SEARCH

SEARCH BY CITATION

Keywords:

  • triple-negative breast cancer;
  • Hispanic;
  • overweight;
  • obesity;
  • menopausal status

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

BACKGROUND:

Triple-negative breast cancer (TNBC) is defined as breast cancer that is negative for estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2. TNBC represents 15% of all invasive breast cancers, but some studies have suggested that its prevalence differs between races. To the authors' knowledge, no previous studies have determined the prevalence of TNBC and its risk factors among Hispanic women.

METHODS:

The authors identified 2074 Hispanic women with breast cancer who attended the National Cancer Institute in Mexico City from 1998 to 2008. All histopathologic and immunohistochemical diagnoses were rereviewed by a breast cancer pathologist. The prevalence of TNBC, its association with clinicopathologic characteristics, and its prognostic impact were determined.

RESULTS:

The median patient age at diagnosis (±standard deviation) was 50 ± 12 years. The overall prevalence of TNBC was 23.1%. Younger age (P < .001), premenopausal status (P = .002), increased parity (P = .029), hormonal contraceptive use (P = .04) high histologic grade (P < .001), and advanced disease (P < .001) were associated independently with TNBC. Postmenopausal patients who had a body mass index (BMI) <25 kg/m2 (P = .027) or <30 kg/m2 (P < .001) were more likely to have TNBC. In multivariate analysis, patients with TNBC had a higher risk of locoregional recurrence (LRR), lower disease-free survival (DFS) (hazard ratio, 1.62; 95% confidence interval, 1.13-2.32; P = .009), and a lower cancer-specific survival (CSS) rate (hazard ratio, 1.66; 95% confidence interval, 1.20-2.30; P = .002) than patients with non-TNBC.

CONCLUSIONS:

The median age at diagnosis of Hispanic women with breast cancer was 11 years younger than the average age reported in the United States. The prevalence of TNBC in this study population was higher than that reported in white women with breast cancer. TNBC was associated with a higher risk of LRR and with lower DFS and CSS than those in patients with non-TNBC. Cancer 2011;. © 2011 American Cancer Society.

Breast cancer is the leading cause of cancer deaths among women in the United States and Mexico.1 Several clinical and histopathologic prognostic factors for breast cancer have been studied; however, among the most important are the expression of 3 biomarkers: estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2).2-6 Negative staining for these 3 markers defines the triple-negative (TN) breast cancer (TNBC) phenotype,7 which represents approximately 15% of all invasive breast cancers.8 TNBCs are characterized by a frequent ductal histology, high grade, and high proliferation and mitotic rates.9 Women with TNBC have high recurrence rates and poor cancer-specific survival (CSS).10 This subtype is associated with high mortality and few therapeutic options, and conventional chemotherapy is the only effective treatment.11

The molecular biology and pathophysiology of TNBC are not completely understood. Nevertheless, TN tumors have a molecular profile associated with biologic aggressiveness, absence of BCL-2 expression, high p53 content, and breast cancer 1 gene (BRCA1) mutation.12, 13 TNBC often is categorized as a basal-like tumor: a distinct biologic subtype that is identified by gene expression profiling.14 Furthermore, patients with TNBC have a response to chemotherapy and a recurrence pattern that differ from those for patients with other breast cancers.15-17

Breast cancer is a genetic, heterogeneous disease and has important variability according to ethnicity and race with respect to incidence, clinical characteristics, and prognosis. Breast cancer in African-American women has unfavorable characteristics, such as absence of ER expression; in addition, TNBC is more prevalent in African-American women with breast cancer, contributing to increased mortality.18

Hypertension, diabetes, overweight, and obesity also have been associated with increased risk of breast cancer and with decreased survival and poor prognosis in premenopausal and postmenopausal women with breast cancer.19, 20 Some studies have associated overweight and obesity with the presence of the TN phenotype.21

To our knowledge, the prevalence and prognosis for Hispanic women with TNBC have not been studied previously. Notably, Mexico has a very high prevalence of overweight and obesity.22, 23 In the current study, we sought to evaluate and compare the demographic, clinical, and pathologic characteristics of breast cancer patients with and without TNBC who were seen at the National Cancer Institute (INCan), which is a national referral center for patients with cancer in Mexico City.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Study Population and Data Collection

We retrospectively reviewed the clinical records from patients with breast cancer who were seen at the INCan between January 1998 and December 2008. Included in our analysis were all Hispanic women who were diagnosed with invasive breast cancer for which ER, PR, and HER2 status was available. For eligible patients, we obtained the following clinical and general demographic data from the patient records: age at diagnosis; diagnosis of arterial hypertension and/or diabetes mellitus; body mass index (BMI) at diagnosis; number of pregnancies; use of hormone contraceptives; clinical stage at diagnosis (according to American Joint Committee on Cancer staging criteria)24; histologic type; Scarff-Bloom-Richardson (SBR) grade; lymphovascular invasion; and ER, PR, and HER2 status. Subsequently, disease progression, recurrence, and vital status also were obtained. The review and analyses of the clinical records were approved by the INCan Internal Review Board.

Morphology and Immunohistochemistry

We obtained specimens from the slide archive at the INCan. Tumor samples were examined and classified by a dedicated breast cancer pathologist who used the breast cancer classification proposed by the World Health Organization and the SBR grade.25 Regarding pathologic margins, considerable controversy exists about the definition of a negative pathologic margin. Traditionally, margins <1 mm are considered inadequate, and margins >10 mm are widely accepted as negative. In our department, adequate margins are defined as >3 mm. Surgical margins were assessed by an expert breast cancer pathologist.

We obtained hormone receptor status by immunohistochemical analyses of formalin-fixed, paraffin-embedded tissue sections from incisional biopsies that were used for diagnosis. We determined ER and PR expression levels by using immunoperoxidase staining (Dako, Glostrup, Denmark) and quantified them with image analysis (Biogenex, San Ramon, Calif); values <5% were categorized as negative. HER2 expression was determined by immunohistochemistry (Dako). To be characterized as HER2-positive for this study, tumors (primary or metastatic) were required to have either v-erb-b2 erythroblastic leukemia viral oncogene homolog 2 (ErbB2) gene amplification, as measured by fluorescence in situ hybridization (Vysis, Des Plaines, Ill), or ErbB2 protein overexpression, as measured by immunohistochemistry (3+) (HercepTest; Dako). One breast cancer pathologist re-evaluated all immunohistochemical results in each sample.

Statistical Analysis

For descriptive purposes, continuous variables were summarized as arithmetic means with standard deviations (SDs) and medians with ranges, and categorical variables were summarized as relative frequencies, proportions, and 95% confidence intervals (CIs). Inferential comparisons were performed with the Student t test or the Mann-Whitney U test according to distribution (normal or non-normal) determined by the Kolmogorov-Smirnov test. Pearson chi-square tests and odds ratios (ORs) were used to compare the clinical and pathologic variables of TNBC. For multivariate analysis, the variables included were those that had clinical significance and those that had significance or near statistical significance (P < .1) in univariate analysis.

Locoregional recurrence-free survival (LRFS) and disease-free survival (DFS) durations were measured from the date of breast surgery to the date of locoregional recurrence (LRR) and the date of locoregional or systemic recurrence, respectively, for patients with stage I, II, and III TNBC. CSS duration was defined as the period from the date of diagnosis to the date of death or last follow-up. LRFS, DFS, and CSS were analyzed with the Kaplan-Meier method, and comparisons among subgroups were performed with the log-rank test or the Breslow test.

All variables were dichotomized for survival analysis. When the median was not reached, we expressed survival as 3-year and 5-year rates and 95% CIs. Adjustment of potential confounders was carried out with Cox proportional hazards regression analysis. All tests were 2-sided, and significance was set at P < .05. SPSS software (version 17.0; SPSS, Chicago, Ill) was used for data analysis.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Patients and Samples

In total, 2074 Hispanic patients with breast cancer were included in this analysis. The median age (±SD) at diagnosis was 50 ± 12 years (range, 19-96 years). A family history of breast cancer was present in 20.4% of patients. Information regarding the number of births and the history of contraceptive use was obtained in 1647 and 1664 of 2074 patients, respectively. The mean number of births was 3.1 (range, 0-15 births), and 24.4% of patients had history of hormone contraceptive use. Ductal carcinoma was the most common histology (82.7%). Clinical stage was distributed as follows: 9.7% of patients had stage I disease, 34.6% of patients had stage II disease, 44% of patients had stage III disease, and 11.7% of patients had stage IV disease. The SBR grades low, intermediate, and high accounted for 12.4%, 36.9%, and 50.8% of patients, respectively. None of the patients who were included in this study had received previous hormone-replacement therapy.

Among the patients with stage I, II and III breast cancer who underwent breast-conserving surgery or mastectomy, positive/close margins (<3 mm) were obtained in 55 patients (0.3%). All patients underwent re-excision to achieve a confirmed negative pathologic margin. Sentinel lymph node mapping and resection were done for 97.5% of patients with clinically lymph node-negative (N0) breast cancer for surgical staging of the axilla. The patients who had axillary lymph node involvement identified on sentinel lymph node biopsy underwent complete axillary lymph node dissection. Sixty-three percent of patients received adjuvant radiotherapy according to established criteria: Patients who had T4 disease, a primary tumor >4 cm, >3 lymph nodes that were positive for metastasis, or 1 to 3 positive lymph nodes in the presence of additional risk factors (aged <35 years, lymphovascular invasion, and a low number of axillary lymph nodes examined). Of the 243 patients who had stage IV disease at diagnosis, 5.3% (13 patients) underwent surgery for local control.

The treatment of patients was determined by clinical stage. Patients with clinical stage IIB and stage III disease received neoadjuvant chemotherapy with anthracyclines and taxanes. After neoadjuvant chemotherapy, if there was significant residual disease (tumor ≥5 cm in greatest dimension, skin edema, ulceration, or skin involvement), then patients received preoperative radiotherapy. A pathologic complete response was achieved by 32.8% of all patients (41% of patients with HER2-positive disease, 29% of patients with TNBC, and 9% of patients with ER and/or PR [hormone receptor]-positive disease). Of the patients with hormone receptor-positive disease, 92.3% received adjuvant hormone treatment.

Patients with hormone-receptor-positive, HER2-positive, and TN tumors made up 56.5%, 20.4%, and 23.1% of patients, respectively. The initial estimated prevalence of TNBC was 28.3% according to the original assessment of breast cancer samples. After the re-evaluation of all immunohistochemical results by our dedicated breast cancer pathologist for the purposes of the current study, the prevalence of TNBC was 23.1%. For patients who were rediagnosed with hormone receptor-positive and/or HER2-positive breast cancer, a note was added to the patients' medical records to let the attending physician know the change in the tumor status for guidance of further treatment.

In univariate analysis, TNBC was associated with younger age (49.2 years vs 52.2 years; P < .001), premenopausal status (OR, 0.72; 95% CI, 0.58-0.88; P = .002), increased parity (OR, 1.32; 95% CI, 1.03-1.69; P = .029), hormone contraceptive use (OR, 1.20; 95% CI, 1.003-1.54; P = .04), high SBR grade (OR, 4.20; 95% CI, 2.50-7.01; P < .001), and advanced disease stage (stages III and IV vs stages I and II: OR, 1.60; 95% CI, 1.33-2.04; P < .001) (Table 1). In multivariate analysis, only premenopausal status (P = .01), the number of births (P = .014), SBR grade (P < .001), and stage (P = .002) were associated significantly with TNBC diagnosis. The frequency of TNBC according to clinical stage I, II, III, and IV at diagnosis was 9.2%, 20.6%, 27.3%, and 23.6%, respectively. TNBC prevalence was not associated significantly with overweight (BMI ≥25 kg/m2; P = .423) or obesity (BMI ≥30 kg/m2; P = .103). Nevertheless, in postmenopausal patients versus premenopausal patients, a BMI <25 kg/m2 (43.4% vs 35%; OR, 0.70; 95% CI, 0.51-0.96; P = .027) and a BMI <30 kg/m2 (78.8% vs 68.8%; OR, 0.54; 95% CI, 0.37-0.78; P < .001) were strongly associated with TNBC (Table 2). A family history of breast cancer was identified in 21% of patients with non-TNBC, which was not different from the 19.5% rate among patients with TNBC (P = .46).

Table 1. Clinical and Pathologic Characteristics of Patients With Nontriple-Negative and Triple-Negative Breast Cancer
CharacteristicPercentage of PatientsUnivariate AnalysisMultivariate Logistic Regression Analysis: P
Nontriple-Negative, n = 1596Triple-Negative, n = 469OR (95% CI)P
  1. OR indicates odds ratio; CI, confidence interval; SD, standard deviation.

Age: Median ± SD, y52.2 ± 12.349.2 ± 12.2 <.001 
Menopausal status
 Premenopausal49.157.40.72 (0.58-0.88).002.01
 Postmenopausal50.942.6   
Births
 <273.7681.32 (1.03-1.69).029.014
 ≥226.332   
Use of contraceptives
 No77.672.41.2 (1.003-1.54).04.079
 Yes19.627.6   
Multicentricity
 No91.994.40.76 (0.68-1.04).07 
 Yes8.15.6   
Histology
 Ductal81.885.3   
 Lobular10.37.20.94 (0.64-1.40).12 
 Other7.87.5   
Scarff-Bloom-Richardson grade
 Low14.94.1   
 Intermediate42.817.54.20 (2.50-7.01)<.001<.001
 High44.378.5   
Stage
 I11.33.7   
 II35.631.4   
 III41.352.71.60 (1.33-2.04)<.001.002
 IV11.612.2   
Lymphovascular invasion
 Yes33.436.10.88 (0.71-1.10).282 
 No66.663.9   
Table 2. Diagnosis of Arterial Hypertension, Diabetes Mellitus, Body Mass Index, Overweight, and Obesity According To Triple-Negative Status
Patient SubsetNontriple-Negative, %Triple-Negative, %Univariate P
All patientsan = 1595n = 479 
  • BMI indicates body mass index; SD, standard deviation.

  • a

    The sum of premenopausal and postmenopausal patients does not match the total sum of the values given because, in some patients, this information was lacking.

 Arterial hypertension
  No81.786.5.015
  Yes18.313.5 
 Diabetes mellitus
  No87.790.3.312
  Yes11.39.7 
 BMI: Mean ± SD, kg/m227.7 ± 5.127.5 ± 4.7.289
 BMI, kg/m2
  <2538.240.3.423
  ≥2561.859.7 
 BMI, kg/m2
  <3070.974.8.103
  ≥3029.125.2 
Premenopausal patientsan = 778n = 267 
 Arterial hypertension
  No92.794.4.337
  Yes7.35.6 
 Diabetes mellitus
  No95.295.1.944
  Yes4.84.9 
 BMI: Mean ± SD, kg/m227.3 ± 4.627.8 ± 4.9.280
 BMI, kg/m2
  <2541.638.249
  ≥2558.462 
 BMI, kg/m2
  <3075.171.8.295
  ≥3024.928.2 
Postmenopausal patientsan = 800n = 199 
 Arterial hypertension
  No7175.9.167
  Yes2924.1 
 Diabetes mellitus
  No82.483.9.606
  Yes17.616.1 
 BMI: Mean ± SD, kg/m228.2 ± 5.527.1 ± 4.8.008
 BMI, kg/m2
  <253543.4.027
  ≥2565.56.6 
 BMI, kg/m2
  <3068.878.8.001
  ≥3033.221.2 

Outcomes

The median follow-up was 17 months (range, 3-120 months). Of the 2074 patients who were included in this study, only 75 patients (3.6%) had a median follow-up <3 months. There were 122 LRRs (7.4%). The median time to local recurrence was 118.7 months (95% CI, 107-129 months). The 3-year LRFS rate was 92.2% (95% CI, 90.5%-93.9%). In multivariate analysis, variables that had a significant independent association with low LRFS were the presence of the TN phenotype (hazard ratio [HR], 2.4; 95% CI, 1.6-3.8; P < .001 (Fig. 1A-C), lymphovascular invasion (HR, 2.7; 95% CI, 1.5-4.9; P < .001), and advanced stage (stages III and IV vs stages I and II; HR, 1.95; 95% CI, 1.30-2.90; P < .001).

thumbnail image

Figure 1. These Kaplan-Meier plots illustrate (A) locoregional recurrence-free survival, (B) disease-free survival, and (C) cancer-specific survival for patients with triple-negative breast cancer (TNBC) and non-TNBC. HR indicates hazard ratio; CI, confidence interval.

Download figure to PowerPoint

There were 235 (14.6%) systemic recurrences. The frequency of TNBC among patients who had systemic recurrences was 34.4%, and the median DFS was 82 months (95% CI, 71.8-92.2 months). In multivariate analysis, factors that had a significant independent association with short DFS were the presence of the TN phenotype (HR, 1.62; 95% CI, 1.13-2.32; P = .009) (Fig. 1B), HER2-positive status (HR, 1.40; 95% CI, 1.04-1.93; P = .025), and advanced stage (stage III vs stages I and II; HR, 2.23; 95% CI, 1.69-2.95; P < .001).

There were 209 (10.2%) deaths. The 3-year survival rate was 84.70% (95% CI, 82.35%-7.05%), and the 5-year survival rate was 76.40% (95% CI, 72.48%-80.32%). In multivariate analysis, factors that were associated with CSS duration were presence of the TN phenotype (HR, 1.66; 95% CI, 1.20-2.30; P = .002) (Fig. 1C), HER2-positive status (HR, 1.30; 95% CI, 1.02-1.80; P = .031), and advanced stage (stage III vs stages I and II; HR, 2.1; 95% CI, 1.7-2.5; P < .001) (Table 3). SBR grade had a trend toward significance (HR, 1.3; 95% CI, 0.98-1.77; P = .062). Particularly in patients with TNBC, the sole prognostic factor associated with poor CSS was advanced stage (stage III vs stages I and II; P < .001) (Table 4).

Table 3. Cancer-Specific Survival Analysis in all Breast Cancer Patients
VariableFive-Year SurvivalMultivariate Analysis
Rate (95% CI), %PHR (95% CI)P
  1. CI indicates confidence interval; HR, hazard ratio; ER, estrogen receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2; BMI, body mass index.

Menopausal status
 Premenopausal77.60 (72.51-82.69)   
 Postmenopausal74.30 (67.25-81.35).412  
Histology
 Ductal77.30 (73.24-81.36)   
 Lobular71.30 (54.64-87.96)   
 Other71.6 (70.0-73.8).603  
Triple-negative subtype
 Yes57.1 (47.3-66.9)<.0011.66 (1.20-2.30).002
 No82.80 (79.76-85.84)   
ER status
 Positive87.50 (83.58-91.42)<.001  
 Negative64.50 (57.45-71.55)   
PR status
 Positive86.90 (82.83-91.64)   
 Negative68.30 (62.03-74.57)<.001  
HER2 status
 Positive69.10 (60.28-77.92)   
 Negative76.70 (71.41-81.99).0481.30 (1.02-1.80).031
Scarff-Bloom-Richardson grade
 I-II79.00 (71.16-86.84)   
 III75.50 (69.62-81.38).0011.30 (0.98-1.77).062
Lymphovascular invasion status
 No78.50 (70.86-86.14)   
 Yes75.7 (70.8-80.6).782  
Stage
 I94.20 (88.91-99.49)<.0012.1 (1.7-2.5)<.001
 II89.90 (84.81-94.99)   
 III73.80 (67.53-80.07)   
 IV33.80 (32.24-35.36)   
Arterial hypertension
 No77.20 (72.89-81.51).942  
 Yes73.70 (72.53-74.87)   
Diabetes mellitus
 No78.30 (73.99-82.61).259  
 Yes65.20 (50.90-79.50)   
BMI, kg/m2
 <2572.40 (65.35-79.45).189  
 ≥2580.3 (75.4-85.2)   
BMI, kg/m2
 <3076.1 (71.4-80.8).121  
 ≥3078.10 (69.09-87.11)   
Table 4. Cancer-Specific Survival Analysis in Patients With Triple-Negative Breast Cancer
VariableFive-Year Survival Rate (95% CI), %P
  1. CI indicates confidence interval; BMI, body mass index.

Menopausal status
 Premenopausal63.40 (51.45-75.35).453
 Postmenopausal42.10 (39.04-45.16) 
Histology
 Ductal59.90 (48.14-65.06) 
 Lobular72.20 (70.24-74.16).390
 Other21.60 (18.08-25.12) 
Scarff-Bloom-Richardson grade
 I-II68.30 (65.56-71.04).161
 III55.20 (54.03-56.37) 
Lymphovascular invasion
 No57.30 (45.54-69.06).948
 Yes50.00 (47.26-52.74) 
Stage
 I100<.001
 II84.3 (74.5-94.1) 
 III60.60 (46.88-74.32) 
 IV0 
Arterial hypertension
 No56.40 (44.64-68.16).985
 Yes64.70 (62.43-66.97) 
Diabetes mellitus
 No58.00 (46.24-69.76).58
 Yes39.60 (35.88-43.32) 
BMI, kg/m2
 <2546.30 (27.68-64.92).765
 ≥2568.90 (59.89-77.91) 
BMI, kg/m2
 <3056.70 (68.46-68.46).73
 ≥3062.50 (46.82-78.18) 

When we conducted a subgroup analysis according to breast cancer subtype (HR-positive/HER2-negative, HER2-positive, and TNBC), CSS was longer in patients with HR-positive/HER2-negative disease compared with the other 2 groups (P = .002) (Fig. 2). Only 10.9% of the whole population received treatment with trastuzumab, which corresponded to 51.2% of the patients with HER2-positive disease. This was because it was not until 2006 that the government covered for trastuzumab therapy. We also compared CSS in the HER2-positive patients who and did not receive trastuzumab treatment. CSS was significantly longer in HER2-positive patients who received trastuzumab (CSS: 91.6% [95% CI, 87.6%-95.5%] vs 75% [95% CI, 63.8%-87.2%]; P < .001) (Fig. 3). All data regarding the treatment of diabetic patients were analyzed according to treatment, and we did not observe any statistically differences between metformin and other kinds of treatments in terms of LRFS, DFS, or CSS (insulin, oral hypoglycemic drugs).

thumbnail image

Figure 2. Cancer-specific survival is illustrated according to breast cancer subtypes (hormone [estrogen and/or progesterone] receptor [HR]-positive/human epidermal growth factor receptor 2 [HER2]-negative [HR+/HER2−], HR-negative/HER2-positive [HR−/HER2+], and triple-negative breast cancer [TNBC]).

Download figure to PowerPoint

thumbnail image

Figure 3. Cancer-specific survival is illustrated for patients with human epidermal growth factor receptor 2-positive breast cancer who did or did not receive trastuzumab.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

To our knowledge, this is the first study to determine the prevalence of TNBC and its risk factors among Hispanic patients with breast cancer. In our study population, the mean age at diagnosis of breast cancer was 11 years younger than the average age reported in the United States. TNBC prevalence in our study population was higher than that reported in white patients with breast cancer. We identified differences in hormone expression profiles of breast cancer according to BMI, menopausal status, and parity. TNBC was associated with a higher risk of LRR and with lower DFS and CSS.

Increasing evidence supports the possibility that it is biologically plausible for specific subtypes of breast cancer to have different etiologies. These findings are based on the current understanding of the heterogeneity of breast cancer development because of hormone and growth factor receptors status.21 Several studies have demonstrated clear differences in clinical, pathologic, and prognostic characteristics between African-American patients and white patients.21, 26-28 Indeed, gene expression analyses of several breast cancer subtypes have demonstrated differences between African-American patients and non-African-American patients, with a higher prevalence of basal-like breast tumors and a lower prevalence of luminal tumors in the African-American patients.29 These factors may contribute to the worse prognosis for African-American women with breast cancer.29

In our study population, after re-evaluating immunohistochemistry results from each sample, we observed that the prevalence of TNBC was 23.1% in Hispanic patients at the INCan. This prevalence was higher than that reported in white patients (range, 10%-13%) but was closer to the prevalence reported in African-American patients (range, 23%-30%).14, 30 Few studies have been performed on Hispanics living in the United States. In a small study of 46 patients, the prevalence of TNBC was 10%,14 although this low prevalence may not be representative because of the small sample size. In a subsequent study of 255 patients with TNBC, 49 patients (19.2%) were Hispanic Americans.31 The prevalence of TNBC varies according to ethnicity and has been reported as high as 82% in Ghana,32 39% in Saudi Arabia,33 19.3% in China,32 and 15.9% in Taiwan.34

The distribution according to clinical disease stage in our study population (localized disease, 44.3%; regional disease, 44%; and distant disease, 11.7%) was similar to that reported in African-American patients (51%, 37%, and 10%, respectively)1 and Hispanic-American patients (54%, 37%, and 7%, respectively) with breast cancer.35 All of these distributions differ significantly from those reported in white patients with breast cancer (63%, 29%, and 5%, respectively).1, 35 Together, these findings suggest a proportional relation between clinical stage and TNBC prevalence, leading to a worse prognosis for African-American and Hispanic-American patients compared with that for white patients with breast cancer. Loss of hormone receptor and HER2 expression has been described as a consequence of tumor progression.36-39

We also observed that TNBC was associated with high histologic grade, as reported previously.20 Another similar feature between African-American patients with breast cancer and our study population is the frequency of high histopathologic grade at diagnosis (48% in African-American patients and 51% in our study population), which is higher than that reported in white patients with breast cancer (32%).14 Several studies have associated high histologic grade with loss of hormone receptor and HER2 expression.40-42 Advanced disease and high grade at diagnosis in both African-American patients with breast cancer and in our study population may explain the high prevalence of TNBC. In addition, lack of health education, deficient of awareness, and suboptimal access to high-quality mammography and prompt referral might also justify the high rates of advanced disease in Mexican patients.

Overweight and obesity have been associated with a high risk of developing breast cancer.43 Obesity is related to a poor prognosis, possibly secondary to estrogenic activity, advanced or aggressive disease at diagnosis, and a high likelihood of both local and systemic treatment failure.43-45 We did not observe a correlation between overweight or obesity and a diagnosis of TNBC when we considered all patients; however, we did observe a significantly higher prevalence of TNBC in postmenopausal patients who had a BMI <25 kg/m2 or a BMI <30 kg/m2.

Hormone receptor status has not been linked directly with overweight or obesity. However, several studies have yielded conflicting results regarding the association between BMI, menopausal status, and hormone receptor expression. Similar to our findings, a prospective cohort of >90,000 postmenopausal patients with breast cancer showed that a higher BMI was associated with advanced disease and with higher expression of hormone receptors.46 Conversely, in a study that was conducted in breast cancer patients between the ages of 20 and 50 years, obesity was associated with TNBC (HR, 1.89; 95% CI, 1.22-2.92).21 TNBC patients had a higher incidence of hypertension, diabetes, overweight, and obesity compared with their non-TNBC counterparts (58.1% vs 36.7%, respectively). However, BMI was not associated independently with TNBC, consistent with our findings when we analyzed the entire breast cancer population (a subanalysis in premenopausal and postmenopausal breast cancer patients was not carried out by the authors).20 Another study demonstrated that the association between ER/PR and HER2 expression and BMI varies according to menopause status.47 In premenopausal women, BMI was associated inversely with hormone receptor expression; but, in postmenopausal women, BMI had a positive association with hormone receptor and HER2 levels. They hypothesized that low serum hormone levels in overweight or obese, premenopausal women cause up-regulation of ER and PR levels in normal breast epithelium, leading to an exaggerated hormone response after menopause.47

In our study population, the median age at diagnosis of breast cancer was 11 years younger than that reported in US patients (50 years vs 61 years).48 Our study supports the finding of Rodriguez-Cuevas et al., who reported that the average age at diagnosis of breast cancer was 51 years in Mexican patients.49 Another study performed in the United States indicated that there were differences in age at presentation of breast cancer: 47% of Hispanic-American patients were aged <50 years at diagnosis versus 25% of white patients.14 These observations have great relevance for evaluating the best screening model for women at risk of breast cancer in Mexico.50

We observed that there was an association of TNBC with poor CSS and an increased risk of LRR. Poor prognosis was evidenced by low DFS and CSS in HER2-positive patients. This might be related to the fact that only 50% of all patients with HER-2 positive breast cancer received treatment with trastuzumab, because it was not until 2006 that the government paid for this treatment. We believe that, if a higher proportion of HER2-positive patients received trastuzumab as indicated, then CSS for the non-TNBC subgroup would have been longer, and more significant differences in CSS would have been observed between the patients with and without TNBC. Studies have reported similar results for CSS and a high LRR rate among patients with TNBC.51-55

Although there are no population-based cancer registries in Mexico, there are some estimates of breast cancer incidence through mortality-specific analyses. On the basis of these findings, the estimated incidence of breast cancer in our country is of 27.2 per 100,000 with a mortality rate of 10.1 per 100,000.56

Several hormone/reproductive/lifestyle risk factors can affect the incidence rates of TNBC between populations, and these may be independent of race/ethnicity and may follow a basic age-related curve. We observed a significant association between TNBC diagnosis and more than 1 birth. The increased number of births in Mexican patients compared with Caucasian-American women may explain the higher incidence of TNBC. The total fertility rate in the United States is 2.06 per women,57 in contrast to our population, who have a mean number of 3.1 births. Millikan et al. and the Carolina Breast Cancer Study Group reported data indicating that reproductive history/parity is a major explanatory factor for race/ethnicity-related differences in the frequency of TNBC.58

In the current univariate analysis, but not in the multivariate analysis, we observed an association between hormone contraceptive use and TNBC diagnosis. In previous studies, oral contraceptive use for more than 1 year was associated with a 2.5-fold increase in the incidence of TNBC.59 It has been proposed that the mechanism through which oral contraceptives use impacts the risk of breast cancer among young women is that estrogen promotes the growth of breast cancer-enhancing angiogenesis and stromal cell recruitment.60

In conclusion, in our study population, the median age at diagnosis of breast cancer was 11 years younger than the average age reported in the United States and in other, similar patient populations. Our results suggest that, in the Hispanic population, screening programs and recommendations need to be analyzed carefully and evaluated from the perspective of cost-effectiveness in middle-income countries like Mexico. TNBC prevalence in our study population was higher than that reported in white patients with breast cancer. This may be explained by racial differences, a more advanced stage at diagnosis, or a higher SBR grade, all of which can cause a change in hormone receptor status. We identified differences in hormone expression profiles in breast cancer according to BMI, menopausal status, and parity, suggesting a heterogeneous etiology. And, as reported in other studies, TNBC was associated a with poor prognosis, a high risk of LRR, and poor DFS and CSS.

CONFLICT OF INTEREST DISCLOSURES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

This research is supported in part by the Universal Health Insurance Program of the Mexican Government for Breast Cancer.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES