SEARCH

SEARCH BY CITATION

Keywords:

  • prophylactic mastectomy;
  • breast cancer;
  • hormone receptor status;
  • epidemiology;
  • propensity score analysis

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. FUNDING SOURCES
  7. REFERENCES

BACKGROUND:

The effect of contralateral prophylactic mastectomy (CPM) on the survival of patients with early-stage breast cancer remains controversial. The objective of this study was to evaluate the benefits of CPM using a propensity scoring approach that reduces selection bias from the nonrandom assignment of patients in observational studies.

METHODS:

A total of 3889 female patients with stage I to III breast cancer were identified who were treated at The University of Texas MD Anderson Cancer Center from 1997 to 2009. We assessed the association between CPM and disease-free (DFS) and overall survival (OS), by using Cox proportional hazards models to estimate hazard ratios (HRs), and by matching patients in the CPM and no-CPM groups using propensity scores (n = 497 pairs).

RESULTS:

With a median follow-up time of 4.5 years, CPM was associated with improved DFS (HR, 0.75; 95% confidence interval [CI], 0.59-0.97) and OS (HR, 0.74; 95% CI, 0.56-0.99), adjusted for prognostic factors. The improved DFS was seen predominantly among hormone receptor–negative (HR, 0.60; 95% CI, 0.38-0.95) compared with hormone receptor–positive patients (HR, 0.80; 95% CI, 0.58-1.10). For the matched patient cohort, stratified survival analysis also showed an improvement in DFS with CPM (HR, 0.48; 95% CI, 0.22-1.01) in hormone receptor–negative patients that was nearly statistically significant.

CONCLUSIONS:

CPM was associated with improved DFS for some patients with hormone receptor–negative breast cancer, after reducing selection bias. Identifying subsets of patients most likely to benefit from CPM may have important implications for a more personalized approach to treatment decisions about CPM. Cancer 2012. © 2012 American Cancer Society.

Contralateral prophylactic mastectomy (CPM) among US patients with unilateral invasive breast cancer increased by 150% from 1993 to 2003, with no evidence of a geographic difference in practice or plateau effect.1 Although CPM has been shown to reduce the risk of developing contralateral breast cancer, it is unclear whether CPM reduces breast cancer–related mortality or improves overall survival (OS) rates.2 The increased use of CPM is especially concerning among women with early-stage sporadic breast cancer, who have a cumulative risk of contralateral breast cancer that ranges from 7.6% to 13.0% for women younger than 50 years at initial diagnosis and 3.5% to 5.3% for women older than 50 years at initial diagnosis.3, 4 For many of these women, the risk of distant metastatic disease from the index tumor outweighs the risk of contralateral breast cancer.5-8

It is also possible that any marginal benefit of CPM on breast cancer survival will be observed only among certain subgroups of patients. For example, using the National Cancer Institute's SEER (Surveillance, Epidemiology, and End Results) database, Bedrosian et al showed that patients younger than 50 years with stage I or II estrogen receptor (ER)-negative breast cancers who underwent CPM had a 4.3% improvement in breast cancer survival compared with those who underwent CPM for ER-positive breast cancer, although information on whether patients received adjuvant endocrine therapy was not included.9 Additional limitations of prior epidemiologic studies evaluating the clinical benefit of CPM have included small sample sizes, short-term follow-up, and incomplete information on tumor characteristics, systemic treatment, and comorbidities.10

Selection bias also plays a major role in studies examining the benefits of CPM because patients who receive CPM are most likely to be white, younger than 50 years, and may have other characteristics that predispose them to better outcomes.11, 12 It is unlikely that a randomized study will ever be conducted to evaluate the clinical benefits of CPM; therefore, statistical methods that balance demographic and clinical characteristics that influence the decision to undergo CPM are needed to reduce selection bias. Our objective in this study was to examine the associations between CPM and DFS and OS in a large cohort of patients with early-stage breast cancer, applying the propensity score method to reduce the impact of selection bias.13

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. FUNDING SOURCES
  7. REFERENCES

Patient Selection

The prospective Breast Cancer Management System (BCMS) database of The University of Texas MD Anderson Cancer Center (MDACC), Houston, Texas, was searched to identify women with clinical stage I to III primary unilateral invasive breast cancer who underwent a mastectomy at MDACC between June 1997 and August 2009. We excluded patients who underwent bilateral mastectomies for metachronous bilateral breast cancer (contralateral breast cancer within 6 months of diagnosis of the primary breast cancer) and patients with contralateral invasive or ductal carcinoma in situ incidentally discovered at the time of CPM.

The BCMS database, which has been previously described,14 contained detailed information on patient (ethnicity, age, menopausal status, comorbidity score) and tumor (clinical and pathologic stage, ER and progesterone receptor [PR] status, nuclear grade, HER2/neu [human epidermal growth factor receptor 2] mutation status) characteristics, chemotherapy and endocrine treatments, and surgery type (segmental mastectomy, mastectomy, CPM). A patient was considered to have received systemic therapy if there was documentation of receipt of treatment in the medical record. HER2/neu status was evaluated by immunohistochemical analysis or by fluorescence in situ hybridization of breast cancer tissue. Tumors considered HER2/neu-positive were those showing a score of 3+ for protein overexpression on immunohistochemical staining and/or gene amplification on fluorescence in situ hybridization testing. Follow-up information for patients in the BCMS database was obtained every 2 years by direct review of the medical records and linkage to the MDACC Tumor Registry, which mailed annual follow-up letters to each patient registered at MDACC and known to be alive to determine their clinical status. The MDACC Tumor Registry checked the social security death index and the Texas Bureau of Vital Statistics for the status of patients who did not respond to the letters.

This study was approved by the MDACC Institutional Review Board.

Statistical Analysis

The primary endpoint of interest was DFS, defined as the time from breast cancer diagnosis to contralateral breast cancer, locoregional or distant recurrence, or breast cancer–related death, whichever event occurred first. OS was defined as the time from breast cancer diagnosis to death from any cause. The exposure variable of interest was CPM status (yes or no). Other prognostic variables of interest included year of diagnosis, age at diagnosis, ethnicity, stage, comorbidity score, nuclear grade, hormone receptor status, HER2/neu status, chemotherapy history, and endocrine therapy history.

Propensity score analysis was undertaken in an attempt to adjust for potential bias associated with factors related to the decision to undergo CPM. This statistical methodology has often been used in observational epidemiologic studies to control for nonrandom treatment assignment of patients by adjusting for differences in covariates between the treatment groups.13 To control for factors that may confound the relationship between CPM and DFS or OS, we determined the propensity to receive CPM for each patient, using multivariable logistic regression that included the following covariates: year of diagnosis, age at diagnosis, ethnicity, comorbidity score, stage at diagnosis, HER2/neu status, hormone receptor status, nuclear grade, and chemotherapy history. Patients with ER-positive and/or PR-positive tumors were classified as hormone receptor–positive, and those with ER-negative and PR-negative tumors were classified as hormone receptor–negative. Given the propensity scores for all patients, we identified pairs of patients, one CPM patient randomly matched with one who did not undergo CPM, using a 5-to-1 digit greedy match algorithm.14 The differences in propensity scores in each pair were no more than 0.01. We used absolute standardized differences to assess balance in the baseline variables between patients who underwent CPM and those who did not undergo CPM in the matched cohort. The absolute standardized differences for all baseline covariates were <10% in the matched cohort.

For the prematching cohort, differences in demographic, clinical, and tumor characteristics between groups of patients who underwent CPM and those who did not undergo CPM were compared, by using the chi-square test. Multivariable Cox proportional hazards models were used to determine the effect of CPM on DFS and OS after adjusting for the prognostic variables. Because hormone receptor status was highly associated with receipt of endocrine therapy (P < .0001), we adjusted only for hormone receptor status rather than for both hormone receptor status and endocrine therapy history. In the process of fitting the multivariable Cox regression model, a backward variable selection was used with P ≤ .05 as the limit for inclusion. We also explored whether the association between CPM and DFS or OS differed by hormone receptor status.

For the matched cohort (n = 994), differences in demographic, clinical, and tumor characteristics between matched pairs were evaluated using McNemar's test for categorical covariates with 2 levels and generalized estimating equation (GEE) methods for categorical covariates with 3 levels.15, 16 Cox proportional hazards models stratified on the matched pairs were fitted to determine the effect of CPM on DFS and OS for the matched cohort. Kaplan-Meier curves by CPM status were estimated among patients with hormone receptor-negative tumors. We created Kaplan-Meier survival curves and log-minus-log survival plots to test for potential violation of the underlying proportional hazards assumption on which the Cox regression model is based. All tests were 2-sided. P values <0.05 were considered statistically significant. All analyses were conducted using SAS (version 9.1; SAS Institute, Cary, NC) and S-plus (version 8.0; TIBCO Software Inc., Palo Alto, Calif) statistical software.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. FUNDING SOURCES
  7. REFERENCES

Patient Characteristics

We used the records of 3889 patients in our analysis, 532 in the CPM group, and 3357 in the no-CPM group. The median follow-up time for the entire study cohort was 4.5 years, with a total of 708 breast cancer–related events (contralateral cancers, recurrences, or breast cancer–related deaths) and 174 deaths not related to breast cancer. The median follow-up time was 4.4 years among patients who underwent CPM and 4.6 years among patients who did not undergo CPM. Sixty-seven contralateral breast cancers occurred in the no-CPM group and 1 contralateral breast cancer occurred in the CPM group. Table 1 lists the characteristics of patients in the CPM and no-CPM groups before and after propensity score matching. Among the 71 patients whose hormone receptor status was not available (1.8%), 40 patients received endocrine therapy and 31 patients did not receive endocrine therapy.

Table 1. Patient and Tumor Characteristics by Contralateral Prophylactic Mastectomy Status Before and After Matching
FactorsAll Patients Matched Patients 
 No CPM no. (%)CPM no. (%)PNo CPM no. (%)CPM no. (%)P
  1. CPM indicates contralateral prophylactic mastectomy; HER2/neu, human epidermal growth factor receptor 2.

Total patients3357532 497497 
Age at diagnosis, y  <.01  1.00
 <501417 (42.2)308 (57.9) 282 (56.7)282 (56.7) 
 50-59998 (29.7)142 (26.7) 139 (28.0)139 (28.0) 
 ≥60942 (28.1)82 (15.4) 76 (15.3)76 (15.3) 
Ethnicity  <.01  1.00
 Black442 (13.2)30 (5.6) 27 (5.4)27 (5.4) 
 White2172 (64.7)444 (83.5) 415 (83.5)415 (83.5) 
 Other743 (22.1)58 (10.9) 55 (11.1)55 (11.1) 
Comorbidity score  .13  .21
 02772 (82.6)450 (84.6) 426 (85.7)419 (84.3) 
 1326 (9.7)54 (10.2) 36 (7.2)50 (10.1) 
 ≥2259 (7.7)28 (5.3) 35 (7.0)28 (5.6) 
Stage at diagnosis  <.01  1.00
 I648 (19.3)151 (28.4) 138 (27.8)137 (27.6) 
 II1622 (48.3)273 (51.3) 257 (51.7)257 (51.7) 
 III1087 (32.4)108 (20.3) 102 (20.5)103 (20.7) 
Hormone receptor status  .33  .80
 Negative691 (20.6)100 (18.8) 90 (18.1)93 (18.7) 
 Positive2604 (77.6)423 (79.5) 407 (81.9)403 (81.1) 
 Unknown62 (1.8)9 (1.7)  1 (0.2) 
HER2/neu status  <.01  1.00
 Negative2432 (72.4)433 (81.4) 433 (87.1)433 (87.1) 
 Positive605 (18.0)64 (12.0) 64 (12.9)64 (12.9) 
 Unknown320 (9.6)35 (6.6)    
Nuclear grade  .30  .60
 I182 (5.4)28 (5.3) 36 (7.2)28 (5.6) 
 II1345 (40.1)234 (44.0) 215 (43.3)221 (44.5) 
 III1767 (52.6)265 (49.8.) 243 (48.9)245 (49.3) 
 Unknown63 (1.9)5 (0.9) 3 (0.6)3 (0.6) 
Chemotherapy  .95  .61
 No608 (18.1)97 (18.2) 95 (19.1)90 (18.1) 
 Yes2749 (81.9)435 (81.8) 402 (80.9)407 (81.9) 
Endocrine therapy  .04  .65
 No959 (28.6)129 (24.2) 124 (24.9)118 (23.7) 
 Yes2398 (71.4)403 (75.8) 373 (75.1)379 (76.3) 

Prior to matching, patients who underwent CPM differed from those who did not undergo CPM in terms of important prognostic factors. Patients who received CPM were more likely to be younger than 50 years (P < .01), white (P < .01), and have stage I to II disease (P < .01) at diagnosis compared with patients who did not undergo CPM. The CPM patients were also more likely to have been diagnosed between 2006 and 2009 and to have HER2/neu-negative tumors. Pairwise matching using the propensity score substantially reduced the imbalance of these important prognostic factors between the CPM and no-CPM patient groups (see propensity score P values in Table 1).

Association Between CPM and DFS or OS

The adjusted multivariable Cox regression analysis showed a statistically significant 25% improvement in DFS for patients who underwent CPM compared with patients who did not undergo CPM (Table 2). When stratified by the hormone receptor status of the tumor, the benefit of CPM was seen predominantly among the hormone receptor–negative (hazard ratio [HR], 0.60; 95% confidence interval [CI], 0.38-0.95) compared with the hormone receptor–positive patients (HR, 0.84; 95% CI, 0.60-1.15), although the interaction term was not statistically significant (P = .29). In the matched multivariable analysis, the HR indicated improved DFS for the CPM group (HR, 0.77; 95% CI, 0.53-1.13), but the difference was not statistically significant (P = .18). Among patients with hormone receptor–negative tumors, the matched-multivariable analysis showed an improvement in DFS for the CPM group that was nearly statistically significant (P = .05).

Table 2. Hazard Ratios of Disease-Free Survival for Patients Who Underwent CPM Compared With Those Who Did Not Undergo CPM in Cox Proportional Hazard Models Before and After Matching
ModelsNo. of PatientsHR (95% CI)P
  • CI indicates confidence interval; CPM, contralateral prophylactic mastectomy; HR, hazard ratio.

  • a

    Model adjusted for stage, nuclear grade, hormone receptor status, and chemotherapy history.

  • b

    Patients matched by propensity score.

  • c

    Model adjusted for stage, nuclear grade, HER2/neu status, endocrine therapy history, and chemotherapy history.

  • d

    Model adjusted for stage and HER2/neu status.

All patients
 Unadjusted38890.70 (0.55-0.89)<.01
 Adjusteda37620.75 (0.59-0.97).03
 Matchedb497 pairs0.77 (0.53-1.13).18
Hormone receptor–positive
 Unadjusted30270.77 (0.57-1.04).09
 Adjustedc27450.84 (0.60-1.15).28
 Matchedb403 pairs0.73 (0.46-1.17).19
Hormone receptor–negative
 Unadjusted7910.56 (0.36-0.87).01
 Adjustedd7470.60 (0.38-0.95).03
 Matchedb93 pairs0.48 (0.22-1.01).05

In the adjusted multivariable models, patients who underwent CPM had longer OS than did patients who did not undergo CPM (Table 3). This benefit was seen predominantly among patients with hormone receptor–negative tumors (HR, 0.58; 95% CI, 0.36-0.96) compared with those with hormone receptor–positive tumors (HR, 0.81; 95% CI, 0.56-1.17, although the interaction term was not statistically significant (P = .83). In the OS analysis for the matched cohort, CPM was associated with improved OS, but this was not statistically significant (HR, 0.50; 95% CI, 0.21-1.17 among hormone receptor–negative patients; HR, 0.63; 95% CI, 0.36-1.13 among hormone receptor–positive patients; Table 3).

Table 3. Hazard Ratios of Overall Survival for Patients Who Underwent CPM Compared With Those Who Did Not Undergo CPM in Cox Proportional Hazard Models Before and After Matching
ModelsNo. of PatientsHR (95% CI)P
  • CI indicates confidence interval; CPM, contralateral prophylactic mastectomy; HR, hazard ratio.

  • a

    Model adjusted for age, ethnicity, Charlson score, stage, nuclear grade, hormone receptor status, and chemotherapy history.

  • b

    Patients matched by propensity score.

  • c

    Model adjusted for age, ethnicity, comorbidity score, stage, nuclear grade, endocrine therapy history, and chemotherapy history.

  • d

    Model adjusted for ethnicity, stage, and HER2/neu status.

All patients
 Unadjusted38890.64 (0.49-0.85)<.01
 Adjusteda37620.74 (0.56-0.99).04
 Matchedb497 pairs0.70 (0.44-1.09).11
Hormone receptor–positive
 Unadjusted30270.66 (0.46-0.95).02
 Adjustedc29820.81 (0.56-1.17).26
 Matchedb403 pairs0.63 (0.36-1.13).12
Hormone receptor–negative
 Unadjusted7910.56 (0.36-0.89).01
 Adjustedd7330.58 (0.36-0.96).03
 Matchedb93 pairs0.50 (0.21-1.17).11

Benefits of CPM Among Subgroups of Hormone Receptor–Negative Patients

We examined subgroups of patients with hormone receptor–negative disease to determine which subgroups would be most likely to benefit from CPM (Figure 1). Kaplan-Meier curves indicated that patients with HER2/neu-negative (triple-negative) tumors were more likely to benefit from CPM (P < .01) than were patients with HER2/neu-positive tumors (P = .52). In addition, patients who were diagnosed with stage I or II disease (P = .08) were more likely to benefit from CPM than patients with stage III disease (P = .52).

thumbnail image

Figure 1. Disease-free survival (DFS) is shown for patients who underwent contralateral prophylactic mastectomy (CPM) compared with those who did not undergo CPM among hormone receptor–negative patients by stage and HER2/neu status. Abbreviations: CI, confidence interval; E, number of events; ER, estrogen-receptor; HR, hazard ratio; N, number of patients at risk; PR, progesterone receptor.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. FUNDING SOURCES
  7. REFERENCES

We applied the propensity score method to reduce the impact of selection bias in this epidemiologic study examining the association between CPM and DFS or OS in a hospital-based cohort of patients with early-stage breast cancer. As expected, a significant imbalance was evident between the demographic and clinical factors of patients who underwent CPM and those of patients who did not undergo CPM. For example, patients who underwent CPM were more likely to be white and younger than 50 years and to have stage I to II disease and HER2/neu-negative disease. In the multivariable models adjusted for significant prognostic factors, CPM was associated with a statistically significant improvement in DFS and OS and the benefit was seen predominantly among patients with hormone receptor–negative disease. In the propensity score–matched multivariable analysis, the benefit of CPM on DFS and OS was not statistically significant, but the improved DFS seen among patients with hormone receptor–negative disease who received CPM was nearly statistically significant (P = .05).

The current body of evidence, albeit limited, is inconclusive regarding the survival benefit of CPM in breast cancer patients, although the use of CPM is estimated at 14% to 28% from 2006 to 2007 at academic medical centers.12, 17, 18 In our cohort, with a median follow-up time of 4.5 years, 68 women developed contralateral breast cancer and only 1 of these women was in the CPM group. Because the benefit of CPM would be evident only as a decreased risk of death from contralateral breast cancer, the study by Narod used actuarial methods to estimate the length of time that would be necessary to demonstrate the impact of CPM on breast cancer–related mortality.19 Among BRCA1 or BRCA2 (breast cancer early onset genes) mutation carriers with breast cancer and an estimated 2% annual risk of developing a contralateral breast cancer, the chance of dying from the contralateral breast cancer was estimated to be 0.4% after 5 years and 6.8% after 20 years of follow-up. These data suggest that any survival benefit from CPM would not be evident until after more than 10 years of follow-up.

The question therefore remains of why some epidemiologic studies have shown a DFS advantage associated with CPM after follow-up periods of <10 years (range, 3.9-6.8 years).9, 20, 21 Certainly, selection bias may play a role; our study and others have shown that patients who undergo CPM are more likely to be younger than 50 years, white, and educated and to have a family history of breast cancer and receive immediate breast reconstruction and BRCA1/2 testing12, 22-25 than are patients who do not undergo CPM. Another possibility is that our study, with its large sample size, may be more likely to detect a small benefit within a follow-up period of <10 years using the endpoint of breast cancer DFS than studies with smaller sample sizes. Although we attempted to control for selection bias using the propensity score method in this study, which is the first study to do so, it is likely that residual confounding remains because of unknown and known patient-related factors, such as increased compliance with breast cancer treatments and overall better health, that predispose some patients to better clinical outcomes from the index breast cancer. Indeed, Bedrosian et al showed that non–breast cancer–related deaths were less common among older patients who underwent CPM than among those who did not undergo CPM, suggesting a selection bias for healthier patients among those who undergo CPM.9

In our subgroup analyses by hormone receptor status, it is worth noting that the sample size in the matched analysis was much smaller than that in the unmatched adjusted multivariable analysis. The HR of 0.48 in the matched analysis of the hormone receptor–negative subgroup indicated a substantial risk reduction with a confidence interval ranging from 0.22 to 1.01, suggesting a possible lack of power to identify this association. Despite the smaller sample size in the matched analysis, there was still a borderline statistically significant benefit of CPM in hormone receptor–negative patients compared with the hormone receptor–positive patients. In an exploratory subgroup analysis, we found that among patients with hormone receptor–negative tumors, those with early-stage and triple-negative disease were most likely to benefit from CPM. The identification of subsets of patients with hormone receptor–negative disease most likely to benefit from CPM may have important implications for a more personalized approach to selecting patients for CPM.

Several limitations of our study should be considered. First, this study includes patients treated at a single institution, and the results may therefore not be generalizable to the whole population. Second, lack of knowledge of the BRCA1/2 germline mutation status of patients may lead to an overestimation of the benefit of CPM in reducing the risk of contralateral breast cancer, because patients with a BRCA1/2 mutation have a higher lifetime risk of contralateral breast cancer compared with patients who have sporadic breast cancer.26 Third, as in other epidemiologic studies, the short median follow-up times may result in an underestimation of the benefit of CPM on DFS.19 However, our cohort consisted of an ethnically diverse patient population with equal access to state-of-the-art breast cancer care and included important information on the receipt of adjuvant endocrine therapy, although data were not available on whether patients were compliant with adjuvant endocrine therapy.

In summary, given the inherent limitations of observational studies in controlling for confounders and eliminating selection bias, patients with sporadic breast cancer should be informed of alternative options to CPM for reducing the risk of developing a contralateral breast cancer, such as adjuvant treatment with tamoxifen or aromatase inhibitors for women with hormone receptor–positive disease. Future studies evaluating the clinical benefit of CPM on breast cancer survival should focus on subsets of patients, such as those with hormone receptor–negative breast cancer who do not receive adjuvant endocrine therapy. This group may benefit most from CPM, provided their risk of death from the index tumor does not outweigh their risk of developing a contralateral breast cancer. The identification of subgroups of patients most likely to benefit from CPM could impact the treatment decision-making process between patients and their physicians and help reduce breast cancer–related mortality.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. FUNDING SOURCES
  7. REFERENCES

National Cancer Institute grant R21 CA149803-01A1 (principal investigators, Abenaa Brewster and Patricia Parker) and National Cancer Institute grant RO1 CA079466 (to Yu Shen).

CONFLICT OF INTEREST DISCLOSURE

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. FUNDING SOURCES
  7. REFERENCES
  • 1
    Tuttle TM, Habermann EB, Grund EH, Morris TJ, Virnig BA. Increasing use of contralateral prophylactic mastectomy for breast cancer patients: a trend toward more aggressive surgical treatment. J Clin Oncol. 2007; 25: 5203-5209.
  • 2
    Lostumbo L, Carbine NE, Wallace J. Prophylactic mastectomy for the prevention of breast cancer. Cochrane Database Syst Rev. 2010; 10( 11): CD002748.
  • 3
    Adami HO, Bergstrom R, Hansen J. Age at first primary as a determinant of the incidence of bilateral breast cancer. Cumulative and relative risks in a population-based case-control study. Cancer. 1985; 55: 643-647.
  • 4
    Schairer C, Brown LM, Mai PL. Inflammatory breast cancer: high risk of contralateral breast cancer compared to comparably staged non-inflammatory breast cancer. Breast Cancer Res Treat. 2011; 129: 117-124.
  • 5
    Healey EA, Cook EF, Orav EJ, Schnitt SJ, Connolly JL, Harris JR. Contralateral breast cancer: clinical characteristics and impact on prognosis. J Clin Oncol. 1993; 11: 1545-1552.
  • 6
    Kollias J, Ellis IO, Elston CW, Blamey RW. Clinical and histological predictors of contralateral breast cancer. Eur J Surg Oncol. 1999; 25: 584-589.
  • 7
    Robbins GF, Berg JW. Bilateral primary breast cancer; a prospective clinicopathological study. Cancer. 1964; 17: 1501-1527.
  • 8
    Rosen PP, Groshen S, Kinne DW, Hellman S. Contralateral breast carcinoma: an assessment of risk and prognosis in stage I (T1N0M0) and stage II (T1N1M0) patients with 20-year follow-up. Surgery. 1989; 106: 904-910.
  • 9
    Bedrosian I, Hu CY, Chang GJ. Population-based study of contralateral prophylactic mastectomy and survival outcomes of breast cancer patients. J Natl Cancer Inst. 2010; 102: 401-409.
  • 10
    Helzlsouer KJ. Contralateral prophylactic mastectomy: quantifying benefits and weighing the harms. J Clin Oncol. 2005; 23: 4251-4253.
  • 11
    Stucky CC, Gray RJ, Wasif N, Dueck AC, Pockaj BA. Increase in contralateral prophylactic mastectomy: echoes of a bygone era? Surgical trends for unilateral breast cancer. Ann Surg Oncol. 2010; 17( suppl 3): 330-337.
  • 12
    Yi M, Hunt KK, Arun BK, Bedrosian I, Barrera AG, Do KA, et al. Factors affecting the decision of breast cancer patients to undergo contralateral prophylactic mastectomy. Cancer Prev Res (Phila). 2010; 3: 1026-1034.
  • 13
    Austin PC, Grootendorst P, Anderson GM. A comparison of the ability of different propensity score models to balance measured variables between treated and untreated subjects: a Monte Carlo study. Stat Med. 2007; 26: 734-753.
  • 14
    Parsons LS. Reducing bias in a propensity score matched-pair sample using greedy matching techniques group. In: Proceedings of the 26th Annual SAS Users Group International Conference; April 22-25, 2001; Long Beach, California. Paper 214-26.
  • 15
    Fleiss JL, Levin B, Paik MC. Statistical Methods for Rates and Proportions. 3rd edition. New York, NY: John Wiley & Sons; 2003.
  • 16
    Diggle P, Liang K-Y, Zeger SL. Analysis of Longitudinal Data. New York, NY: Oxford University Press, 1994.
  • 17
    Arrington AK, Jarosek SL, Virnig BA, Habermann EB, Tuttle TM. Patient and surgeon characteristics associated with increased use of contralateral prophylactic mastectomy in patients with breast cancer. Ann Surg Oncol. 2009; 16: 2697-2704.
  • 18
    Jones NB, Wilson J, Kotur L, Stephens J, Farrar WB, Agnese DM. Contralateral prophylactic mastectomy for unilateral breast cancer: an increasing trend at a single institution. Ann Surg Oncol. 2009; 16: 2691-2696.
  • 19
    Narod SA. The impact of contralateral mastectomy on mortality in BRCA1 and BRCA2 mutation carriers with breast cancer. Breast Cancer Res Treat. 2011; 128: 581-583.
  • 20
    Herrinton LJ, Barlow WE, Yu O, Geiger AM, Elmore JG, Barton MB, et al. Efficacy of prophylactic mastectomy in women with unilateral breast cancer: a cancer research network project. J Clin Oncol. 2005; 23: 4275-4286.
  • 21
    Peralta EA, Ellenhorn JD, Wagman LD, Dagis A, Andersen JS, Chu DZ. Contralateral prophylactic mastectomy improves the outcome of selected patients undergoing mastectomy for breast cancer. Am J Surg. 2000; 180: 439-445.
  • 22
    Polednak AP. Frequency of prophylactic contralateral mastectomy among breast cancer patients. J Am Coll Surg. 2001; 192: 804-805.
  • 23
    Sorbero ME, Dick AW, Beckjord EB, Ahrendt G. Diagnostic breast magnetic resonance imaging and contralateral prophylactic mastectomy. Ann Surg Oncol. 2009; 16: 1597-1605.
  • 24
    Yao K, Stewart AK, Winchester DJ, Winchester DP. Trends in contralateral prophylactic mastectomy for unilateral cancer: a report from the National Cancer Data Base, 1998-2007. Ann Surg Oncol. 2010; 17: 2554-2562.
  • 25
    King TA, Sakr R, Patil S, et al. Clinical management factors contribute to the decision for contralateral prophylactic mastectomy. J Clin Oncol. 2011; 29: 2158-2164.
  • 26
    Metcalfe K, Lynch HT, Ghadirian P, et al. Contralateral breast cancer in BRCA1 and BRCA2 mutation carriers. J Clin Oncol. 2004; 22: 2328-2335.