A, Overview of tumor in patient 9 (original magnification ×20) and B, cellular detail at a higher magnification (original magnification ×400). C, The tumor cells show uniform expression of CD30 (original magnification ×400) and D, CD5 expression in some tumor cells (original magnification ×400). The presence of CD30 on the tumor cells and CD5 protein on tumor cells and reactive T-cells is visualized by the brown chromogen. Immunohistochemistry was performed with BerH2 monoclonal antibody for CD30 (DAKO, Glostrup, Denmark) and CD5-4C7 antibody for CD5 (Novocastra, Newcastle Upon Tyne, England), biotin peroxidase detection system (Powervision, Immunologic, Duiven, the Netherlands), diaminobenxidine as chromogen, and hematoxylin counterstain.
de Jong D, Vasmel WLE, Paul de Boer J, Verhave G, Barbé E, Casparie MK, van Leeuwen FE. Anaplastic Large-Cell Lymphoma in Women With Breast Implants. JAMA. 2008;300(17):2030–2035. doi:10.1001/jama.2008.585
Author Affiliations: Departments of Pathology (Dr de Jong), Hematology (Dr de Boer), and Epidemiology (Dr van Leeuwen), the Netherlands Cancer Institute, Amsterdam, the Netherlands; Department of Internal Medicine (Drs Vasmel and Verhave), Department of Pathology (Dr Barbé), St Lucas-Andreas Hospital, Amsterdam, the Netherlands; and The Nationwide Network and Registry of Histopathology and Cytopathology in the Netherlands (PALGA, Pathologisch Anatomisch Landelijk Geautomatiseerd Archief), Utrecht, the Netherlands (Dr Casparie).
Context Recently, we identified 2 patients with anaplastic large T-cell lymphoma (ALCL) negative for tyrosine kinase anaplastic lymphoma kinase (ALK-negative) in the fibrous capsule of silicone breast prostheses, placed for cosmetic reasons. Similar cases have been reported in the literature. Although an increased risk of ALCL in patients with breast prostheses has been speculated, no studies have been conducted so far.
Objective To determine whether ALCL risk is associated with breast prostheses.
Design A search for all patients with lymphoma in the breast diagnosed in the Netherlands between 1990 and 2006 was performed through the population-based nationwide pathology database. Subsequently, we performed an individually matched case-control study. Conditional logistic regression analysis was performed to estimate the relative risk of ALCL associated with breast prostheses.
Setting and Patients Eleven patients with breast ALCL were identified in the registry. For each case patient with ALCL in the breast, we selected 1 to 5 controls with other lymphomas in the breast, matched on age and year of diagnosis. For all cases and controls (n = 35), pathological and clinical information was obtained with special emphasis on the presence of a breast prosthesis.
Main Outcome Measure Association between breast implants and ALCL of the breast.
Results The 11 patients with ALCL of the breast were diagnosed between 1994 and 2006 at a median age of 40 years (range, 24-68 years). In 5 of these patients, bilateral silicone breast prostheses had been placed 1 to 23 years before diagnosis. All received prostheses for cosmetic reasons. Lymphoma classes of 35 eligible control patients were 12 diffuse large B-cell lymphomas, including 1 T-cell rich B-cell lymphoma; 5 Burkitt lymphomas; 10 mucosa-associated lymphoid tissue–type lymphoma; 3 follicular lymphomas; 3 peripheral T-cell lymphomas; and 2 indolent B-cell lymphomas, unclassified. One of 35 control patients had a breast implant placed before diagnosis of lymphoma. The odds ratio for ALCL associated with breast prostheses was 18.2 (95% confidence interval, 2.1-156.8).
Conclusions These preliminary findings suggest an association between silicone breast prostheses and ALCL, although the absolute risk is exceedingly low due to the rare occurrence of ALCL of the breast (11 cases in the Netherlands in 17 years). These findings require confirmation in other studies.
Since the late 1970s, silicone breast implants have been under constant challenge for suspected association with systemic disease and malignancy.1- 4 Although no health risk had been proven, the use of silicone-filled breast implants was banned by the US Food and Drug Administration in 1992. Saline-filled, silicone-covered implants stayed on the market. Also with these implants, contracture and rupture are frequent events, and interference with breast cancer detection may be a problem. Large observational epidemiological studies in populations in Canada and Sweden have not shown consistent associations with breast cancer or other specific cancer sites or with autoimmune disease.5,6
Several cases of non-Hodgkin lymphoma in women with breast implants have been described. Of these, the majority were anaplastic large-cell lymphoma (ALCL), negative for anaplastic lymphoma kinase (ALK-negative). Only a single case of follicular lymphoma, 1 of lymphoplasmacytic lymphoma, and 1 of primary effusion lymphoma, human herpesvirus 8–associated, have been reported.7- 16 In population-based studies, ALCL at all sites represents only 0.5% to 3% of non-Hodgkin lymphoma in adults (eg, 0.6% in the population-based database of the Comprehensive Cancer Center West, 3% in the World Health Organization blue book).17 Moreover, primary lymphomas of the breast are reported to be predominantly of B-cell type and T-cell lymphomas in general are exceedingly rare. Therefore, an increased risk of ALCL in patients with breast prostheses has been speculated. However, to our knowledge, no epidemiological studies examining this association have been conducted so far.
We recently observed 2 patients diagnosed with ALCL located in the fibrous capsule of a saline-filled silicone breast implant at primary diagnosis. Reports on similar cases in the literature initiated a comprehensive search to identify all patients with biopsy-proven primary non-Hodgkin lymphoma of the breast diagnosed between 1990 and 2006 in the Netherlands.7,10- 16 Since 1971, all reports on cytological and histological diagnoses generated by all pathology departments (academic and nonacademic) in the Netherlands are centrally archived with complete national coverage since 1989.18 Standardized coding allows for anonymized comprehensive searches for specific diagnoses and patient cohorts.
From this population-based database (PALGA, Pathologisch Anatomisch Landelijk Geautomatiseerd Archief), 429 cases were retrieved with a histologically proven diagnosis of lymphoma in the breast between 1990 and 2006 without a previously listed diagnosis of lymphoma at another site (389 women, 40 men). Of the 389 female patients, 11 patients had a diagnosis of ALCL, including the 2 initially observed patients. All histological material was retrieved for confirmation of the diagnosis, additional immunohistochemical analysis, and molecular studies. Information on the complete medical history, including staging results and mammography results, was collected. Immunohistochemistry and immunoglobulin and T-cell receptor rearrangement analyses were performed according to standard methods.19,20
Subsequently, we performed an individually matched case-control study, nested in the same cohort of 389 female patients. For each case patient with ALCL in the breast, we attempted to select 3 to 7 controls with other lymphomas in the breast, matched on age at diagnosis (±5 years) and year of diagnosis (±2 years). For all 47 potential controls, we obtained pathology reports. Furthermore, for all cases and controls, we sent a standardized questionnaire to the treating physician to obtain information on medical history, including previous malignancies, staging results, and presence of a breast prosthesis, including mammography results.
Conditional logistic regression analysis was performed to estimate the odds ratio (OR) of ALCL associated with breast prosthesis, using EGRET for Windows, 1999 (CYTEL Inc, Cambridge, Massachusetts).21 The OR was used as a valid risk estimate of relative risk and is therefore referred to as such. An estimate for absolute risk was made based on breast prosthesis sales figures for 1999 to extrapolate the number of women with breast prostheses.
This study was performed according to the current Dutch laws and regulations for medical record–based research with coded data and approved by the scientific advisory council of PALGA.
Our search in the national database for all patients with a histologically proven diagnosis of lymphoma in a breast specimen retrieved 389 female patients with probable primary disease diagnosed between 1990 and 2006. All 11 patients with ALCL in the breast included in our study were histologically confirmed based on review of the slides and with further immunohistochemistry and T-cell receptor analysis. In 1 additional patient, a suggested diagnosis of ALCL could not be confirmed and that case was excluded from this study.
Clinical and histological features are listed in Table 1. Patients were diagnosed between 1994 and 2006 and were a median age of 40 years (range, 24-68 years). Eight patients had unilateral breast involvement and 3 had bilateral involvement. Eight patients had limited-stage disease with breast involvement with or without axillary lymph node localizations (stage I and II); 3 patients had more disseminated disease with the dominant lymphoma localizations in the breast. In all cases, ALCL was diagnosed on the basis of an infiltrate of large, polymorphic lymphoid tumor cells with a varying background of small lymphocytes, macrophages, and eosinophils (Figure). There was uniform expression of CD30 on tumor cells, and T-cell immunophenotype could be confirmed in all cases on the basis of expression of CD3, CD2, and/or granzyme B expression. There was no expression of ALK-1. In 5 patients, bilateral silicone-covered, saline-filled breast prostheses were placed 1, 3, 4, 13, and 23 years before diagnosis, all for cosmetic reasons. Lymphoma was seen in the fibrous capsules of these prostheses that were all in situ at the time of diagnosis.
Forty-seven potential control women with breast lymphomas other than ALCL were retrieved from the cohort of 389 patients as described above. After receiving all requested clinical and pathological information, 12 of these were excluded: for 6, the lymphoma localization in the breast was found to be recurrent disease after a previous episode of nodal or extranodal lymphoma; for 4, the diagnosis of lymphoma could not be confirmed; and for 2, double entry in the database occurred.
Patient characteristics of the remaining 35 patients are listed in Table 2 and represent 1 to 5 controls for each ALCL patient. Ten patients were diagnosed with marginal-zone lymphoma, mucosa-associated lymphoid tissue–type, 12 with diffuse large B-cell lymphoma, including 1 T-cell rich B-cell lymphoma; 5 with Burkitt lymphoma; 3 with indolent follicular lymphoma; 3 with peripheral T-cell lymphoma; unclassified (CD30 negative); and 2 with indolent B-cell lymphoma, not further classified. Three patients had developed primary breast lymphoma during pregnancy (2 Burkitt lymphoma and 1 diffuse large B-cell lymphoma). Among control patients, 1 patient was identified with bilateral silicone breast implants, placed 14 years before the diagnosis of indolent follicular lymphoma with a dominant localization in the right breast and further nodal localization above and below the diaphragm as well as microscopic bone marrow involvement (Table 2). The odds ratio for ALCL in the breast associated with silicone breast prosthesis placed for cosmetic reasons was 18.2 (95% confidence interval, 2.1-156.8).
Several small series of patients and case reports have been published on lymphoma in the fibrous capsules of breast implants. Of the 16 reported cases that we are aware of, the majority are of ALK-1–negative ALCL nodal-type (n = 10)10- 15 with the exception of 1 case of indolent follicular lymphoma,8 1 case of human herpesvirus 8–associated primary effusion lymphoma,9 1 case of lymphoplasmacytic lymphoma,16 and 3 cases of primary cutaneous ALCL.7 Because nodal-type ALCL is a very rare disease with a frequency in population-based studies of approximately 3%, the unusual distribution of this lymphoma type in case reports supports an association between ALCL and breast implants. A recent epidemiological study by McLaughlin et al6 did not support an increased risk for lymphoma. However, this study did not have sufficient power to address subtypes of non-Hodgkin lymphoma, with only 3 unspecified cases of non-Hodgkin lymphoma included.
Strengths of our study include the complete identification of women with ALCL in the breast in the period 1990-2006, the selection of control subjects from the same cohort, and the complete retrieval of information on breast prosthesis in all women, rendering selection bias implausible. We selected as controls women who had another type of lymphoma in the breast because we assumed that the treating physicians of cases and controls with breast lymphoma would be equally likely to be aware of a patient's breast prosthesis; in addition, mammographic information could be collected on all patients. A limitation of our design is that, if silicone breast prostheses are also associated with breast lymphomas other than ALCL, we have underestimated the strength of association between breast prostheses and ALCL in the breast. Moreover, our study concerns a small number of ALCL cases included, despite nationwide coverage, due to the rarity of the disease.
To explain the biological mechanism of the association of ALCL and breast implants, 3 hypotheses may be proposed: direct immunological drive, indirect cytokine-mediated drive, and toxic damage by silicone products. Although a specific immunological drive by silicone components would be suggestive as an explanation, this is not very likely. Importantly, it has been shown that ALCL lacks a functional T-cell receptor signaling system either by a defect of the T-cell receptor molecules or of downstream signaling components, precluding direct antigenic stimulation.22 The typical organ-related autoimmunity-driven lymphoma of marginal zone B-cell lymphoma of mucosa-associated lymphoid tissue–type has been described in the breast but not in relation to silicone breast implants, further excluding this model for ALCL.23 An indirect association driven by a specific cytokine response may be assumed and indeed, extensive inflammatory and histiocytic reactions to silicone are seen both in fibrous capsules as well as in draininglymph nodes.24 Alternatively, toxic substances from the prostheses could be directly oncogenic to a precursor population of ALCL but apparently not to other organ-specific lymphoid cell populations, including B-cells. Indeed, outside the context of breast implants, primary lymphoma of the breast is mostly of B-cell type with 49% diffuse large B-cell lymphoma and 17% marginal zone lymphoma, mucosa-associated lymphoid tissue–type in our series of 389 women diagnosed between 1990 and 2006. Similar frequencies are reported by others, suggesting for B-cell lymphomas, other risk factors—including autoimmune-related factors—apply more than may be presumed for ALCL.
Although an 18-fold increased odds for the development of a specific lymphoma in the breast may cause significant concern among women with breast prostheses, it should be realized that the absolute risk remains very low due to the exceedingly rare occurrence of ALCL in the population (estimated incidence at all sites 0.1/100 000 per year). Indeed, only 11 cases of breast ALCL occurred in the Netherlands (population of 8 million women) in 17 years. Based on rather uncertain data (sales figures for 1999), we estimate that there may be 100 000 to 300 000 women with cosmetic breast prostheses in the Netherlands.25 This implies that the incidence of ALCL in the breast would vary between 0.1 to 0.3 per 100 000 women with prostheses per year (5 cases in 1.7-5.1 million person-years). Therefore, the absolute risk of developing breast cancer in a prosthesis-containing breast is much higher than the risk of ALCL, despite the absence of a prosthesis-related increased risk of breast cancer (http://www.ikcnet.nl).
In conclusion, these preliminary findings suggest an association between silicone breast prosthesis and ALCL, although the absolute risk is exceedingly low due to the rare occurrence of ALCL. These findings require confirmation by other studies.
Corresponding Author: Daphne de Jong, MD, PhD, Department of Pathology, The Netherlands Cancer Institute, Plesmanlaan 121, Amsterdam 1066CX, the Netherlands (firstname.lastname@example.org).
Author Contributions: Drs de Jong and van Leeuwen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: de Jong, Vasmel, de Boer, van Leeuwen.
Acquisition of data: de Jong, Vasmel, de Boer, Verhave, Barbé, Casparie, van Leeuwen.
Analysis and interpretation of data: de Jong, Vasmel, de Boer, van Leeuwen.
Drafting of the manuscript: de Jong, Vasmel, de Boer, van Leeuwen.
Critical revision of the manuscript for important intellectual content: de Boer, Verhave, Barbé, Casparie.
Statistical analysis: van Leeuwen.
Administrative, technical, or material support: de Boer, Verhave, Barbé, Casparie.
Study supervision: de Jong.
Financial Disclosures: None reported.
Funding/Support: This study was supported by the Netherlands Cancer Institute without external funding.
Role of the Sponsor: The Netherlands Cancer Institute did not participate in the design and conduct of the study, in the collection of data, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript.
Additional Contributions: We thank Carla Wauters, MD, Canisius-Wilhelmina Hospital, Nijmegen, the Netherlands, for her keen attention to this unique feature in patients with breast implants and providing patient material for this study and Willem Kolkman, MSc, Department of Epidemiology, the Netherlands Cancer Institute, Amsterdam, for statistical support. We thank the many clinicians and pathologists who provided clinical information and pathological data on all patients. None of these acknowledged contributors have received any financial compensation.