Mammography is the least sensitive imaging method for examining breast implant rupture with a sensitivity of 11-69% described in studies comparing imaging methods [9,10,11]. MR imaging has been reported to have a sensitivity of 39-76% when radiologists used a body coil, and from 52-95% when they used a breast coil [7, 11,12,13,14,15,16]. In recent MR imaging studies in which radiologists used a breast surface coil and had validated signs of rupture, the sensitivity was 74-94% and the specificity was 85-98% [7, 13, 16]. The specificity reported in these papers is likely underestimated and the sensitivity overestimated because early "learning curve" cases in which rupture was less completely understood than it is now were included.
In our study, an unreferred population of women underwent MR imaging with a breast coil to ascertain the current status of their silicone gel breast implants. Three radiologists independently examined the images for evidence of breast implant rupture. Radiologists also evaluated whether silicone gel had migrated outside the fibrous capsule that forms around the breast implant. Rupture prevalence by implant type (single or standard double lumen), manufacturer, implant location (subglandular or submuscular), and implant age was also evaluated.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The protocol for this study was reviewed and approved by five institutional review boards (NCI Special Studies IRB; Food and Drug Administration Research Involving Human Subjects Committee; University of California, San Diego, Human Subjects Committee; University of Alabama at Birmingham Institutional Review Board; and Abt Associates Institutional Review Board). All participants signed a detailed informed consent document. A report with results of the MR imaging was mailed to each participant and to a physician of her choice. A certificate of confidentiality for the study was obtained from the United States Department of Health and Human Services.
MR Imaging
Women were scheduled for MR imaging at the Kirklin Clinic at the University
of Alabama at Birmingham. We scanned on a 1.5-T scanner (revision 8.2, SIGNA
Horizon; General Electric Medical Systems, Milwaukee, WI) using a dedicated
General Electric bilateral phased array breast surface coil. The goals were to
determine whether implants were ruptured and whether any extracapsular
silicone was present. After a T2-weighted scout sequence, four sequences were
performed on each breast independently, for a total of nine pulse sequences
per patient (duration, about 60 min). Four sequences were used. The first was
an axial T2-weighted fast spin-echo inversion-recovery sequence with water
suppression over a 16.4-cm cephalocaudad distance centered on the breast
(TR/TE, 3000/156; inversion time, 180 msec; echo train length, 16; field of
view, 20 cm; slice thickness, 4 mm; matrix, 256 x 192; excitations,
one). The second sequence was an axial T2-weighted fast spin-echo sequence
with silicone suppression over the same slices as prior series (3000/156;
excitations, one). The third was a sagittal T2-weighted fast spin-echo
sequence with water suppression including the portion of the implant showing
folds (3000/224; field of view, 16 cm; slice thickness, 3 mm; matrix, 256
x 256; excitations, two). The fourth sequence was an axial fast
spin-echo T2-weighted sequence with water suppression including portions of
the implant showing folds with the same parameters as the third sequence. The
goal of this sequence was to look carefully at high resolution in folds
outside the implant for signs of silicone gel, which is the most sensitive
sign of rupture. An additional initial short tuning sequence was obtained for
the final 101 patients to offset a problem with one of the shim gradient
coils. Sequences were repeated as necessary (average, 10.1 sequences per
initial study). Technologists were trained in the study protocol by the study
radiologist. After the in-person training, the study radiologist was available
by telephone for consultations at any time. Images were sent to the study
radiologist by express mail at least once per week.
MR Evaluation of Rupture
The study radiologist and two consulting radiologists reviewed the 359
patients independently. Criteria to determine rupture were discussed by the
radiologists before the consulting radiologists began their evaluation.
Radiologists examined images for signs of rupture, including
"linguine" sign
[14,
18], "wavy line"
sign or double "wavy line" sign
[7], "anterior
spiculation" sign [7],
"subcapsular line" sign
[15,
16], "keyhole"
sign [7,
15,
16], "inverted
teardrop" sign [15,
16,
18], "noose" sign
[13,
19], "pull-away"
sign [7], and the "open
loop" sign [10]. The
signs previously noted, with different terminology from different authors,
describe only two basic phenomena that, considered together, are the criteria
we used to describe ruptured implants. The first appearance is that of an
implant elastomer shell fully collapsed and enveloped by the silicone gel it
once contained. This shows up on MR images as a wavy internal dark line
representing the implant elastomer shell, surrounded by the silicone gel that
has escaped from the implant. The second appearance is of a thin layer of
silicone gel between the implant elastomer shell and the internal surface of
the fibrous capsule, or of silicone gel outside the implant as a whole, yet
collecting within infoldings of the implant elastomer shell that protrude into
the implant itself. All degrees of collapse were grouped together for purposes
of analysis. Double-lumen implants were considered ruptured by the same
criteria: the appearance of silicone gel outside the implant as a whole. Fluid
signal mixing with gel alone was not considered indicative of rupture,
although for double-lumen implants this feature may indicate isolated failure
of the inner of the two shells. MR imaging only infrequently showed evidence
of silicone-fluid bleeding through an intact shell. Findings indeterminate for
rupture included one or two images with a possibility of gel in a fold. The
presence of extracapsular silicone was noted when observed.
When available, implant type, manufacturer, style, catalog number, and serial number were made known to each radiologist before evaluation. In cases in which the imaging contradicted any provided information, the implant-type evidence from the images was used, and that information was provided by the study radiologist to the other radiologists. Single-lumen silicone gel—filled implants were graded as "ruptured" when silicone gel was seen outside the implant. Standard double-lumen implants were graded as ruptured when silicone gel was seen outside the outer shell. Implants were graded "indeterminate" when there was suspicion, but not certainty, of rupture. Implants were graded as "no evidence of rupture" when they appeared intact.
Readings from each radiologist on rupture status were evaluated for agreement with other radiologists in a pairwise fashion with the weighted kappa statistic. A consensus reading was computed by voting the readings from radiologists and reporting the majority consensus. In the event that readings spanned the range from no evidence, to indeterminate, to ruptured, the consensus was "indeterminate." Likewise a consensus reading on the presence of extracapsular silicone was determined by majority vote.
Assignment of Implant Generation
Each implant was classified according to implant generation as defined by
Peters et al. [20] when enough
information was available. Implant generation was determined on the basis of
implant characteristics, with first generation implants being early
thick-shell, mostly thick-gel implants; second generation being thin and
intermediate shells that replaced the first generation, excluding "low
bleed" implants; and third generation being low bleed implants.
Information used to define the generation was implant model, implant catalog
number, implant serial number, and, in some cases, MR imaging appearance.
Statistical Methods
Logistic regression was used to model the probability that at least one of
the implants in a woman was ruptured or was indeterminate (i.e. the units of
observation were women, not implants). Factors considered for the logistic
regression model were implant age, implant type, location and manufacturer,
site of the surgery practice, and their two-way interactions. Another factor
considered was implant generation. In the analysis, we used only women with
two implants that had the same values for all factors. We chose a final model
on the basis of tests of significance using analysis of deviance. To interpret
the odds ratios from the regression, we determined that an odds ratio of 1.0
indicates no association between the characteristic and implant rupture. An
odds ratio of 2.0 would indicate that women with the characteristic have twice
the odds of occurrence of a ruptured implant as those who do not have the
characteristic.
Using the logistic regression models, we estimated the implant age at which the probability of rupture was 0.5 and called this the median age of rupture. We estimated this quantity by backsolving for implant age in the equation, relating the log odds of rupture, or logit, to the linear predictor. This method is commonly used to estimate the median effective dose in dose-response studies [21]. We furthermore estimated the implant age for probabilities of rupture other than 0.5 and plotted the probabilities against these estimates to form a survival curve. Standard errors of the estimated ages were computed by applying the delta method to the function isolating age and its theoretic variance matrix evaluated at the maximum likelihood estimates of the parameters. These standard errors were used to form 95% confidence bounds on the survival curve.
The logistic regression model used is equivalent to a survival model on the time to rupture. Time to rupture was either left-censored, when the implant was ruptured, or right-censored, when the implant was not ruptured. In the logistic regression model, the log odds of rupture was modeled as linear in the log of implant age. The logistic model corresponds to assuming that left- and right- censored times to rupture have a log—logistic distribution [22].
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
25 = 3.827,
p = 0.575) or year during which their initial implantation occurred
(24 = 1.571, p = 0.814). Neither were
there differences in women undergoing MR imaging compared with the others with
respect to the first implant type (single lumen, double lumen, not known)
(22 = 1.210, p = 0.546) or with respect to
the manufacturer of their first implant (26 =
3.720, p = 0.714). The 359 women who accepted and under-went the MR imaging did not differ, with respect to their opinion on whether their current implants were ruptured, from the 86 women who either actively refused the examination or scheduled the examination but did not come to their appointment. Although 31 women (8.6%) who subsequently under-went the examination reported that they thought their current implants were ruptured, seven women (8.0%) invited but declining the examination reported that they thought their implants might be ruptured (chi-square test, p = 0.674). This similarity indicates that the population accepting the MR imaging was not biased with respect to suspecting implant rupture.
Women in the MR imaging cohort had a mean age of 51.4 ± 8.4 years (range, 33-76 years) at the time of the examination. The reason for mammoplasty in the medical record was cosmetic for most women (85%), but 14% had the implants for medical reasons, usually for fibrocystic breasts. Women received their first implants between 1970 and 1988, with a median year of 1981. One woman had only one implant (single lumen), and two women had a single-lumen silicone gel implant in one breast and a standard double-lumen implant in the other. Twenty-four women (7%) had reported that one or both original implants had been surgically removed and replaced. The average breast implant age in this study was 16.5 ± 3.4 years for the 677 implants that had this information available, with a range from 6.4 to 28.0 years since implantation and a median age of 16.4 years.
Table 1 compares the interpretations from the three radiologists. Overall, the radiologists' consensus was that 378 (55%) of the 687 implants in this study were ruptured and that another 50 (7.2%) were indeterminate. Two hundred fifty-nine implants (37.7%) were intact. When viewed on the basis of women, 236 women (68.6%) had at least one ruptured implant. Overall, 108 women (31.4%) had both implants intact, 94 women (27.3%) had one ruptured implant, and 142 women (41.3%) had two ruptured implants. If women with ruptured implants and implants suspicious for rupture were included, then 265 women (77.0%) were affected. A high level of agreement was found between radiologists when compared in pairwise fashion as measured by the weighted kappa statistic. In no case was the kappa statistic less than 0.88, indicating almost perfect agreement [23]. Figures 1,2,3 are representative images of single- and standard double-lumen implants that were rated as ruptured by all three radiologists.
|
|
|
|
Migration of silicone beyond the fibrous capsule was observed in 85 breasts. Although affecting 85 (12.4%) of 687 breasts in this study, silicone migration beyond the fibrous capsule in one or both breasts was found in 73 women (21.2%). Figure 4 is a representative MR image of migration of silicone gel from the intracapsular to the extracapsular space. In all but one breast with extracapsular silicone, the radiologists' consensus was that the implant was ruptured or suspicious for rupture. The prevalence of extracapsular silicone for ruptured implants was 84 (22.2%) of 378 ruptured implants. The agreement between radiologists on extracapsular silicone was not as high as for rupture. When compared in a pairwise fashion, the kappa statistic for extracapsular silicone in the left or right breast was between 0.50 and 0.65 (moderate to substantial agreement) [23].
|
Table 2 shows implant status by implant type, implant age (shown in 5-year groups), implant location (subglandular, submuscular), implant manufacturer, and implant generation. The prevalence of rupture in each of the implants and in women is shown. In this univariate analysis, implant age, location, and manufacturer were all factors that had a bearing on rupture. In particular, rupture prevalence increased as implant age increased from 6 to 20 years, but then decreased as age increased to greater than 20 years. Most implants more than 20 years old were from Dow Corning (Midland, MI); Dow Corning implants ruptured less often than implants from other manufacturers. Few first and third generation implants were found in this study; the majority of implants were from the second generation. The average age of each generation from first to third was 25.4 ± 1.1, 16.3 ± 3.3, and 7.4 ± 2.6 years, respectively.
|
A multivariate logistic regression model was developed and used to estimate odds ratios of rupture, with each factor adjusting for the other factors in the model. Table 3 lists the estimated odds ratios given by the final model for the outcomes of rupture and of rupture or indeterminate (suspicion of). The estimated odds ratio for a 33% increase in implant age was significantly greater than one for both outcomes according to the corresponding 95% confidence intervals and indicated that increases in implant age increase the risk of these outcomes. Each odds ratio for implant age was calculated from a log odds ratio that was an average of log odds ratios specific to each manufacturer, with the average based on weights proportional to the number of women with implants from these manufacturers. The manufacturers Cox-Uphoff International (Carpenteria, CA) and McGhan/3M (Santa Barbara, CA) were excluded from the model because of the small number of women with implants from these manufacturers (Table 2). Odds ratios for implant location were computed for each site because of variation by site. For the outcome of rupture or indeterminate rupture, the estimated odds ratio for implants located in the submuscular position compared with the subglandular position was significantly greater than 1 for site 2, but not for site 1. For the outcome of rupture, the estimated odds ratio was significantly greater than 1 for both sites.
|
The logistic model was further used to compute a survival curve for implants (Figs. 5 and 6). Unlike the survival curves from survival analyses in which probabilities of surviving are estimated at fixed times, in these survival curves, times (i.e., implant ages) are estimated at fixed probabilities. The curves plotted are estimates of implant age obtained at average values of the other factors, in which the average is based on weights for groups within factors that are proportional to the number of women within the groups. (For location effects by site, the weights for location are additionally weighted by site.) The estimated median age of rupture was 10.8 years with a 95% confidence interval of 8.4-13.9, and that of rupture or indeterminate was 9.7 years with a 95% confidence interval of 7.2-13.0.
|
|
The logistic regression model was fit to all women (n = 304) who had two implants with the same values for site and for implant age, type, location, and manufacturer (Cox-Uphoff International and McGhan/3M excluded). The final model was determined by first including all main effects for these factors and then by using analysis of deviance to test for significant two-way interactions. Significant interactions of implant age by manufacturer and implant location by site were included in the model. The interaction of implant location by implant type was also significant but not included because it led to instability in model estimates. Implant generation was also significant but not included because it was highly correlated with implant age, and most implants were of the second generation.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Factors that were associated with implant rupture were implant age and location. These factors were also associated with an MR imaging outcome of ruptured or indeterminate. A conservative estimate of median age of implant rupture was 10.8 years. The association of rupture with implant location indicates that submuscular implants were more likely to be ruptured than subglandular implants.
For some manufacturers, rupture prevalence was based on few implants (e.g., Cox-Uphoff International and McGhan/3M), and these data should be interpreted cautiously. Other manufacturers such as Surgitek (Medical Engineering, Racine, WI) had a large number of implants in this study population. It is likely that local variations among surgeons in the preference for manufacturers played a role in the uneven distribution of implants from different manufacturers. Because of limited resources for this study, we could not pursue additional sites that may have provided equal representation of implant manufacturers. A high prevalence of rupture was seen across all manufacturers, excluding McGhan, ranging between 45% and 82%.
Peters et al. [20] addressed the concept of implant generation as a possible factor in implant rupture. Difficulties with this approach have not yet been addressed fully in the literature and are beyond the scope of this work, mainly involving the definitions of their generations and the ranges of years over which they extended. Notwithstanding those objections, using their basic definitions, Peters et al. thought that 91.8% (631/687) of the implants in this study were second generation (i.e., thin or intermediate shell thickness, not early thick shell and thick gel, and not low bleed). We considered generation a factor in a logistic regression model but excluded it because implant age is associated with calendar year and therefore measures similar information. In this study all first generation implants (thick shell and thick gel) were manufactured by Dow Corning. Reported to have lower rupture rates than succeeding generations [20], first generation implants (11/24) ruptured more frequently than did second generation implants manufactured by Dow Corning (10/30). Nonetheless, the first generation rupture prevalence of 11 (45.8%) of 24 is less than the overall rupture prevalence of 55.0%. All 24 first generation implants were more than 20 years old, which contributed to the relatively lower rupture prevalence observed among implants more than 20 years old (Table 3). Another possible explanation is that many of the older implants that were inferior have been explanted and explanted implants would not be included in the MR study. This hypothesis is consistent with the observation that the proportion of women reporting an explantation surgery was less for women in the MR imaging study (7%) compared with all women completing the questionnaire (33%).
Another finding of concern was the high prevalence of migration of silicone gel from the fibrous capsule that surrounds the implant. Silicone had migrated beyond the capsule in at least one breast in 21% of the women in this study. Numerous cases of silicone migration from the capsule have been reported in the literature [2, 7]. Any association of silicone migration or breast implant rupture with disease has not been specifically investigated because the status of women with respect to implant rupture or silicone migration has been unknown in studies that have examined risk of disease in women after mammoplasty with silicone implants [24, 25]. The issue of health consequences of free silicone is poorly understood because of the lack of studies in which implant status and health status are both known.
The agreement among the three radiologists in this study as to the status of implants was quite high. MR imaging has been shown to be the most sensitive and specific method of imaging for breast implant rupture [7, 9,10,11, 13, 16]. These results on implant status may be viewed with a high degree of confidence particularly because the results are the consensus of at least two of the three reviewing radiologists. The radiologists' agreement on extracapsular silicone was moderate to substantial. A study to examine the reason for this is planned.
This study had some shortcomings. We could not rule out all sources of bias. We cannot know what effect litigation had on participation in this study: anecdotal evidence suggests that some litigants were encouraged to participate in the NCI study and others were discouraged by their attorneys or others. Although 303 (33.4%) of 907 of women interviewed for this study indicated that they had their implant removed, only 7% of those receiving MR imaging reported having their implants removed and replaced. This discrepancy is because women who had implants removed and not replaced or removed and replaced with saline implants would not be eligible to participate in the MR imaging study. This requirement would result in culling out older implants that were removed because they were ruptured or for other reasons and may have led to an underestimation of the rupture prevalence. However, participants were similar to those declining to participate with respect to their opinion on the status of their current implants. As previously mentioned, implant manufacturers were not represented equally in this study. Although MR imaging is considered the best method for imaging breast implants for rupture, it is not perfect. This study did not include ascertaining implant status for women who subsequently decided to have their implants explanted. Because of the lower sensitivity of MR imaging in depicting uncollapsed rupture, it is likely that these interpretations present an underestimate of the rupture prevalence [7, 13, 16].
Given the past belief that implant rupture was rare, but the current evidence that the prevalence of implant rupture is high, we believe it is time to reevaluate the need to screen women for implant rupture and to develop recommendations for implant removal or replacement in the event of a rupture. Considerable disagreement exists over the appropriate treatment for women with ruptured implants. Individual plastic surgeons have recommended prophylactic explanation of silicone gel breast implants before 8 years after implantation, regardless of evidence of rupture, to avoid the increasing risk of rupture as the implant ages [5]. Others have advised that silicone should be removed from patients symptomatic for connective tissue disease or with other nonspecific illnesses like chronic fatigue syndrome or fibromyalgia [26]. There is agreement that if implant rupture is symptomatic for local complications (breast deformity, siliconoma or granuloma, pain, and migration), the implant should be explanted with the option of replacement [27]. Plastic surgeons will remove an implant if it is ruptured and causes a cosmetic defect, but some may question the health benefit of removing a ruptured implant if it is contained within the scar capsule [28]. Conversely, some plastic surgeons have argued that rupture should always be treated aggressively to prevent extracapsular spread because silicone gel is more difficult to remove once it has migrated [29, 30]. Both the potential for distant migration and subsequent inflammatory reactions have been cited as a reason to explant ruptured implants [31, 32]. Other studies have maintained that silicone poses little or no health risk [33] or that the potential for unnecessary surgical procedures is worrisome [34].
Women considering silicone gel breast implants, which are presently available to some women in clinical studies, should be informed of the possible risk of implant rupture and the possibility that this may necessitate additional surgery.
Acknowledgments
We thank women who participated in this study, some of whom received
unsettling news on the status of their breast implants. We also appreciate the
contributions of Louise Brinton and Jay Lubin from the NCI, who graciously
provided access to the NCI cohort and advice on study design. Mary Cay Burich,
Kathryn Vargish, Marilyn Sawyer, and Jon Schmalz from Abt Associates, Inc. in
Chicago are to be commended for their diligent work on this complex study. We
thank those plastic surgeons who shared their records with the NCI, making
this study possible. Finally, we thank all the sponsors of this research in
the Department of Health and Human Services, National Institutes of Health,
and particularly the Office of Women's Health at the Food and Drug
Administration.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J.-W. Kwek, H. Choi, J. Ma, and M. J. Miller Gel-Gel Double-Lumen Silicone Breast Implant: Mimic of Intracapsular Implant Rupture Am. J. Roentgenol., October 1, 2006; 187(4): W436 - W437. [Full Text] [PDF]
|
|
|
|
 |
|
 R. L. Levine, T. C. Allen, J. Cartwright Jr, and P. T. Cagle Silicone Thorax Due to a Ruptured Breast Implant Chest, May 1, 2005; 127(5): 1854 - 1857. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. R. Holmich, S. Friis, J. P. Fryzek, I. M. Vejborg, C. Conrad, S. Sletting, K. Kjoller, J. K. McLaughlin, and J. H. Olsen Incidence of Silicone Breast Implant Rupture Arch Surg, July 1, 2003; 138(7): 801 - 806. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. A. Berg, T. K. Nguyen, M. S. Middleton, M. S. Soo, G. Pennello, and S. L. Brown MR Imaging of Extracapsular Silicone from Breast Implants: Diagnostic Pitfalls Am. J. Roentgenol., February 1, 2002; 178(2): 465 - 472. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
