SEARCH

SEARCH BY CITATION

Abstract

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

BACKGROUND:

Randomized clinical trials (RCT) have demonstrated equivalent survival for breast-conserving therapy with radiation (BCT) and mastectomy for early-stage breast cancer. A large, population-based series of women who underwent BCT or mastectomy was studied to observe whether outcomes of RCT were achieved in the general population, and whether survival differed by surgery type when stratified by age and hormone receptor (HR) status.

METHODS:

Information was obtained regarding all women diagnosed in the state of California with stage I or II breast cancer between 1990 and 2004, who were treated with either BCT or mastectomy and followed for vital status through December 2009. Cox proportional hazards modeling was used to compare overall survival (OS) and disease-specific survival (DSS) between BCT and mastectomy groups. Analyses were stratified by age group (< 50 years and ≥ 50 years) and tumor HR status.

RESULTS:

A total of 112,154 women fulfilled eligibility criteria. Women undergoing BCT had improved OS and DSS compared with women with mastectomy (adjusted hazard ratio for OS entire cohort = 0.81, 95% confidence interval [CI] = 0.80-0.83). The DSS benefit with BCT compared with mastectomy was greater among women age ≥ 50 with HR-positive disease (hazard ratio = 0.86, 95% CI = 0.82-0.91) than among women age < 50 with HR-negative disease (hazard ratio = 0.88, 95% CI = 0.79-0.98); however, this trend was seen among all subgroups analyzed.

CONCLUSIONS:

Among patients with early stage breast cancer, BCT was associated with improved DSS. These data provide confidence that BCT remains an effective alternative to mastectomy for early stage disease regardless of age or HR status. Cancer 2013. © 2012 American Cancer Society.


INTRODUCTION

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

For more than 2 decades, it has been believed that early-stage breast cancer outcomes were equivalent with either mastectomy or breast conserving surgery with adjuvant radiation (breast conserving therapy [BCT]). Accordingly, BCT rates have increased throughout the United States.1, 2 The proportion of women choosing BCT has also increased among breast cancer patients of all ages3 and racial/ethnic groups, although important disparities still persist.4, 5

However, recent reports have suggested a rising use of mastectomy in more affluent regions,6 in those undergoing magnetic resonance imaging,7, 8 in younger women,6 and among those with in situ disease.6 Although the reasons contributing to this observed trend toward mastectomy remain largely speculative, they may relate to improvements in reconstruction techniques, or to women's attitudes toward mastectomy, including the desire to reduce anxiety associated with long-term surveillance. In addition, there has been renewed interest in mastectomy based on recent studies reporting higher recurrence rates among BCT recipients with tumor characteristics suggesting increased locoregional recurrence risk.9, 10 In light of the reports showing renewed interest in mastectomy, and because there have been few opportunities to observe long-term treatment-associated outcomes in the general population, we asked whether the comparable survival outcome of BCT compared with mastectomy as seen in randomized controlled trials could be generalized to the non–clinical trial population. We used the large, population-based prospective California Cancer Registry (CCR) to explore observational associations between surgery type, overall survival (OS) and breast cancer–specific survival among women diagnosed with stage I or II breast cancer to determine how the choice of mastectomy might affect long-term outcomes for early-stage disease, and importantly, whether these results persisted despite age at diagnosis or hormone receptor (HR) status.

MATERIALS AND METHODS

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

Case Ascertainment and Data Collection

Cases eligible for this study were all female residents of California newly diagnosed with invasive breast cancer (International Classification of Diseases-Oncology, 3rd edition [ICD-O-3]) morphology codes C50.0-50.9 and reported to the state-mandated CCR with a new, unilateral T1/T2 stage I or II disease diagnosed during the period January 1, 1990, through December 31, 2004. Eligible cases were treated with either definitive BCT defined by receipt of both lumpectomy and radiation, or mastectomy without radiation and had at least 5 years of follow-up. Patients treated with either lumpectomy without radiation or mastectomy with radiation were excluded. Vital status and underlying cause of death if applicable, routinely determined by the CCR through hospital follow-up and routine linkage with state and national death indices, were collected up to December 31, 2009. A total of 215,049 women aged 18 to 80 years who were diagnosed with unilateral breast cancer between 1990 and 2004 were identified. Cases were excluded sequentially as follows: 29,713 for a diagnosis of a previous primary cancer; 36,115 for treatment with lumpectomy without adjuvant radiation (23,617) or mastectomy with adjuvant radiation (11,002) or no/unknown/other treatment (1496); 30,870 women for unknown hormone receptor (HR) status; 4074 for tumor size missing or > 5 cm; and 2123 for date of surgery more than 120 days after the initial date of diagnosis, based on the concern that these cases may have been treated with neoadjuvant chemotherapy. Node status was treated as a categorical variable according to proportion of positive nodes, because current nodal staging of N2 and N3 disease includes internal mammary node data, and this information was not reliably collected by the registry. The final cohort fulfilling all eligibility criteria consisted of 112,514 women.

Tumor Subtype Information

The CCR has collected information on ER and progesterone receptor (PR) expression of breast cancers since 1990, the first year of cohort for this study. Each marker was classified as positive (at least 5% nuclear staining), negative, borderline, not tested, not recorded, or unknown on the basis of pathology or medical record information. Cases were deemed HR-positive if tumors expressed ER, PR, or both. If both receptors were negative, cases were classified as HR-negative. Patients whose receptor status was borderline, not tested, or not recorded were excluded. Human epidermal growth factor receptor 2 (HER2) receptor testing results were not available uniformly for years prior to 2006, and were therefore not included in the analysis.

Neighborhood Socioeconomic Information

We used a measure of neighborhood-level socioeconomic status (SES) assigned to patients' residential address at diagnosis, based on a previously described principal components–derived index that incorporates census-block group level Census 1990 or 2000 data on education, income, occupation, and housing costs, among others.11 Each case was assigned a neighborhood SES quintile based on the distribution of the composite SES index across the state of California, with 1 representing the lowest SES quintile.

Statistical Analysis

Overall survival between surgery groups was compared using a Kaplan-Meier estimator. Outcome between surgery groups was compared using the log-rank comparison. Cumulative incidence functions were used to estimate the probability of 5-year breast cancer–specific survival in the presence of competing risks. Adjusted Cox proportional hazards models were used to estimate hazard ratios for disease-specific survival (DSS) or OS between BCT and mastectomy groups; the full models included surgery type, tumor grade, tumor size (continuous variable), proportion of positive lymph nodes, age at diagnosis, year of diagnosis, SES, and race. For cause-specific survival, cases who were alive or had died of other causes were censored at time of last follow-up or death. Separate analyses were conducted for each combination of HR status and age group < 50 versus ≥ 50 based on age at diagnosis. To test the proportionality assumption we graphed the log(-log) of survival versus the log of time. No violations were found. Competing risk analyses were conducted using the cmprsk package implemented in R (R Foundation for Statistical Computing, Vienna, Austria). All other statistical analyses were conducted in SAS version 9.2 (SAS Institute, Cary, NC).

RESULTS

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

The cohort of 112,154 eligible women included 61,771 (55%) who received lumpectomy and radiation and 50,383 (45%) who had mastectomy without radiation. Median follow-up time was 110.6 months.

Distributions of Patient/Tumor Factors by Surgery Type

Table 1 shows demographic and clinical characteristics of the patient cohort. A total of 25% of the BCT group and 26% of the mastectomy group were younger than 50 years of age at diagnosis; only 6.0% of the cohort overall were younger than 40 years, and 23% were between 70 and 80 years at diagnosis. Nearly a quarter of the cohort was nonwhite, including 11.5% Hispanic, 8.6% Asian/Pacific Islander, and 5.0% black.

Table 1. Distribution of Demographic and Clinical Characteristics Among Women in California Diagnosed With Stage I or II Breast Cancer, by Surgical Treatment, 1990-2004
 Lumpectomy GroupMastectomy Group
 N = 61,771 (55%)N = 50,383 (45%)
CharacteristicN%N%
Cohort Characteristics    
Age at examination    
 ≤3934305.5532836.52
 40-4912,19419.74993519.72
 50-5916,84427.2711,69023.2
 60-6915,91825.7712,54524.9
 70-8013,38521.6712,93025.66
Year of diagnosis    
 1990-199250188.12850916.89
 1993-199510,63617.2210,95421.74
 1996-199813,95022.5810,49520.83
 1999-200115,81725.6110,35120.54
 2002-200416,35026.4710,07419.99
Race    
 White, non-Hispanic47,40476.7435,89071.23
 Black, non-Hispanic31255.0625335.03
 Hispanic653010.57632312.55
 Asian/native Hawaiian/Pacific islander43547.05530210.52
 American indian/Alaska native1730.281410.28
 Other (include mixed)/unknown1850.31940.39
Socioeconomic status quintile    
 1580111.5147837.74
 2885117.57852813.81
 310,53120.9012,25919.85
 411,72523.2715,54025.16
 513,47526.7520,66133.45
Tumor Characteristics    
Stage    
 I40,26665.1922,43644.53
 IIA16,43326.618,22436.17
 IIB50728.21972319.3
Hormone receptor status    
 Negative996916.14978519.42
 Positive51,80283.8640,59880.58
Histology    
 Ductal47,87077.537,45574.34
 Lobular39086.3344448.82
 Both48967.9349409.8
 Other50978.2535447.03
Grade    
 I15,00724.29755314.99
 II24,63839.8919,13637.98
 III or IV16,61726.916,58132.91
 Unknown55098.92711314.12
Node status    
 Negative47,26276.5132,52364.55
 Positive13,64222.0817,47034.67
 Unknown8671.43900.77
Tumor size    
 <0.5 cm18092.9313082.6
 0.5-0.9 cm1180019.1523210.38
 1-1.4 cm1688427.33925218.36
 1.5-1.9 cm1368322.15951818.89
 2-2.9 cm1277220.681389527.58
 3-3.9 cm35365.72680713.51
 4-5 cm12872.0843718.68

Surgical practice patterns for early-stage breast cancers differed by year of diagnosis between 1990 and 2004. In the overall cohort, the proportion of patients treated with BCT increased from 37% between 1990 and 1992 to 62% between 2002 and 2004. The proportion of women undergoing lumpectomy was nonlinear with respect to decade of age at diagnosis; 51% of women under the age of 40 underwent BCT compared with 59% of women aged 50 to 59 years and 51% of women aged 70 to 80 years. Non-Hispanic whites had the highest rate of BCT (57%) followed by non-Hispanic blacks, Hispanics, and non-Hispanic Asians/Pacific Islanders (55%, 51%, and 45%, respectively). The mastectomy rate declined among all racial groups during the study period.

Median tumor size was 1.5 cm, with larger average tumor size seen among the mastectomy group, as expected. The rate of BCT among tumor size categories was compared among age groups (< 40, 40-54, 55-69, 70-80) and decreased with tumor size across all age groups (Fig. 1). Interestingly, the use of BCT varied by age even among tumors ≤ 2 cm where the youngest and oldest age groups had the lowest BCT rate. In larger tumors (> 2 cm), BCT rate declined by age (Fig. 1).

thumbnail image

Figure 1. Graph shows the impact of tumor size and age at diagnosis on rate of breast-conserving therapy (BCT) in women with stage I or II breast cancer, from 1990-2004.

Download figure to PowerPoint

Differences in surgery type were seen among tumor histologies. Lobular cancers were associated with a higher mastectomy rate when compared to ductal tumors (53% compared to 44%, respectively), a result that was seen even when limiting the analysis to T1 lesions only (44% compared to 38%). A greater proportion of grade I tumors were more likely to be treated with BCT than non–low grade tumors. Patients with grade III and HR-negative tumors were more likely to undergo mastectomy than the overall population (50% in both groups).

Impact of Surgery Type on Survival Endpoints

A total of 31,416 deaths were identified in the cohort during the study period, including 39% due to breast cancer as the underlying cause. The 5-year OS in the group was 89.3% (95% confidence interval [CI] = 89.2%-89.5%); DSS was 94.4% (95% CI = 94.2%-94.5%)

Cox multivariable analysis was performed to compare OS among 4 groups: 1) age ≥ 50, HR-negative; 2) age ≥ 50, HR-positive; 3) age < 50, HR-negative; and 4) age < 50, HR-positive using the full model described above (data not shown). Among all patient groups, BCT was consistently associated with improved OS when compared with mastectomy. The group achieving greatest benefit in OS with BCT relative to mastectomy were women aged 50 and older at diagnosis with HR-positive tumors (hazard ratio = 0.81, 95% CI = 0.79-0.84). The smallest benefit was seen among those women younger than age 50 at diagnosis with HR-positive tumors (hazard ratio = 0.93, 95% CI = 0.86-0.99). Furthermore, the greatest benefit of BCT for DSS was seen among women age 50 and older at diagnosis with HR-positive disease (hazard ratio = 0.87, 95% CI = 0.82-0.92), with other groups showing minimal or no significant benefit of BCT over mastectomy (Table 2).

Table 2. Cox Multivariable Analysis: Breast Cancer–Specific Survival (DSS) of Women Diagnosed With Stage I or II Breast Cancer (1990-2004) With at Least 5 Years of Follow-Up
  Age ≥50, HR-NegativeAge ≥50, HR-PositiveAge <50, HR-NegativeAge <50, HR-Positive
  Hazard Ratio95% Confidence LimitsHazard Ratio95% Confidence LimitsHazard Ratio95% Confidence LimitsHazard Ratio95% Confidence Limits
  1. Abbreviations: BCS, breast-conserving surgery; HR, hazard ratio; NH, non-Hispanic; SES, socioeconomic status.

SurgeryMastectomy1.0 1.0 1.0 1.0 
 BCS+radiation0.93(0.85-1.02)0.86(0.82-0.91)0.88(0.79-0.98)0.94(0.87-1.02)
GradeGrade I1.0 1.0 1.0 1.0 
 Grade II1.99(1.43-2.77)1.81(1.65-1.99)1.34(0.81-2.21)2.45(2.01-2.98)
 Grade III or IV2.51(1.82-3.46)3.11(2.82-3.43)1.41(0.87-2.29)3.76(3.09-4.57)
 Unknown1.75(1.24-2.48)1.77(1.58-1.97)1.03(0.62-1.73)2.11(1.69-2.64)
Proportion nodes positiveNone1.0 1.0 1.0 1.0 
 <10% positive1.45(1.25-1.67)1.63(1.49-1.78)1.71(1.44-2.02)1.69(1.50-1.92)
 10%-24% positive2.17(1.90-2.47)2.00(1.85-2.16)2.31(1.99-2.69)2.25(2.01-2.52)
 25+% positive3.50(3.14-3.90)3.78(3.53-4.04)3.74(3.27-4.29)3.60(3.25-3.99)
 Unknown1.37(1.10-1.70)1.51(1.35-1.68)2.1(1.57-2.79)2.15(1.73-2.67)
Race/ethnicityNH white1.0 1.0 1.0 1.0 
 Asian/Pacific islander0.62(0.52-0.75)0.87(0.78-0.97)0.86(0.71-1.05)1.00(0.88-1.13)
 Hispanic1.03(0.91-1.18)1.04(0.96-1.14)1.02(0.89-1.18)0.95(0.85-1.06)
 NH American indian/other/unknown0.49(0.23-1.03)0.67(0.44-1.01)0.89(0.44-1.79)0.86(0.5-1.49)
 NH black1.18(1.02-1.37)1.32(1.17-1.47)1.39(1.18-1.65)1.30(1.13-1.50)
SES quintile5 (highest)1.0 1.0 1.0 1.0 
 1 (lowest)1.10(0.95-1.29)1.27(1.15-1.4)1.30(1.08-1.57)1.46(1.26-1.68)
 21.14(1.00-1.30)1.19(1.10-1.29)1.28(1.08-1.52)1.48(1.31-1.67)
 31.08(0.95-1.23)1.15(1.07-1.24)1.09(0.93-1.29)1.19(1.06-1.33)
 41.06(0.94-1.20)1.11(1.03-1.19)1.11(0.95-1.30)1.10(0.99-1.23)
Tumor size, cm 1.35(1.30-1.41)1.42(1.39-1.46)1.19(1.13-1.25)1.32(1.27-1.37)
Age at diagnosis, y 1.01(1.00-1.01)1.01(1.01-1.02)0.99(0.98-1.00)0.96(0.95-0.97)

Kaplan-Meier survival estimates showed significantly greater OS and breast cancer–specific survival favoring BCT over mastectomy. This benefit was seen in women in younger and older age groups, as well as those with HR-positive and HR-negative disease (Table 2, Fig. 2). The interaction between tumor size and surgical treatment was further evaluated. The OS and DSS benefit of BCT were greater among T1 than T2 tumors in all age and HR status strata analyzed. However, OS favored the BCT group even among those with T2 tumors. The advantage favoring BCT was greatest in women age 50 and older at diagnosis with HR-positive disease (hazard ratio = 0.86, 95% CI = 0.82-0.91).

thumbnail image

Figure 2. (A) Kaplan-Meier overall survival estimates are shown for hormone receptor (HR)-negative patients, age < 50 years at diagnosis: comparison of breast-conserving therapy (breast-conserving surgery [BCS] and radiation [R]) and mastectomy (M) groups. (B) Kaplan-Meier overall survival estimates are shown for HR-positive patients, age ≥ 50 years at diagnosis: comparison of breast-conserving therapy (BCS+R) and mastectomy (M) groups.

Download figure to PowerPoint

In order to estimate whether greater burden of underlying comorbidities could have affected treatment choice of BCT or mastectomy, several different competing causes of mortality were evaluated. We examined several outcomes independently including OS, DSS, heart disease–specific survival, other cancer-specific survival, cerebrovascular disease-–specific survival, and respiratory disease–specific survival. To more specifically evaluate those comorbidities likely to have been present at the time of breast cancer diagnosis, we examined mortality by truncating follow-up at 3 years after diagnosis (Table 3). Notably, BCT was associated with significantly lower 3-year mortality rates from all causes, with the lowest hazard seen for heart disease–specific survival (hazard ratio = 0.51, 95% CI = 0.44-0.60) and respiratory disease–specific survival (hazard ratio = 0.46, 95% CI = 0.34-0.62). The smallest reduction in hazard favoring BCT was seen for breast cancer–specific survival (hazard ratio = 0.85, 95% CI = 0.78-0.91); however, BCT was associated with improved survival for all outcomes.

Table 3. Cox Multivariable Analysis: Overall Survival Among Women Diagnosed With Stage I or II Breast Cancer and Dying Within 3 Years of Diagnosis, by Cause of Death
 OverallBreast CancerHeart Disease
Cause of Death Hazard Ratio95% Confidence LimitsHazard Ratio95% Confidence LimitsHazard Ratio95% Confidence Limits
SurgeryMastectomy1.0 1.0 1.0 
 BCS+radiation0.72(0.68-0.76)0.84(0.78-0.91)0.51(0.43-0.59)
GradeGrade I1.0 1.0 1.0 
 Grade II1.11(1.01-1.22)2.29(1.87-2.82)0.97(0.79-1.19)
 Grade III or IV1.70(1.55-1.86)4.78(3.91-5.86)1.25(1.00-1.56)
 Unknown1.17(1.04-1.30)2.51(2.00-3.14)1.11(0.86-1.43)
ProportionNone1.0 1.0 1.0 
nodes positive<10% positive1.31(1.20-1.44)1.84(1.64-2.07)1.00(0.76-1.32)
 10%-24% positive1.71(1.58-1.85)2.45(2.21-2.72)0.88(0.67-1.15)
 25+% positive2.76(2.58-2.95)4.28(3.92-4.67)1.40(1.12-1.76)
 Unknown2.05(1.86-2.25)1.99(1.69-2.35)2.07(0.76-1.32)
Race/ethnicityNH white1.0 1.0 1.0 
 Asian/Pacific Islander0.72(0.65-0.80)0.71(0.61-0.82)0.74(0.53-1.03)
 Hispanic0.93(0.86-1.01)1(0.90-1.10)0.63(0.47-0.83)
 NH American Indian1.30(0.86-1.96)1.08(0.61-1.91)0.97(0.24-3.92)
 NH black1.19(1.08-1.31)1.19(1.05-1.34)1.46(1.12-1.91)
 Other/unknown0.30(0.13-0.68)0.22(0.05-0.89)0.70(0.17-2.84)
SES quintile5 (highest)1.0 1.0 1.0 
 1 (lowest)1.55(1.42-1.70)1.30(1.15-1.47)2.48(1.93-3.17)
 21.43(1.32-1.55)1.33(1.20-1.48)2.04(1.63-2.55)
 31.28(1.19-1.38)1.19(1.07-1.32)1.65(1.32-2.06)
 41.15(1.07-1.24)1.07(0.97-1.19)1.26(1.00-1.59)
Tumor size (cm) 1.32(1.29-1.35)1.45(1.40-1.49)1.20(1.12-1.29)
Age at<501.0 1.0 1.0 
Diagnosis, y50-540.91(0.82-1.02)0.87(0.77-0.98)1.76(0.96-3.22)
 55-591.17(1.06-1.29)1.03(0.91-1.16)3.47(2.07-5.81)
 60-641.46(1.32-1.60)1.14(1.01-1.29)5.50(3.42-8.85)
 65-691.85(1.69-2.03)1.26(1.12-1.42)10.74(6.92-16.68)
 70-742.42(2.22-2.63)1.34(1.19-1.52)16.90(11.01-25.94)
 75-803.41(3.15-3.7)1.65(1.47-1.86)30.64(20.15-46.59)
HR statusPositive1.0 1.0 1.0 
 Negative2.16(2.04-2.29)3.03(2.81-3.26)0.96(0.78-1.18)
  Cerebrovascular DiseaseChronic Lower Respiratory Disease
Cause of Death Hazard Ratio95% Confidence LimitsHazard Ratio95% Confidence Limits
  1. Abbreviations: BCS, breast-conserving surgery; HR, hazard ratio; NH, non-Hispanic; SES, socioeconomic status.

SurgeryMastectomy1.0 1.0 
 BCS+radiation0.64(0.48-0.86)0.45(0.33-0.62)
GradeGrade I1.0 1.0 
 Grade II0.80(0.57-1.12)0.88(0.62-1.26)
 Grade III or IV0.56(0.36-0.86)0.68(0.44-1.06)
 Unknown0.88(0.57-1.37)0.59(0.34-1.00)
ProportionNone1.0 1.0 
nodes positive<10% positive1.02(0.61-1.70)0.60(0.30-1.19)
 10%-24% positive0.85(0.51-1.42)0.87(0.50-1.50)
 25+% positive0.82(0.49-1.37)1.35(0.85-2.14)
 Unknown1.96(1.34-2.88)3.39(2.32-4.95)
Race/ethnicityNH white1.0 1.0 
 Asian/Pacific Islander1.58(1-2.50)0.23(0.08-0.63)
 Hispanic0.88(0.53-1.47)0.40(0.21-0.75)
 NH American Indian2.08(0.29-14.93)0(0-1.03E+29)
 NH black1.64(0.96-2.80)0.29(0.10-0.79)
 Other/unknown1.35(0.19-9.68)0(0-4.32E+24)
SES quintile5 (highest)1.0 1.0 
 1 (lowest)1.51(0.95-2.39)2.69(1.67-4.33)
 20.93(0.60-1.46)1.67(1.07-2.61)
 31.39(0.96-2.02)1.57(1.03-2.40)
 41.16(0.79-1.69)1.44(0.95-2.19)
Tumor size, cm 1.34(1.18-1.52)1.23(1.07-1.42)
Age at diagnosis, y<501.0 1.0 
 50-541.28(0.30-5.38)2.42(0.73-7.94)
 55-593.86(1.29-11.55)1.94(0.56-6.73)
 60-648.70(3.25-23.26)8.05(3.02-21.41)
 65-6912.17(4.69-31.58)18.17(7.22-45.74)
 70-7423.79(9.46-59.85)19.13(7.61-48.07)
 75-8049.25(19.92-121.75)20.4(8.13-51.15)
HR statusPositive1.0 1.0 
 Negative1.25(0.84-1.84)1.40(0.94-2.08)

DISCUSSION

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

Numerous prospective randomized trials suggest that BCT and mastectomy result in equivalent OS. Despite this, recent studies have shown an increased rate of mastectomy for patient subgroups including younger women with early-stage tumors, many of which would have presumably been amenable to BCT. One likely contributing factor leading to an increased uptake of mastectomy is a perception of worse outcomes among women receiving BCT for tumors with unfavorable factors (eg, young age, ER-negative disease, Her2-positive disease).12 The goal of this study was to determine whether the equivalent outcomes from BCT and mastectomy predicted from randomized controlled trials were achieved in the general population, and to find whether there existed a subgroup for which this equivalence was not demonstrated. We observed the unanticipated finding that patients treated with mastectomy had a significantly lower OS and breast cancer–specific survival than women who underwent BCT, even after adjusting for tumor grade, proportion of positive nodes, race, SES, tumor size, age at diagnosis, and year of diagnosis.

Our findings have important implications for understanding the overall benefit of BCT at the population level. Long-term breast cancer–specific survival after BCT was either equivalent or superior to mastectomy in all age groups and HR tumor types, with almost 10 years of median follow-up. The strongest association with outcome was observed in those women older than 50 years with HR-positive tumors who had a 13% lower breast cancer mortality and 19% lower overall mortality with BCT than with mastectomy (Table 2). These findings support the notion that BCT, when combined with radiation, confers at least equivalent and perhaps even superior survival to mastectomy as definitive breast cancer treatment. The greatest potential benefit to be gained was among postmenopausal women with ER-positive T1 tumors, but at least equivalent outcomes were seen among all groups on multivariable analysis, with the exception of women who had tumor sizes of 4 to 5 cm.

Determinants of Locoregional Treatment

To put these findings into context, it is important to consider the factors determining receipt of BCT or mastectomy, including both patient- and disease-related variables. We noted that non-Hispanic whites and women residing in higher SES neighborhoods were more likely to undergo lumpectomy between 1990 and 2004. Although we adjusted for age, race, and SES, there are additional patient-level variables that we could not account for in our model. These include health care access issues such as health insurance coverage, distance to closest radiation facility, availability of reconstructive surgeons, and provider-related factors, all of which have been shown to affect rates of BCT.1, 2, 13-17 Further, the cancer registry does not collect information regarding individual health history such as comorbid conditions, previous chest wall radiation, tumor-to-breast ratio, and other medical contraindications to radiation or lumpectomy. In aggregate, it is likely that these unmeasured variables could have resulted in higher mastectomy rates in those patients with more limited access to medical care, biasing our findings toward higher overall mortality and specific-cause mortality among women undergoing mastectomy. The magnitude and consequence of these potential biases is difficult to estimate but may in part account for the observed findings.

Among the disease-related factors that were collected, larger tumor size was associated with greater likelihood of undergoing mastectomy, as expected. However, other adverse tumor features that would not be expected to affect choice of breast surgery, including positive node status and higher tumor grade, also increased the likelihood of mastectomy. In addition, factors such as node status are commonly not known at the time that surgery is planned; nonetheless, 44% of women with node-positive disease underwent BCT compared with 59% who had node-negative cancer. This difference can be partly attributed to the larger tumor size in the mastectomy group. Recognizing this potential confounder, the final model was adjusted for tumor size as a continuous variable as well as for proportion of positive nodes, because the latter is a more precise measure of nodal tumor burden than node status as a binomial variable. Other pathologic variables such as lymphovascular invasion and extranodal extension, both measures of tumor aggressiveness, were not accounted for, nor was Her2neu status, because these data were not routinely reported to cancer registries prior to 2004. Although we were not able to adjust for these factors, it is unlikely that their cumulative impact would override the influence of tumor size, node status, and tumor grade.

Alternate Explanations

Many patient or tumor characteristics not reported to the cancer registry may have influenced the recommendation for and choice of lumpectomy or mastectomy. One important source of confounding could have been competing causes of mortality from concurrent comorbidities, which could not directly be assessed with the CCR data. Indeed, Giordano et al cautions against misinterpreting treatment and mortality effects based on population-based cancer registry data, due to concerns about confounding by indication.18 However, we conducted several analyses to directly address these concerns, including evaluating mortality due to varying causes of death within 3 years of the index breast cancer diagnosis as a proxy for preexisting comorbidities that may have influenced surgical treatment decisions. Interestingly, for every cause of mortality that we evaluated, women who had mastectomy were more likely to die within 3 years of their breast cancer diagnosis than women who chose BCT. Based on these findings, it is reasonable to infer that the mastectomy group was likely to have a greater burden of nonfatal comorbidities at presentation, and that this factor may well have influenced surgical decision-making. Nevertheless, this factor alone cannot account for why women with mastectomy had lower DSS after adjusting for age and tumor characteristics.

Strengths and Limitations of Study Design

The CCR is a large and representative source of long-term outcome data for women diagnosed and treated for breast cancer within the state of California. The strength of such a data source is that it allows a greater understanding of how BCT and mastectomy are used in a general and ethnically diverse population, not specific to particular health care institutions or settings. The long duration of follow-up provided in this study is important in order to minimize over-representation of breast cancer mortality by those tumor subtypes known to recur early.

There are clearly important limitations inherent in using a large population-based dataset. As evidenced above, cancer registries such as the CCR are not equipped to collect granular clinical data such as coexisting comorbidities, detailed pathologic data, and individual patient and provider biases, all of which may affect choice of treatment. Cancer registry data on radiation therapy are known to be undercaptured, due to radiation that is administration in free-standing facilities, which is not always reported to the regional cancer registry.19, 20 However, none of these factors would be expected to result in a sufficiently strong bias to override the known impact of tumor size, node status, age at diagnosis, race, and SES, all of which were included in the final multivariable model.

Conclusions

In a population-based cohort with early-stage breast cancer, BCT was independently associated with an advantage in breast cancer–specific survival at almost 10 years of follow-up. The magnitude of this benefit was greatest among women 50 years or older at diagnosis with HR-positive tumors, although this effect was seen regardless of HR status and age at diagnosis. These results provide confidence in the efficacy of BCT even among younger patients with HR-negative disease thought to be at relatively higher risk for local failure.

FUNDING SOURCES

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

This research was supported by the National Cancer Institute's Surveillance, Epidemiology and End Results Program under contract HHSN261201000140C awarded to the Cancer Prevention Institute of California. The collection of cancer incidence data used in this study was supported by the California Department of Health Services as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885.

CONFLICT OF INTEREST DISCLOSURE

The authors made no disclosure.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SOURCES
  8. REFERENCES
  • 1
    Albain KS, Green SR, Lichter AS, et al. Influence of patient characteristics, socioeconomic factors, geography, and systemic risk on the use of breast-sparing treatment in women enrolled in adjuvant breast cancer studies: an analysis of two intergroup trials. J Clin Oncol. 1996; 14: 3009-3017.
  • 2
    Nattinger AB, Gottlieb MS, Veum J, Yahnke D, Goodwin JS. Geographic variation in the use of breast-conserving treatment for breast cancer. N Engl J Med. 1992; 326: 1102-1107.
  • 3
    Gilligan MA, Kneusel RT, Hoffmann RG, Greer AL, Nattinger AB. Persistent differences in sociodemographic determinants of breast conserving treatment despite overall increased adoption. Med Care. 2002; 40: 181-189.
  • 4
    Freedman RA, He Y, Winer EP, Keating NL. Trends in racial and age disparities in definitive local therapy of early-stage breast cancer. J Clin Oncol. 2009; 27: 713-719.
  • 5
    Gross CP, Smith BD, Wolf E, Andersen M. Racial disparities in cancer therapy: did the gap narrow between 1992 and 2002? Cancer. 2008; 112: 900-908.
  • 6
    Gomez SL, Lichtensztajn D, Kurian AW, et al. Increasing mastectomy rates for early-stage breast cancer? Population-based trends from California. J Clin Oncol. 2010; 28: e155-e157.
  • 7
    Itakura K, Lessing J, Sakata T, et al. The impact of preoperative magnetic resonance imaging on surgical treatment and outcomes for ductal carcinoma in situ. Clin Breast Cancer. 2011; 11: 33-38.
  • 8
    Katipamula R, Degnim AC, Hoskin T, et al. Trends in mastectomy rates at the Mayo Clinic Rochester: effect of surgical year and preoperative magnetic resonance imaging. J Clin Oncol. 2009; 27: 4082-4088.
  • 9
    van der Sangen MJ, van de Wiel FM, Poortmans PM, et al. Are breast conservation and mastectomy equally effective in the treatment of young women with early breast cancer? Long-term results of a population-based cohort of 1,451 patients aged ≤ 40 years. Breast Cancer Res Treat. 2011; 127: 207-215.
  • 10
    Jones HA, Antonini N, Hart AA, et al. Impact of pathological characteristics on local relapse after breast-conserving therapy: a subgroup analysis of the EORTC boost versus no boost trial. J Clin Oncol. 2009; 27: 4939-4947.
  • 11
    Yost K, Perkins C, Cohen R, Morris C, Wright W. Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes Control. 2001; 12: 703-711.
  • 12
    Nguyen PL, Taghian AG, Katz MS, et al. Breast cancer subtype approximated by estrogen receptor, progesterone receptor, and HER-2 is associated with local and distant recurrence after breast-conserving therapy. J Clin Oncol. 2008; 26: 2373-2378.
  • 13
    Boscoe FP, Johnson CJ, Henry KA, et al. Geographic proximity to treatment for early stage breast cancer and likelihood of mastectomy. Breast. 2011; 20: 324-328.
  • 14
    Smith GL, Shih YC, Xu Y, et al. Racial disparities in the use of radiotherapy after breast-conserving surgery: a national Medicare study. Cancer. 2010; 116: 734-741.
  • 15
    Chagpar AB, Studts JL, Scoggins CR, et al. Factors associated with surgical options for breast carcinoma. Cancer. 2006; 106: 1462-1466.
  • 16
    Elward KS, Penberthy LT, Bear H, Swartz DM, Boudreau RM, Cook SS. Variation in the use of breast-conserving therapy for Medicare beneficiaries in Virginia: clinical, geographic, and hospital characteristics. Clin Perform Qual Health Care. 1998; 6: 63-69.
  • 17
    Morrow M, Jagsi R, Alderman AK, et al. Surgeon recommendations and receipt of mastectomy for treatment of breast cancer. JAMA. 2009; 302: 1551-1556.
  • 18
    Giordano SH, Kuo YF, Duan Z, Hortobagyi GN, Freeman J, Goodwin JS. Limits of observational data in determining outcomes from cancer therapy. Cancer. 2008; 112: 2456-2466.
  • 19
    Jagsi R, Abrahamse P, Hawley ST, Graff JJ, Hamilton AS, Katz SJ. Underascertainment of radiotherapy receipt in Surveillance, Epidemiology, and End Results registry data. Cancer. 2012; 118: 333-341.
  • 20
    Virnig BA, Warren JL, Cooper GS, Klabunde CN, Schussler N, Freeman J. Studying radiation therapy using SEER-Medicare-linked data. Med Care. 2002; 40( 8 suppl): IV-49 to IV-54.