Rietjens, Mario; Villa, Gaetano; Toesca, Antonio; Rizzo, Stefania; Raimondi, Sara; Rossetto, Fabio; Sangalli, Claudia; De Lorenzi, Francesca; Manconi, Andrea; Matthes, Angelo Gustavo Zucca; Chahuan, Badir; Brenelli, Fabricio; Bellomi, Massimo; Petit, Jean Yves
Implant-based reconstruction is the most popular technique of breast reconstruction after mastectomy. Implant rupture is one of the most important sequelae, resulting in a significant deterioration of aesthetic outcomes and requiring a further surgical procedure. Although extracapsular silicone leakage has been investigated in the past for correlation with autoimmune diseases, a rare form of lymphoma, or other locoregional or systemic events, no increased risk of connective tissue diseases or cancer is evident in women with extracapsular ruptures.1–6 However, revision procedures after ruptured implants may increase the risk of local adverse events such as capsular contracture, scarring, pain, and aesthetic failure.2,7,8
The true prevalence of implant ruptures in asymptomatic patients is unknown.1,7 Implant rupture can occur in the early period after implantation or following a long interval. Its prevalence increases with implant age. Hölmich et al.9 found 2 percent of ruptured prostheses at 5 years’ and 15 to 17 percent at 10 years’ follow-up, whereas Robinson and colleagues10 observed 11 percent of ruptured implants at 8 years, reaching 95 percent at 20 years.
Although breast symptoms and abnormal physical examination can lead to implant rupture diagnosis,11 neither breast symptoms nor abnormal physical examination represents a reliable predictive value.12 Recently, a more accurate diagnosis was achieved by radioimaging techniques.13–15 Ultrasonography presents a variable sensitivity and specificity, ranging from 54 to 67 percent and 64 to 92 percent, respectively.13,16,17 Mammography is more reliable in cases of long-term implant rupture with extracapsular silicone diffusion.18 However, all of the above-mentioned radiodiagnostic methods depend on the radiologist’s experience; thus, the results are highly variable in the literature.13 Currently, magnetic resonance imaging is considered the best method for implant rupture diagnosis; it is considerably useful for extracapsular silicone leakage and even better for intracapsular rupture, with a superior sensitivity and specificity.12,14 However, there is still a large variability in its accuracy and, furthermore, only a few studies have classified the suspicious and certain radiological signs of implant rupture.11
The primary aim of this study was to investigate the accuracy of magnetic resonance imaging and ultrasound in detecting silicone breast implant rupture in postmastectomy asymptomatic patients requiring secondary surgery, comparing magnetic resonance imaging and ultrasound with intraoperative findings. We also compare the two radiological techniques and provide a reliability description of the occurrence of each radiological sign.
PATIENTS AND METHODS
From April of 2005 to June of 2008, we prospectively recruited a consecutive series of 102 postmastectomy patients requiring implant change for aesthetic purposes at the European Institute of Oncology. None of them suffered any specific symptom of implant rupture or had any trauma. The study was approved by an institutional review committee. Only the patients who signed the informed consent document were eligible for this study. This study is a prospective cohort study with level II evidence according to the American Society of Plastic Surgeons Rating Levels of Evidence and Grading Recommendations.
Preoperative magnetic resonance imaging was performed in every single case, but for five implants, an evaluation was not feasible. Ultrasound was performed on a subset of 107 implants. The radiological signs of magnetic resonance imaging were noted by a single experienced radiologist and were recorded. The examined implants were classified as “undamaged” or “ruptured” according to preoperative magnetic resonance imaging and ultrasound radiological findings. These findings were compared with the intraoperative surgical findings. Intraoperatively, implants were defined as “ruptured” when silicone gel was present outside the implant shell, whereas they were defined as “undamaged” when no free silicone was detected outside their shell.
Magnetic Resonance Imaging Technique and Evaluation of Images
Magnetic resonance imaging examinations were performed on a 1.0-T scanner (Signa Horizon; General Electric, Fairfield, Conn.) or on a 1.5-T scanner (Avanto Siemens, Erlangen, Germany), equipped with a dedicated breast coil (seven channels; Invivo, Gainesville, Fla.). Patients were positioned prone, feet first, and were scanned independently of the day of menstrual cycle. Magnetic resonance imaging sequences acquired were as follows: axial spin echo T1-weighted sequences (repetition time/echo time, 563/9.7 msec; field of view, 320 mm; slice thickness, 3.0 mm; voxel size, 1.4 × 1.0 × 5.0 mm; integrated parallel acquisition techniques factor, 2); sagittal short T1 inversion recovery–weighted sequences (repetition time/echo time, 4120/72 msec; field of view, 250 mm; slice thickness, 3 mm; voxel size, 1.2 × 1.0 × 3.0 mm; integrated parallel acquisition techniques factor, 2); sagittal short T1 inversion recovery water saturation sequences (repetition time/echo time, 6740/79 msec; field of view, 250 mm; slice thickness, 3 mm; voxel size, 1.0 × 0.8 × 3.0 mm; integrated parallel acquisition techniques factor, none), and sagittal T1-weighted spin echo sequence (repetition time/echo time, 350/8.5 msec; field of view, 250 mm; slice thickness, 3.0 mm; voxel size, 1.2 × 1.0 × 3.0 mm; integrated parallel acquisition techniques factor, none). Sagittal sequences were acquired on the side of the implant being investigated. No oral or intravenous contrast was given. Images were evaluated by one experienced magnetic resonance imaging reader, aware of the patient’s name and age and of the type of breast implant. The reader completed a standardized form, according to predetermined criteria for rupture. According to Hölmich et al.,19 imaging signs indicating rupture of the breast implant were the linguine sign (Fig. 1), referring to folded wavy lines within silicone gel, often arranged more or less parallel to the fibrous capsule; the teardrop sign (Fig. 2), referring to an invagination of the silicone membrane containing a droplet of silicone; the train rail sign (Fig. 2), referring to the presence of two hypointense parallel lines forming a double contoured subcapsular line within the silicone gel; water droplets (Fig. 3), referring to the presence of small hypodense elements within the silicone, with a variant showing “salt-and-pepper” appearance (Fig. 4), referring to the presence of tiny punctuate hypointense droplets within the silicone gel; inhomogeneity of intensity of the silicone gel within the implant; and double contour of the implant, referring to the presence of a hypointense line running almost parallel to the fibrous capsule.
Magnetic resonance image evaluation indicated rupture (further differentiated into intracapsular or extracapsular rupture, depending on silicone within or outside the fibrous capsule) or intact implant. If the magnetic resonance imaging evaluation was altered by artifacts or other limitations for further evaluation, it was considered not assessable.
Ultrasound Technique and Evaluation of Images
One hundred seven implants were also examined by ultrasound, which was performed using a high-resolution ultrasound scanner equipped with an 8- to 10-MHz transducer (Technos Esaote machine; Esaote, Genoa, Italy). Suspicious areas were also scanned by radial and antiradial orientation with and without compression. Ultrasonography was performed by experienced, highly specialized radiologists exclusively dedicated full time to clinical breast radiology.
In the ultrasound reports relating to images, the following features were considered for extracapsular rupture: “snow storm” (multiple hyperechoic elements in the breast parenchyma caused by microscopic accumulations of silicone outside the fibrous capsule, which determine dispersion of the ultrasound beam), hyperechoic and hypoechoic assets (caused by a leakage of the intraprosthetic content, resulting in inflammation of the periprosthetic tissue), discontinuity of the breast implant capsule, siliconomas, and granulomas containing large silicone parts (which result in the transmission of an ultrasound beam similar to that in fluids, with minimal fibrous reaction and the appearance of complex cysts). Features considered indicative of intracapsular rupture were hyperechoic “serpentine” and silicone nodules in the axillary cords.20
Surgery was performed under general anesthesia and all implants were removed. The type and status of the implant were assessed once the periprosthetic capsule was entered. Ruptured implants were replaced with new implants or autologous flaps.
Sensitivity, specificity, positive predictive value, negative predictive value, and the overall accuracy of magnetic resonance imaging and ultrasound in detecting ruptured implants were calculated, along with their corresponding 95 percent confidence intervals. The accuracy of magnetic resonance imaging and ultrasound were compared by using the chi-square test. The agreement between the two radiological techniques was evaluated by using the K statistic.
The percentage of ruptured implants and median time from implantation for anatomical and rounded implants were compared by using the chi-square test and the two-sample Wilcoxon test, respectively. For each implant, duration was defined as the time (in months) between implantation and surgery.
In our series, 102 patients with 130 implants had undergone breast implant replacement for aesthetic improvement. Implant reconstruction was unilateral in 75 patients and bilateral in 26 patients. One patient had an initial unilateral reconstruction; on a subsequent occasion, she underwent contralateral mastectomy and implant reconstruction and implant replacement. The mean age of the patients was 50 years (range, 25 to 73 years). The median time of implantation was 57 months (range, 6 to 166 months). All were single-lumen silicone gel implants, including 71 round and 59 anatomically shaped implants. At the intraoperative evaluation, 36 implants (27.7 percent) were found to be ruptured and 94 implants (72.3 percent) were undamaged.
The median duration of undamaged implants was 42.5 months (range, 6 to 164 months), whereas the median duration of the damaged implants was 81 months (range, 11 to 166 months; p = 0.0003, two-sample Wilcoxon test). The total number of damaged implants at 4, 5, and 10 years was seven of 130 (5 percent), 12 of 130 (9 percent), and 32 of 130 (25 percent), respectively.
The preoperative magnetic resonance imaging scans of 125 implants for which an evaluation was feasible showed signs of implant rupture in 31 (24.8 percent) and no rupture in 94 (75.2 percent). When we compared magnetic resonance imaging scans and surgery for these 125 implants, we observed six false-negatives (4.8 percent) and two false-positives (1.6 percent) (Table 1). The magnetic resonance imaging sensitivity was 83 percent (95 percent CI, 66 to 93 percent), the specificity was 98 percent (95 percent CI, 92 to 100 percent), the positive predictive value was 94 percent (95 percent CI, 79 to 99 percent), and the negative predictive value was 94 percent (95 percent CI, 87 to 98 percent). The overall accuracy of magnetic resonance imaging in detecting ruptured implants was 94 percent (95 percent CI, 88 to 97 percent).
The preoperative ultrasound images of 107 implants for which an evaluation was made showed signs of implant rupture in 42 implants (39.3 percent) and no rupture in 65 implants (60.7 percent). When we compared ultrasound and surgery for these 107 implants, we observed 10 false-negatives (9.3 percent) and 20 false-positives (18.7 percent) (Table 2). The ultrasound sensitivity was 69 percent (95 percent CI, 50 to 84 percent), the specificity was 73 percent (95 percent CI, 62 to 83 percent), the positive predictive value was 52 percent (95 percent CI, 36 to 68 percent), and the negative predictive value was 85 percent (96 percent CI, 74 to 92 percent). The overall accuracy of ultrasound in detecting ruptured implants was 72 percent (95 percent CI, 62 to 80 percent).
The diagnostic accuracy of magnetic resonance imaging was significantly higher than that of ultrasound (94 percent versus 72 percent; p < 0.0001). The agreement between the two radiological techniques was moderate (K statistic, 0.48; 95 percent CI, 0.31 to 0.65).
The abnormal magnetic resonance imaging findings are described in Table 3. Of the 31 implants classified as ruptured with preoperative magnetic resonance imaging, seven implants (22.6 percent) presented a single suspicious sign of rupture, 15 implants (48.4 percent) presented two signs, and nine implants (29.0 percent) presented three signs. For the two false-positive cases, the observed signs were silicone heterogeneity in both cases and water droplets in one of the two.
The present work assesses the accuracy of magnetic resonance imaging in detecting silicone breast implant rupture in postmastectomy patients. The median time to explantation was 57 months (range, 6 to 166 months) and the rupture rate was 27.7 percent. Previous studies reported silicone shell ruptures in 15 to 33 percent. This was dependent on implant generation and length of implantation.9,13,21,22
In our series, magnetic resonance imaging sensitivity was 83 percent and specificity was 98 percent. These data are consistent with those in the literature (sensitivity range, 72 to 94 percent; specificity range, 85 to 100 percent).16 Moreover, we observed a 94 percent positive predictive value and a 94 percent negative predictive value.
Our results show that magnetic resonance imaging performs better than ultrasound for diagnosis of breast implant rupture, with an overall accuracy of 94 percent and 72 percent, respectively. As shown by Hold et al.,23 magnetic resonance imaging should be considered the investigation of choice for implant rupture where the resource is available. However, if magnetic resonance imaging is contraindicated or unavailable, ultrasound can be considered as an alternative method. Despite the results of Hold et al., we did not find comparable specificities of the two methods (98 percent versus 73 percent for magnetic resonance imaging and ultrasound, respectively), probably because of the operator dependency of ultrasound. Indeed, in our cohort, magnetic resonance imaging examinations were all read by the same radiologist, whereas the ultrasound procedures were performed by different readers. In this study, the negative predictive value of ultrasound was 85 percent, meaning that in the case of negative ultrasound findings, magnetic resonance imaging may be avoided. However, the concordance of the two investigations may be helpful to ensure the appropriate use of the available resources.
Our study population includes postmastectomy patients who underwent implant-based reconstruction and required secondary surgery for aesthetic improvement (e.g., asymmetry and unsatisfactory outcomes). None of them suffered any specific symptoms of implant rupture. Estimates of the diagnostic accuracy of magnetic resonance imaging may be different for symptomatic patients, who have a higher prevalence of ruptures. Furthermore, other variables may interfere with magnetic resonance imaging reliability, such as time of implantation, time of magnetic resonance imaging evaluation after implant damage, and intracapsular or extracapsular ruptures.24 The variability of magnetic resonance imaging findings also arises from the lack of standardized radiological images. Finally, even the macroscopic definition of implant rupture is not always clear: some surgeons define a rupture as when there is a visible hole or a gross rupture site in the implant shell; others, however, consider the implant ruptured even when silicone “bleeds” through an intact shell.6,11 Moreover, surgical manipulation may result in implant shell breakdown during surgery, resulting in an unpredictable alteration of prevalence and magnetic resonance imaging accuracy.11
Preoperative magnetic resonance imaging showed six radiological signs predicting 31 ruptured implants. For 24 implants (77 percent), more than one radiological sign was present. The teardrop sign is the most frequent sign (34.4 percent), followed by water droplets (29.7 percent), the linguine sign (15.6 percent), the railway sign (12.5 percent), heterogeneous silicone (6.2 percent), and the salt-and-pepper sign (1.6 percent). All of the signs observed on magnetic resonance imaging were considered predictive of implant rupture. There were two false-positive signs, both including heterogeneous silicone imaging, and water droplets in just one of the two cases. According to Hölmich,25 these suspicious signs are not certain signs of rupture, and are mainly considered as an isolated finding, which possibly results in an overinterpretation and may reduce the specificity.25 This different classification of magnetic resonance imaging radiological images and lacking of standardization may affect magnetic resonance imaging accuracy. Strict diagnostic criteria to define ruptured, suspicious. and undamaged implants would represent a promising development. In the series of Hölmich et al.,26 by comparison with our study, the linguine sign and subcapsular line sign were observed more frequently and were considered as certain signs of rupture. Conversely, in our series, the teardrop sign and water droplets are the most common findings. As these authors intended to be more conservative than us by searching for a lower false-positive rate (high specificity) rather than few false-negative rates (high sensitivity), they considered the teardrop sign and water droplets as suspicious, but not certain, signs of rupture; instead, we considered all of the above-mentioned abnormal radiological findings as certain rupture signs. However, in their statistical analysis, they classified all the signs as a diagnosis of rupture. Following this strategy for calculation, they found results comparable to ours, achieving an accuracy of 92 percent, a sensitivity of 89 percent, and a specificity of 97 percent and corresponding to a positive predictive value of 99 percent and a negative predictive value of 79 percent.26
In our opinion, magnetic resonance imaging remains a second-stage confirmatory diagnostic evaluation and should not be used to screen asymptomatic women.11 It is unclear whether the potential benefits of magnetic resonance imaging screening overcome patient risks and costs or lead to a reduction in morbidity.27 Despite the consensus on explantation of symptomatic and extracapsular implant rupture, no evidence suggests the same clinical approach for intracapsular or asymptomatic ruptures.27 In fact, there is a lack of evidence that silent ruptures will progress to clear or symptomatic implant ruptures.2,25 The decision between “watchful waiting” and explantation is based on the balance between potential symptoms and risks correlated with silicone migration and potential risks correlated with secondary surgery.27 We recommend a standardization and classification of abnormal magnetic resonance imaging signs predictive for implant rupture to improve magnetic resonance imaging accuracy. In fact, magnetic resonance imaging guidelines could decrease radiologist bias and could be useful both in further resources and as institutional protocols.
Based on the indications of the U.S. Food and Drug Administration, which recommend screening with magnetic resonance imaging 3 years after breast augmentation and every 2 years thereafter,28 Chung et al.29 performed an economic analysis to determine optimal screening strategies by considering the diagnostic accuracy of the screening tests (ultrasound versus magnetic resonance imaging), costs of the tests, and subsequent implant removal, for asymptomatic patients who had undergone breast augmentation or postmastectomy reconstruction with a silicone gel breast implant. They showed that screening women with ultrasound followed by magnetic resonance imaging just in case of positive ultrasound findings is the least expensive strategy with which to detect silent rupture. Considering the high false-positive rate of ultrasound, the average cost of a single positive ultrasound followed by surgery is more expensive when compared with magnetic resonance imaging screening after a positive ultrasound ($1089 versus $637 per rupture detected and managed). Even if we did not perform a direct analysis of costs, the study of Chung et al. reinforces our results on postmastectomy reconstruction patients, confirming that the presence of a positive ultrasound examination requires magnetic resonance imaging before surgery to avoid an unnecessary operation and to contain costs.
Our series demonstrated an acceptable percentage of implant rupture after 5 years in asymptomatic patients (9 percent). Considering the high negative predictive value of ultrasound, we would recommend yearly screening of asymptomatic women with ultrasound, avoiding the use of more expensive examinations such as magnetic resonance imaging. Because magnetic resonance imaging has a higher diagnostic accuracy than ultrasound, we suggest magnetic resonance imaging assessment after 5 years, when implant rupture becomes more frequent.
Magnetic resonance imaging should be considered the investigation of choice for silicone gel implant rupture in postmastectomy patients, and standardization of magnetic resonance imaging criteria may improve magnetic resonance imaging accuracy. Considering the high negative predictive value of ultrasound, magnetic resonance imaging can be avoided in the case of a negative ultrasound exmaination. We therefore suggest a strategy of screening asymptomatic women with ultrasound every year and with magnetic resonance imaging every 5 years.
1. Brown SL. Epidemiology of silicone-gel breast implants. Epidemiology. 2002;13(Suppl 3):S34–S39
2. Hölmich LR, Vejborg IM, Conrad C, et al. Untreated silicone breast implant rupture. Plast Reconstr Surg. 2004;114:204–214; discussion 215
3. Janowsky EC, Kupper LL, Hulka BS. Meta-analyses of the relation between silicone breast implants and the risk of connective-tissue diseases. N Engl J Med. 2000;342:781–790
4. Jewell M, Spear SL, Largent J, Oefelein MG, Adams WP Jr. Anaplastic large T-cell lymphoma and breast implants: A review of the literature. Plast Reconstr Surg. 2011;128:651–661
5. Lipworth L, Holmich LR, McLaughlin JK. Silicone breast implants and connective tissue disease: No association. Semin Immunopathol. 2011;33:287–294
6. Brown SL, Silverman BG, Berg WA. Rupture of silicone-gel breast implants: Causes, sequelae, and diagnosis. Lancet. 1997;350:1531–1537
7. Hölmich LR, Kjøller K, Fryzek JP, et al. Self-reported diseases and symptoms by rupture status among unselected Danish women with cosmetic silicone breast implants. Plast Reconstr Surg. 2003;111:723–732; discussion 733
8. Vermeulen RC, Scholte HR. Rupture of silicone gel breast implants and symptoms of pain and fatigue. J Rheumatol. 2003;30:2263–2267
9. Hölmich LR, Friis S, Fryzek JP, et al. Incidence of silicone breast implant rupture. Arch Surg. 2003;138:801–806
10. Robinson OG Jr, Bradley EL, Wilson DS. Analysis of explanted silicone implants: A report of 300 patients. Ann Plast Surg. 1995;34:1–6; discussion 6
11. Cher DJ, Conwell JA, Mandel JS. MRI for detecting silicone breast implant rupture: Meta-analysis and implications. Ann Plast Surg. 2001;47:367–380
12. Kreymerman P, Patrick RJ, Rim A, Djohan R, Crowe JP. Guidelines for using breast magnetic resonance imaging to evaluate implant integrity. Ann Plast Surg. 2009;62:355–357
13. Goodman CM, Cohen V, Thornby J, Netscher D. The life span of silicone gel breast implants and a comparison of mammography, ultrasonography, and magnetic resonance imaging in detecting implant rupture: A meta-analysis. Ann Plast Surg. 1998;41:577–585; discussion 585
14. Di Benedetto G, Cecchini S, Grassetti L, et al. Comparative study of breast implant rupture using mammography, sonography, and magnetic resonance imaging: Correlation with surgical findings. Breast J. 2008;14:532–537
15. Gorczyca DP, Gorczyca SM, Gorczyca KL. The diagnosis of silicone breast implant rupture. Plast Reconstr Surg. 2007;120(Suppl 1):49S–61S
16. Ikeda DM, Borofsky HB, Herfkens RJ, Sawyer-Glover AM, Birdwell RL, Glover GH. Silicone breast implant rupture: Pitfalls of magnetic resonance imaging and relative efficacies of magnetic resonance, mammography, and ultrasound. Plast Reconstr Surg. 1999;104:2054–2062
17. Reynolds HE, Buckwalter KA, Jackson VP, Siwy BK, Alexander SG. Comparison of mammography, sonography, and magnetic resonance imaging in the detection of silicone-gel breast implant rupture. Ann Plast Surg. 1994;33:247–255; discussion 256
18. Everson LI, Parantainen H, Detlie T, et al. Diagnosis of breast implant rupture: Imaging findings and relative efficacies of imaging techniques. AJR Am J Roentgenol. 1994;163:57–60
19. Hölmich LR, Vejborg I, Conrad C, Sletting S, McLaughlin JK. The diagnosis of breast implant rupture: MRI findings compared with findings at explantation. Eur J Radiol. 2005;53:213–225
20. Colombo G, Ruvolo V, Stifanese R, Perillo M, Garlaschi A. Prosthetic breast implant rupture: Imaging–pictorial essay. Aesthetic Plast Surg. 2011;35:891–900
21. Brown SL, Middleton MS, Berg WA, Soo MS, Pennello G. Prevalence of rupture of silicone gel breast implants revealed on MR imaging in a population of women in Birmingham, Alabama. AJR Am J Roentgenol. 2000;175:1057–1064
22. Hölmich LR, Kjøller K, Vejborg I, et al. Prevalence of silicone breast implant rupture among Danish women. Plast Reconstr Surg. 2001;108:848–858; discussion 859
23. Hold PM, Alam S, Pilbrow WJ, et al. How should we investigate breast implant rupture? Breast J. 2012;18:253–256
24. Song JW, Kim HM, Bellfi LT, Chung KC. The effect of study design biases on the diagnostic accuracy of magnetic resonance imaging for detecting silicone breast implant ruptures: A meta-analysis. Plast Reconstr Surg. 2011;127:1029–1044
25. Paetau AA, McLaughlin SA, McNeil RB, et al. Capsular contracture and possible implant rupture: Is magnetic resonance imaging useful? Plast Reconstr Surg. 2010;125:830–835
26. Hölmich LR, Vejborg I, Conrad C, Sletting S, McLaughlin JK. The diagnosis of breast implant rupture: MRI findings compared with findings at explantation. Eur J Radiol. 2005;53:213–225
27. McCarthy CM, Pusic AL, Kerrigan CL. Silicone breast implants and magnetic resonance imaging screening for rupture: Do U.S. Food and Drug Administration recommendations reflect an evidence-based practice approach to patient care? Plast Reconstr Surg. 2008;121:1127–1134
28. U.S. Food and Drug Administration. . 2006 guidance for industry and FDA in saline, silicone gel, and alternative breast implants; 8.5 safety assessment: Rupture. http://www.fda.gov/
29. Chung KC, Malay S, Shauver MJ, Kim HM. Economic analysis of screening strategies for rupture of silicone gel breast implants. Plast Reconstr Surg. 2012;130:225–237