- Email: [email protected]
Breast Cancer Subtype Variation by Race and Ethnicity in a Diverse Population in British Columbia
Original Study
Breast Cancer Subtype Variation by Race and Ethnicity in a Diverse Population in British Columbia Dante Wan,1 Diego Villa,1 Ryan Woods,2 Rinat Yerushalmi,1,3 Karen Gelmon1 Abstract The distribution of breast cancer subtypes varied significantly by race/ethnicity within our cross-sectional cohort from a multicultural population. Our data suggest that race/ethnicity plays a significant role in the biology of invasive breast cancer and that certain groups might experience different outcomes. These results have important implications for resource allocation and clinical care in multiethnic settings with universal healthcare. Background: Breast cancer subtypes occur differentially across different racial and ethnic groups. However, their distribution within a multicultural population is unknown. Materials and Methods: Patients with invasive breast cancer newly diagnosed in 2006 and referred to the British Columbia Cancer Agency were identified from the Breast Cancer Outcomes Unit database. Race/ethnicity data were abstracted from a patient-completed health assessment questionnaire completed at the initial consultation, and grouped as white, East Asian, Aboriginal, South Asian, Southeast Asian, and other. Breast cancer subtypes were created using available data on estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) status. Results: A total of 1829 women had complete data. Of these women, 73% were white, 8% were East Asian, 4% Aboriginal, 3% South Asian, 3% Southeast Asian, and 3% other. The median age at diagnosis was 60 years; the youngest were Southeast Asian (51 years) and the oldest were white (60 years; P < .001). The overall ER positivity rate was 81%, highest in East Asian women (89%) and lowest in South Asian women (73%). The HER2 positivity rate was 16% for all patients and was highest in the South Asian (28%), Southeast Asian (28%), and Aboriginal (24%) women and lowest in the white women (14%; P < .001). Triple-negative (ER-, PR-, and HER2-negative) breast cancer was uncommon in East Asian women (5%) but more common in South Asian women (19%; P < .001). The 5-year breast cancer-specific survival was 90% (95% confidence interval, 89%-92%), with no significant difference among the racial/ethnic groups (P ¼ .136). Conclusion: Breast cancer subtypes varied by race/ethnicity in our cross-sectional cohort of a multicultural population, suggesting that race/ethnicity plays a significant role in the biology of invasive breast cancer. Clinical Breast Cancer, Vol. -, No. -, --- ª 2015 Elsevier Inc. All rights reserved. Keywords: Aboriginal, Asian, Breast cancer subtypes, Estrogen receptor, HER2
Introduction Since the initial description of breast cancer subtypes by Perou et al1 in 2000, our concept of breast cancer has evolved to a better understanding of the unique biology, response, and prognosis of the specific 1 Division of Medical Oncology, British Columbia Cancer Agency, Vancouver, BC, Canada 2 Cancer Control Research, British Columbia Cancer Agency, Vancouver, BC, Canada 3 Division of Medical Oncology, Davidoff Center, Petah Tikvah, Israel
Submitted: Jun 1, 2015; Revised: Aug 31, 2015; Accepted: Sep 11, 2015 Address for correspondence: Diego Villa, MD, MPH, British Columbia Cancer Agency, 600 West 10th Avenue, Vancouver, BC, Canada E-mail contact: [email protected]
1526-8209/$ - see frontmatter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.clbc.2015.09.005
cancer subtypes. Various reports have suggested that the distribution of breast cancer subtypes varies by race/ethnicity. For example, epidemiologic data have suggested an increased prevalence of human epidermal growth factor receptor 2 (HER2)-positive breast cancer in Asians living in the United States.2,3 Specifically, HER2-positive breast cancer appears to be overrepresented in Chinese, Korean, Filipino, and Vietnamese women compared with non-Hispanic white women.4 Population-based studies have also suggested a greater prevalence of basal-like cancer in African5 and African-American women6-8 and in a number of other countries, including Brazil.9 Simultaneously, race/ethnicity has been associated with different breast cancer outcomes in retrospective and prospective studies that have not evaluated breast cancer subtypes. Compared with white
Clinical Breast Cancer Month 2015
-1
Distribution of Breast Cancer Subtypes by Race/Ethnicity women, African-American women present with a more advanced stage at diagnosis,10,11 higher grade tumors,12 and worse survival after controlling for prognostic factors.13,14 The age-adjusted breastcancer specific mortality rate in the United States in 2010 for white women was 21.3 per 100,000 compared with 30.2 per 100,000 in black women.15 However, the differences in the clinical outcomes among racial/ethnic groups are likely multifactorial, including socioeconomic factors, access to screening programs and treatment, inherent ethnic biology, and variable distributions of breast cancer subtypes. British Columbia (BC) is a multiethnic province with a predominantly white population and a significant proportion of East Asian, Aboriginal, South Asian, and Southeast Asian individuals. Residents of BC have access to universal, publicly funded healthcare, including breast cancer screening, diagnosis, and treatment. The objective of the present study was to assess the distribution of breast cancer subtypes by race/ethnicity in our population and evaluate whether race/ethnicity is associated with different breast cancer outcomes.
Materials and Methods Study Population All patients diagnosed with invasive breast cancer from January 1, 2006 to December 31, 2006 were identified in the Breast Cancer Outcomes Unit (BCOU) database. The year 2006 was chosen for 3 reasons. First, 2006 was the first year when HER2 testing was consistently performed on all new breast cancer biopsy specimens at our institution. Second, that year allowed a sufficient follow-up length to examine the 5-year survival outcomes. Finally, the race/ethnicity data were extracted manually from the paper medical records, posing incremental logistical and resource challenges if additional years were included. The BCOU database is a population-based registry that contains prospectively collected demographic, pathologic, staging, and initial treatment data for patients referred to the BC Cancer Agency (BCCA) for breast cancer management. The outcomes for all patients were obtained by an active letter of follow-up to the general practitioners and through periodic direct linkage to the provincial vital statistics agency. The outcomes information collected included first local, regional, and distant relapse and, if applicable, the date and cause of death.
Baseline Characteristics and Breast Cancer Subtypes
2
-
Variables, including breast cancer subtype, patient age, tumor grade, and tumor stage, were obtained from the BCOU database. The patients’ height and weight, measured at the initial consultation, were abstracted from the medical records to calculate the body mass index (BMI) for each patient. The breast cancer subtype was assigned using standard immunohistochemistry (IHC) to determine the estrogen receptor (ER), progesterone receptor (PR), and HER2 status on core biopsy or excision specimens. HER2 was considered positive if the IHC score was 3þ or if the HER2-to-chromosome 17 centromere ratio was 2 using fluorescent in situ hybridization (FISH). FISH was only performed in cases with indeterminate or inconclusive HER2 expression by IHC. For the present study, the combinations of IHC markers and FISH data were used as surrogates for the following
Clinical Breast Cancer Month 2015
intrinsic breast cancer subtypes: luminal A or B (ERþ or PRþ, HER2), HER2 enriched (ERþ or ER or PRþ or PR, and HER2þ), and basal-like (ER, PR, and HER2). The distinction between luminal A and B subtypes was not made using the available IHC results.
Race and Ethnicity At the initial consultation at a BCCA center, all patients completed the Health Assessment Form (HAF). The HAF explores patients’ medical and sociodemographic background, including selfreported race/ethnicity through the question “To which ethnic or cultural group do you belong?” Each patient was asked to self-assign to 1 of the following major groups: Aboriginal (eg, North American First Nations, Inuit, Metis), black (eg, African, Haitian), white or European, Latin American, or Asian. The Asian group was further subdivided geographically into West Asian (Arab), East Asian (Chinese, Japanese, Korean), South Asian (East Indian, Pakistani), and Southeast Asian (Indonesian, Laotian, Filipino). Patients were assigned an “other” ethnicity if they listed 2 ethnicities (mixed heritage) or if they self-identified with a category not listed (eg, Pacific Islander). Patients with incomplete data were excluded. For the present analysis, patients were grouped into the following 6 categories: white, East Asian, Aboriginal, South Asian, Southeast Asian, and other.
Statistical Analysis The main study endpoint was the distribution of breast cancer subtypes according to race/ethnicity, reported as frequencies. The proportion of ERþ and HER2þ cases according to breast cancer subtype was also calculated. Categorical variables were compared among the ethnic groups using the c2 test and continuous variables using the Kruskal-Wallis test. Because this was a retrospective, population-based study, no formal sample size calculations were performed. As a secondary endpoint, breast cancer-specific survival (BCSS) was calculated as the time from the date of diagnosis to death from a primary or secondary cause related to breast cancer. Patients who were alive at the last follow-up visit or who had died of nonebreast cancer causes were censored at the last follow-up visit. Overall survival (OS) was calculated from the date of the initial diagnosis to the date of the last follow-up visit or death from any cause. BCSS and OS were estimated using the Kaplan-Meier method and compared among groups using the log-rank test. The 5-year BCSS and OS rates are presented by ethnic group and breast cancer subtype with the 95% confidence intervals (CIs). Cox proportional hazards regression models were used to examine the relationship between race/ethnicity and survival outcomes, adjusted for breast cancer subtype, stage, and grade and patient factors, including age, smoking history, and BMI. Treatment information was not included in the Cox regression model, because it was not available. The proportional hazards assumption was examined using Schoenfeld residual plots, and the sensitivity of the results to departures from this assumption was assessed using models stratified by any terms of concern. All data analysis was performed using SAS, version 9.3, and the R statistical language, version 2.15. The University of BC/BCCA research ethics board approved the present study.
Dante Wan et al Results Patient Population A total of 2222 patients with invasive breast cancer newly diagnosed during 2006 were identified in the BCOU database. The following 393 patients (18%) were excluded (Figure 1): 150 without available HAF data, 135 with missing self-reported race/ethnicity on the HAF, and 108 with the information on ER, PR, and HER2 status required to assign a breast cancer subtype missing. The final analysis included 1829 patients. The distribution of race/ethnicity in descending order was white (n ¼ 1435; 73%), East Asian (n ¼ 155; 8%), Aboriginal (n ¼ 71; 4%), South Asian (n ¼ 65; 3%), and Southeast Asian (n ¼ 51; 3%). A total of 52 patients (3%) had a selfreported race/ethnicity that was either “other” or an uncommon choice with too few occurrences to be analyzed as a separate group (eg, African American; n ¼ 5).
Clinical and Demographic Variables The patient characteristics were unevenly distributed across the groups (Table 1). The median age at diagnosis was 58 years (interquartile range, 49-69); however, the age varied considerably by ethnicity, with white women presenting at a typically older age and Southeast Asian and other patients presenting with a lower median age at diagnosis. The proportion of patients aged < 40 years was not markedly different across the groups; these proportions were determined from very small counts for most groups (n < 10 for all groups except white). A greater proportion of Southeast Asian
Figure 1 Patient Identification and Exclusions
2222 pts with IBC idenƟfied in BCOU database 150 charts either unretrievable or lacking HAF forms 2072 medical charts with HAF forms for review 135 forms with missing race/ ethnicity data 1937 HAF forms with complete race/ ethnicity data
women were premenopausal (63%), more than twice the proportion for white women (29%). No significant difference was found in the stage distribution among the groups, although South Asian patients had a greater proportion of high-grade disease. Aboriginal women had the greatest rates of smoking and elevated BMI. East Asian women had the lowest BMI. South Asian patients had the lowest current and ever smoking rates.
Subtype Distribution by Ethnicity The ERþ/HER2 subtype was the most common breast cancer subtype (n ¼ 1320; 72%), followed by the HERþ (n ¼ 287; 16%) and ER/HER2 subtypes (n ¼ 222; 12%; Table 2). The white and East Asian patients had the greatest proportions of ERþ/HER2 breast cancer, and the Southeast Asian, South Asian, and Aboriginal patients had the greatest proportions of HER2þ breast cancer. South Asian and Aboriginal patients had a greater proportion of ER/HER2 disease, but this breast cancer subtype was very uncommon in East Asian patients. Overall, white and East Asian women had the most favorable subtype distribution, with a relatively greater proportion of ERþ/HER2 tumors and lower HER2þ and ER/HER2þ tumors. In contrast, South Asian women had relatively greater frequencies of the less favorable breast cancer subtype.
Survival Outcomes With a median follow-up of 5.5 years (range, 5-6) for the living patients, the 5-year BCSS for the entire cohort was 90% (95% CI, 89-92), with no significant difference among groups (P ¼ .136; Figure 2). The 5-year OS for the entire cohort was 87% (95% CI, 85-88), with a significant difference among groups likely related to the difference in age at diagnosis (P ¼ .030; Figure 3). The 5-year BCSS and OS rates for the cohort stratified by both race/ethnicity and breast cancer subtype are listed in Table 3. Cox proportional hazards regression analysis for the BCSS endpoint showed no significant difference in outcomes across the groups after adjustment (likelihood ratio test [LRT], P ¼ .131). This result was also consistent across the sensitivity analyses that stratified models by key prognostic variables. Models for the OS endpoint demonstrated a persistent significant difference in outcome across groups (LRT, P ¼ .053) after adjustment for other prognostic variables. This difference was driven largely by differences between the white and East Asian groups (hazard ratio, 0.40; 95% CI, 0.18-0.88; P ¼ .022). This result was consistent in sensitivity analyses that stratified models by key prognostic factors.
Discussion 108 pts with missing subtype informaƟon
1829 pts with known race/ethnicity and subtype data for final analysis
Abbreviations: BCOU ¼ Breast Cancer Outcomes Unit; HAF ¼ Health Assessment Form; IBC ¼ Invasive Breast Cancer; pts ¼ patients.
In our specific population, the distribution of breast cancer subtypes varied significantly by race/ethnicity. Consistent with previous studies, the ERþ/HER2 subtype was the most common subtype across all ethnicities, with greater frequencies in white and East Asian women (w70%) and lowest in Southeast Asian and South Asian women (w60%). These results suggest that race/ ethnicity plays a significant role in the biology of invasive breast cancer. To our knowledge, ours is the first study examining breast cancer subtypes in Aboriginal women (n ¼ 71). Consistent with findings from the Surveillance, Epidemiology, and End Results program, we observed that South Asian women had a greater incidence of ER/PR breast cancers than did white
Clinical Breast Cancer Month 2015
-3
Distribution of Breast Cancer Subtypes by Race/Ethnicity Table 1 Patient Characteristics at Diagnosis According to Ethnicity Race/Ethnicity Variable
White (n [ 1435)
East Asian (n [ 155)
Aboriginal (n [ 71)
South Asian (n [ 65)
SE Asian (n [ 51)
Other (n [ 52)
60.0
54.0
55.0
54.0
50.0
49.0
50.0-70.0
47.0-68.0
47.0-65.0
47.0-64.0
44.0-61.0
43.5-64.5
<.0001
Age at diagnosis (years) Median IQR
<.0001
Age group (years) <40
68 (4.7)
8 (5.2)
6 (8.5)
4 (6.2)
4 (7.8)
7 (13.5)
40-59
647 (45.1)
91 (58.7)
39 (54.9)
36 (55.4)
34 (66.7)
29 (55.8)
60
720 (50.2)
56 (36.1)
26 (36.6)
25 (38.5)
13 (25.5)
16 (30.8) <.0001
Menopausal status Premenopausal
420 (29.3)
66 (42.6)
27 (38)
27 (41.5)
32 (62.7)
27 (51.9)
Postmenopausal
994 (69.3)
89 (57.4)
44 (62)
38 (58.5)
19 (37.3)
25 (48.1)
Unknown
21 (1.5)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
Stage at diagnosis
.2267
I
590 (41.1)
65 (41.9)
24 (33.8)
15 (23.1)
20 (39.2)
21 (40.4)
II
563 (39.3)
58 (37.4)
29 (40.8)
37 (56.9)
22 (43.1)
21 (40.4)
III
219 (15.3)
28 (18.1)
15 (21.1)
11 (16.9)
7 (13.7)
8 (15.4)
IV
59 (4.1)
3 (1.9)
3 (4.2)
1 (1.5)
1 (2)
2 (3.8)
3 (0.2)
1 (0.6)
0 (0)
1 (1.5)
1 (2)
0 (0)
1
357 (24.9)
23 (14.8)
8 (11.3)
10 (15.4)
6 (11.8)
11 (21.2)
2
553 (38.5)
70 (45.2)
30 (42.3)
14 (21.5)
17 (33.3)
19 (36.5)
3
496 (34.6)
59 (38.1)
32 (45.1)
39 (60)
28 (54.9)
20 (38.5)
3 (1.9)
1 (1.4)
2 (3.1)
85 (54.8)
33 (46.5)
31 (47.7)
30 (58.8)
30 (57.7)
Unknown
<.0001
Grade
Unknown
29 (2)
0 (0)
2 (3.8)
Nodal status Negative
.4182 795 (55.4)
Positive
478 (33.3)
60 (38.7)
26 (36.6)
29 (44.6)
16 (31.4)
15 (28.8)
Unknown
162 (11.3)
10 (6.5)
12 (16.9)
5 (7.7)
5 (9.8)
7 (13.5)
Current
175 (12.2)
2 (1.3)
16 (22.5)
2 (3.1)
0 (0)
4 (7.7)
Former
646 (45)
13 (8.4)
34 (47.9)
2 (3.1)
10 (19.6)
13 (25)
Never
600 (41.8)
138 (89)
20 (28.2)
61 (93.8)
39 (76.5)
34 (65.4)
2 (3.9)
1 (1.9)
<.0001
Smoking status
Unknown
14 (1)
2 (1.3)
1 (1.4)
0 (0)
26.6
23.2
28.9
27.1
25.6
26.5
23.4-30.7
21.3-25.4
25.5-31.6
24.5-29.0
22.8-28.8
23.8-30.9
<.0001
BMI Median IQR
P Value
Data presented as n (%), unless noted otherwise. Abbreviations: BMI ¼ body mass index; IQR ¼ interquartile range; SE ¼ Southeast.
4
-
women.16 The rate of HER2þ cancer among South Asian, Southeast Asian, and Aboriginal women was significantly greater (w25%) but was much lower in white women. These findings are concordant with data from the California Cancer registry, in which Asian (Korean, Filipino, Vietnamese, and Chinese) women had greater rates of HER2þ breast cancer compared with non-Hispanic white women.4 Our results have important implications for resource allocation and clinical care in multiethnic settings with universal healthcare. Although we did not specifically examine the trend of breast cancer subtype distribution by race/ethnicity over time, it is likely that with the ever-increasing proportions of visible immigrants, a gradual
Clinical Breast Cancer Month 2015
increase in the cost associated with the treatment of certain breast cancer subtypes, especially the HER2þ subtype, would occur. This is particularly relevant in regions with a high rate of immigration. It is also important in terms of health education and the role of prevention and screening. Understanding patient age at presentation and the subtype prevalence could aid in targeting immigrant groups with rational educational materials. Approximately 1,162,900 foreign-born people arrived in Canada between 2006 and 2011.17 These recent immigrants accounted for 17% of the total foreign-born population in Canada and 3.5% of the total population. Of the Group 8 countries, in 2010, Canada
Dante Wan et al Table 2 Breast Cancer Subtype According to Ethnicity Race/Ethnicity Cancer Subtype
White
East Asian
Aboriginal
South Asian
SE Asian
P Value
Other
Hormone status
.0003
ERþ or PRþ, HER2
1058 (73.7)
120 (77.4)
45 (63.4)
35 (53.8)
31 (60.8)
31 (59.6)
ERþ or PRþ, HER2þ
130 (9.1)
19 (12.3)
12 (16.9)
12 (18.5)
10 (19.6)
11 (21.2)
ER, PR, HER2þ
67 (4.7)
9 (5.8)
5 (7)
6 (9.2)
4 (7.8)
2 (3.8)
ER, PR, HER2
180 (12.5)
7 (4.5)
9 (12.7)
12 (18.5)
6 (11.8)
8 (15.4)
1188 (82.8)
139 (89.7)
57 (80.3)
47 (72.3)
41 (80.4)
42 (80.8)
.0548
197 (13.7)
28 (18.1)
17 (23.9)
18 (27.7)
14 (27.5)
13 (25)
.0002
ERþ and/or HERþ ERþ HER2þ
Data presented as n (%). Abbreviations: ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; PR ¼ progesterone receptor; SE ¼ Southeast.
higher BMI, which could affect the outcomes.18,19 Further studies of Aboriginal women should add more information. Although we included a diverse population, this particular composition might not necessarily mirror that of other areas within North America or the rest of the world, limiting its generalizability. Our population was multiethnic but had a predominance of persons of either European or Asian origin. We did not consider finer subgroups of Europeans nor did we have a sizable population with African roots. Also, 243 of 2222 women (11%) were excluded from the analysis because of an unknown race/ethnicity or breast cancer subtype, as were 52 of 1829 women (3%) with self-reported “other” race/ ethnicity. Despite the large sample size, most of our defined race/ ethnicity categories contained a small number of patients; thus, our observed breast cancer subtype distributions within these groups were imprecise from a statistical viewpoint. Just as with a number of other studies, the use of IHC receptor profiles served only as a surrogate for the molecular breast cancer subtype as determined by gene expression profiling.
0.9 0.7
0.8
Caucasian East Asian Aboriginal South Asian SE Asian Other
P=0.136
0.6
Breast Cancer-Specific Survival
1.0
Figure 2 Breast Cancer-Specific Survival According to Ethnicity
0.5
had the greatest proportion of foreign-born peoples (21%), followed by Germany (13%) and the United States (13%). In the 2011 Canadian census, nearly 6,264,800 people identified themselves as a member of the visible minority population, accounting for 19% of the country’s total population, an increase from 16% in the 2006 census. The 3 largest visible minority groups—South Asians, Chinese, and blacks—accounted for 61% of the visible minority population.17 Knowledge of the racial/ethnic differences in cancer biology could, therefore, be potentially helpful in planning healthcare resources, including screening and treatment. Also, economic planning might be facilitated by knowing the rates of different subtypes in the population, especially because costly targeted therapies are determined by subtype. Many reports have suggested that a differential distribution of breast cancer subtypes across geographic areas or within specific races/ ethnicities is present. However, these studies were limited by nonuniversal access to healthcare, which might have led to the underrepresentation of large subgroups of the general population. Our study was performed in a Canadian province with universal healthcare and a centralized cancer system, reducing this potential bias. Although the referral rate for newly diagnosed invasive breast cancer to the BCCA is not absolute, our study captured 85% to 90% of all invasive breast cancer cases in BC. However, residual confounding secondary to referral bias would be unlikely to alter the study findings. The present study provided a sufficient sample size and follow-up length to evaluate the 5-year outcomes, especially in the curative setting. This is the first study to evaluate the distribution of breast cancer subtypes in a diverse population that includes North American Aboriginal women. Additionally, our system, with universal access and screening mammography program, might provide a more comprehensive patient population than could similar reports. Despite these strengths, visible minorities have continued to be underrepresented in most breast cancer clinical trials. Further understanding of the biology, response to treatments, and toxicity profile stratified by race/ethnicity might better inform the optimal management strategies for these particular subpopulations. Our results have shown that Aboriginal women present with the highest median BMI and are more likely to be current or recent smokers at diagnosis. This is of interest, in particular, because other recent studies have suggested that other races/ethnicities, specifically African-American women, might have a
0
1
2
3
4
5
Years since diagnosis
Abbreviation: SE ¼ Southeast.
Clinical Breast Cancer Month 2015
-5
Distribution of Breast Cancer Subtypes by Race/Ethnicity We chose to use self-reported race/ethnicity as the exposure of interest, realizing that no reference standard exists for defining a particular individual. Although the HAF explicitly queried for selfreported ethnicity, we realize that in multicultural settings, some people might identify with > 1 racial or ethnic category and that others might not be able to make a clear distinction between their own race and ethnicity. Although self-identified ethnicity might not be consistent with “biologic ethnicity,” “genetic ethnicity,” or race, it is possible that many patients used these terms interchangeably when completing the HAF. Because of the overlap between race and ethnicity, and because the distinction between Hispanic and nonHispanic ethnicity is not relevant in Canada, we chose to use the term “race/ethnicity,” consistent with other recent, large cohorts evaluating a similar research question.21,22
0.7
0.8
Caucasian East Asian Aboriginal South Asian SE Asian Other
P=0.030
0.5
0.6
Overall Survival
0.9
1.0
Figure 3 Overall Survival According to Ethnicity
0
1
2
3
4
5
Years since diagnosis
Abbreviation: SE ¼ Southeast.
Although different racial/ethnic groups might present with different biologic cancer factors, other factors might affect the outcome. Recent data have suggested that body composition and obesity vary across race/ethnicity, which could, in turn, affect the outcomes, independently of breast cancer subtype distributions.20 In our cohort, we found that our Aboriginal cohort presented with the highest median BMI and were more likely to be current or recent smokers at diagnosis. How these factors might affect outcomes requires study. Also, owing to the very small number of study subjects in some racial/ethnic subsets, overfitting of the statistical model and the generalizability of this research are potential limitations of our research. Furthermore, residual confounding might be present because treatment, socioeconomic status, and comorbidity data were not collected.
Table 3 Five-Year Breast Cancer-Specific and Overall Survival According to Ethnicity and Breast Cancer Subtype Variable
5-year BCSS (%)
Clinical Practice Points Breast cancer subtypes occur differentially across different racial/
ethnic groups. The ERþ/HER2 was the most common subtype, with a
White
89.5 (87.9-91.1)
85.5 (83.7-87.3)
East Asian
96.1 (93.1-99.2)
94.2 (90.6-97.9)
Aboriginal
92.8 (86.9-99.1)
88.7 (81.7-96.4)
South Asian
90.4 (83.3-98.0)
84.6 (76.3-93.9)
SE Asian
94.1 (87.9-100.0)
94.1 (87.9-100.0)
Other
92.3 (85.3-99.8)
90.4 (82.7-98.8)
Breast cancer subtype
-
Although the ethnicity subsets were small, we found differences in the rates of ER and HER2 positivity by ethnicity, with ER disease more frequent in Aboriginal and South Asian women. These groups and Southeast Asians also had a greater incidence of HER2þ disease. These results are provocative and suggest further studies and validation on larger populations are necessary. Studies that also include genomic analysis on both blood and tumor samples might provide more robust epidemiologic information.
5-year OS (%)
Race/ethnicity
6
Conclusion
ERþ or PRþ, HER2
93.8 (92.4-95.1)
90.1 (88.5-91.7)
ERþ or PRþ, HER2þ
88.8 (84.4-93.4)
83.0 (77.9-88.4)
ER, PR, HER2þ
78.5 (70.6-87.3)
78.5 (70.6-87.3)
ER, PR, HER2
77.0 (71.6-82.8)
73.4 (67.8-79.5)
Data in parentheses are 95% confidence intervals. Abbreviations: BCSS ¼ breast cancer-specific survival; ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; PR ¼ progesterone receptor; SE ¼ Southeast.
Clinical Breast Cancer Month 2015
greater frequency in white and East Asian women (w70%) and lowest in Southeast Asian and South Asian women (w60%). Significantly greater rates of HER2þ cancer was found among South Asian, Southeast Asian, and Aboriginal women (w25%), with much lower rates in white women. Race/ethnicity has been associated with differential breast cancer outcomes in prospective and retrospective studies, including the present study. Differences in clinical outcomes among ethnicities are likely multifactorial, including socioeconomic factors, access to screening programs and treatment, inherent racial biologic factors, and variable distributions of breast cancer subtypes. Our results have important implications for resource allocation and clinical care in multiethnic settings with universal healthcare. In areas with high immigration rates, a gradual increase in the cost might be associated with treatment of certain breast cancer subtypes, especially HER2þ. Understanding patient age at presentation and the subtype prevalence could also aid in targeting immigrant groups with rational educational materials.
Disclosure The authors have stated that they have no conflicts of interest.
References 1. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature 2000; 406:747-52.
Dante Wan et al 2. Kwan ML, Kushi LH, Weltzien E, et al. Epidemiology of breast cancer subtypes in two prospective cohort studies of breast cancer survivors. Breast Cancer Res 2009; 11:R31. 3. Parise CA, Bauer KR, Brown MM, Caggiano V. Breast cancer subtypes as defined by the estrogen receptor (ER), progesterone receptor (PR), and the human epidermal growth factor receptor 2 (HER2) among women with invasive breast cancer in California, 1999-2004. Breast J 2009; 15:593-602. 4. Telli ML, Chang ET, Kurian AW, et al. Asian ethnicity and breast cancer subtypes: a study from the California Cancer Registry. Breast Cancer Res Treat 2011; 127: 471-8. 5. Huo D, Ikpatt F, Khramtsov A, et al. Population differences in breast cancer: survey in indigenous African women reveals over-representation of triple-negative breast cancer. J Clin Oncol 2009; 27:4515-21. 6. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA 2006; 295:2492-502. 7. Bauer KR, Brown M, Cress RD, Parise CA, Caggiano V. Descriptive analysis of estrogen receptor (ER)-negative, progesterone receptor (PR)-negative, and HER2negative invasive breast cancer, the so-called triple-negative phenotype: a populationbased study from the California Cancer Registry. Cancer 2007; 109:1721-8. 8. Lund MJ, Trivers KF, Porter PL, et al. Race and triple negative threats to breast cancer survival: a population-based study in Atlanta, GA. Breast Cancer Res Treat 2009; 113:357-70. 9. Carvalho LV, Pereira EM, Frappart L, et al. Molecular characterization of breast cancer in young Brazilian women. Rev Assoc Med Bras 2010; 56:278-87. 10. Furberg H, Millikan R, Dressler L, Newman B, Geradts J. Tumor characteristics in African American and white women. Breast Cancer Res Treat 2001; 68:33-43. 11. Eley JW, Hill HA, Chen VW, et al. Racial differences in survival from breast cancer: results of the National Cancer Institute Black/White Cancer Survival Study. JAMA 1994; 272:947-54. 12. Porter PL, Lund MJ, Lin MG, et al. Racial differences in the expression of cell cycle-regulatory proteins in breast carcinoma. Cancer 2004; 100:2533-42.
13. Clegg LX, Li FP, Hankey BF, Chu K, Edwards BK. Cancer survival among US whites and minorities: a SEER (Surveillance, Epidemiology, and End Results) program population-based study. Arch Intern Med 2002; 162:1985-93. 14. Shavers VL, Brown ML. Racial and ethnic disparities in the receipt of cancer treatment. J Natl Cancer Inst 2002; 94:334-57. 15. Surveillance, Epidemiology, and End Results Program. SEER stat database: mortalityeALL COD, public-use with state, total U.S. for expanded races/Hispanics (1975-2010). Bethesda, MD: National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, Cancer Statistics Branch. Available at: https://www.seer.cancer.gov. Accessed: May 1, 2015. 16. Kakarala M, Rozek L, Cote M, Liyanage S, Brenner DE. Breast cancer histology and receptor status characterization in Asian Indian and Pakistani women in the U. S.—a SEER analysis. BMC Cancer 2010; 10:191. 17. Statistics Canada. 2011 Census of Population. Available at: http://www12.statcan. gc.ca/census-recensement/index-eng.cfm. Accessed: May 1, 2015. 18. Kwan ML, John EM, Caan BJ, et al. Obesity and mortality after breast cancer by race/ethnicity: the California Breast Cancer Survivorship Consortium. Am J Epidemiol 2014; 179:95-111. 19. Lu Y, Ma H, Malone KE, et al. Obesity and survival among black women and white women 35 to 64 years of age at diagnosis with invasive breast cancer. J Clin Oncol 2011; 29:3358-65. 20. Sparano JA, Gray RJ, Zujewski JA, et al. Recurrence score and clinicopathologic characteristics of TAILORx participants by race and ethnicity. J Clin Oncol 2014; 32(suppl 26), abstract 36. 21. Warner ET, Tamimi RM, Hughes ME, et al. Racial and ethnic differences in breast cancer survival: mediating effect of tumor characteristics and sociodemographic and treatment factors. J Clin Oncol 2015; 33:2254-61. 22. Iqbal J, Ginsburg O, Rochon PA, Sun P, Narod SA. Differences in breast cancer stage at diagnosis and cancer-specific survival by race and ethnicity in the United States. JAMA 2015; 313:165-73.
Clinical Breast Cancer Month 2015
-7
Breast Cancer Subtype Variation by Race and Ethnicity in a Diverse Population in British Columbia Dante Wan,1 Diego Villa,1 Ryan Woods,2 Rinat Yerushalmi,1,3 Karen Gelmon1 Abstract The distribution of breast cancer subtypes varied significantly by race/ethnicity within our cross-sectional cohort from a multicultural population. Our data suggest that race/ethnicity plays a significant role in the biology of invasive breast cancer and that certain groups might experience different outcomes. These results have important implications for resource allocation and clinical care in multiethnic settings with universal healthcare. Background: Breast cancer subtypes occur differentially across different racial and ethnic groups. However, their distribution within a multicultural population is unknown. Materials and Methods: Patients with invasive breast cancer newly diagnosed in 2006 and referred to the British Columbia Cancer Agency were identified from the Breast Cancer Outcomes Unit database. Race/ethnicity data were abstracted from a patient-completed health assessment questionnaire completed at the initial consultation, and grouped as white, East Asian, Aboriginal, South Asian, Southeast Asian, and other. Breast cancer subtypes were created using available data on estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) status. Results: A total of 1829 women had complete data. Of these women, 73% were white, 8% were East Asian, 4% Aboriginal, 3% South Asian, 3% Southeast Asian, and 3% other. The median age at diagnosis was 60 years; the youngest were Southeast Asian (51 years) and the oldest were white (60 years; P < .001). The overall ER positivity rate was 81%, highest in East Asian women (89%) and lowest in South Asian women (73%). The HER2 positivity rate was 16% for all patients and was highest in the South Asian (28%), Southeast Asian (28%), and Aboriginal (24%) women and lowest in the white women (14%; P < .001). Triple-negative (ER-, PR-, and HER2-negative) breast cancer was uncommon in East Asian women (5%) but more common in South Asian women (19%; P < .001). The 5-year breast cancer-specific survival was 90% (95% confidence interval, 89%-92%), with no significant difference among the racial/ethnic groups (P ¼ .136). Conclusion: Breast cancer subtypes varied by race/ethnicity in our cross-sectional cohort of a multicultural population, suggesting that race/ethnicity plays a significant role in the biology of invasive breast cancer. Clinical Breast Cancer, Vol. -, No. -, --- ª 2015 Elsevier Inc. All rights reserved. Keywords: Aboriginal, Asian, Breast cancer subtypes, Estrogen receptor, HER2
Introduction Since the initial description of breast cancer subtypes by Perou et al1 in 2000, our concept of breast cancer has evolved to a better understanding of the unique biology, response, and prognosis of the specific 1 Division of Medical Oncology, British Columbia Cancer Agency, Vancouver, BC, Canada 2 Cancer Control Research, British Columbia Cancer Agency, Vancouver, BC, Canada 3 Division of Medical Oncology, Davidoff Center, Petah Tikvah, Israel
Submitted: Jun 1, 2015; Revised: Aug 31, 2015; Accepted: Sep 11, 2015 Address for correspondence: Diego Villa, MD, MPH, British Columbia Cancer Agency, 600 West 10th Avenue, Vancouver, BC, Canada E-mail contact: [email protected]
1526-8209/$ - see frontmatter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.clbc.2015.09.005
cancer subtypes. Various reports have suggested that the distribution of breast cancer subtypes varies by race/ethnicity. For example, epidemiologic data have suggested an increased prevalence of human epidermal growth factor receptor 2 (HER2)-positive breast cancer in Asians living in the United States.2,3 Specifically, HER2-positive breast cancer appears to be overrepresented in Chinese, Korean, Filipino, and Vietnamese women compared with non-Hispanic white women.4 Population-based studies have also suggested a greater prevalence of basal-like cancer in African5 and African-American women6-8 and in a number of other countries, including Brazil.9 Simultaneously, race/ethnicity has been associated with different breast cancer outcomes in retrospective and prospective studies that have not evaluated breast cancer subtypes. Compared with white
Clinical Breast Cancer Month 2015
-1
Distribution of Breast Cancer Subtypes by Race/Ethnicity women, African-American women present with a more advanced stage at diagnosis,10,11 higher grade tumors,12 and worse survival after controlling for prognostic factors.13,14 The age-adjusted breastcancer specific mortality rate in the United States in 2010 for white women was 21.3 per 100,000 compared with 30.2 per 100,000 in black women.15 However, the differences in the clinical outcomes among racial/ethnic groups are likely multifactorial, including socioeconomic factors, access to screening programs and treatment, inherent ethnic biology, and variable distributions of breast cancer subtypes. British Columbia (BC) is a multiethnic province with a predominantly white population and a significant proportion of East Asian, Aboriginal, South Asian, and Southeast Asian individuals. Residents of BC have access to universal, publicly funded healthcare, including breast cancer screening, diagnosis, and treatment. The objective of the present study was to assess the distribution of breast cancer subtypes by race/ethnicity in our population and evaluate whether race/ethnicity is associated with different breast cancer outcomes.
Materials and Methods Study Population All patients diagnosed with invasive breast cancer from January 1, 2006 to December 31, 2006 were identified in the Breast Cancer Outcomes Unit (BCOU) database. The year 2006 was chosen for 3 reasons. First, 2006 was the first year when HER2 testing was consistently performed on all new breast cancer biopsy specimens at our institution. Second, that year allowed a sufficient follow-up length to examine the 5-year survival outcomes. Finally, the race/ethnicity data were extracted manually from the paper medical records, posing incremental logistical and resource challenges if additional years were included. The BCOU database is a population-based registry that contains prospectively collected demographic, pathologic, staging, and initial treatment data for patients referred to the BC Cancer Agency (BCCA) for breast cancer management. The outcomes for all patients were obtained by an active letter of follow-up to the general practitioners and through periodic direct linkage to the provincial vital statistics agency. The outcomes information collected included first local, regional, and distant relapse and, if applicable, the date and cause of death.
Baseline Characteristics and Breast Cancer Subtypes
2
-
Variables, including breast cancer subtype, patient age, tumor grade, and tumor stage, were obtained from the BCOU database. The patients’ height and weight, measured at the initial consultation, were abstracted from the medical records to calculate the body mass index (BMI) for each patient. The breast cancer subtype was assigned using standard immunohistochemistry (IHC) to determine the estrogen receptor (ER), progesterone receptor (PR), and HER2 status on core biopsy or excision specimens. HER2 was considered positive if the IHC score was 3þ or if the HER2-to-chromosome 17 centromere ratio was 2 using fluorescent in situ hybridization (FISH). FISH was only performed in cases with indeterminate or inconclusive HER2 expression by IHC. For the present study, the combinations of IHC markers and FISH data were used as surrogates for the following
Clinical Breast Cancer Month 2015
intrinsic breast cancer subtypes: luminal A or B (ERþ or PRþ, HER2), HER2 enriched (ERþ or ER or PRþ or PR, and HER2þ), and basal-like (ER, PR, and HER2). The distinction between luminal A and B subtypes was not made using the available IHC results.
Race and Ethnicity At the initial consultation at a BCCA center, all patients completed the Health Assessment Form (HAF). The HAF explores patients’ medical and sociodemographic background, including selfreported race/ethnicity through the question “To which ethnic or cultural group do you belong?” Each patient was asked to self-assign to 1 of the following major groups: Aboriginal (eg, North American First Nations, Inuit, Metis), black (eg, African, Haitian), white or European, Latin American, or Asian. The Asian group was further subdivided geographically into West Asian (Arab), East Asian (Chinese, Japanese, Korean), South Asian (East Indian, Pakistani), and Southeast Asian (Indonesian, Laotian, Filipino). Patients were assigned an “other” ethnicity if they listed 2 ethnicities (mixed heritage) or if they self-identified with a category not listed (eg, Pacific Islander). Patients with incomplete data were excluded. For the present analysis, patients were grouped into the following 6 categories: white, East Asian, Aboriginal, South Asian, Southeast Asian, and other.
Statistical Analysis The main study endpoint was the distribution of breast cancer subtypes according to race/ethnicity, reported as frequencies. The proportion of ERþ and HER2þ cases according to breast cancer subtype was also calculated. Categorical variables were compared among the ethnic groups using the c2 test and continuous variables using the Kruskal-Wallis test. Because this was a retrospective, population-based study, no formal sample size calculations were performed. As a secondary endpoint, breast cancer-specific survival (BCSS) was calculated as the time from the date of diagnosis to death from a primary or secondary cause related to breast cancer. Patients who were alive at the last follow-up visit or who had died of nonebreast cancer causes were censored at the last follow-up visit. Overall survival (OS) was calculated from the date of the initial diagnosis to the date of the last follow-up visit or death from any cause. BCSS and OS were estimated using the Kaplan-Meier method and compared among groups using the log-rank test. The 5-year BCSS and OS rates are presented by ethnic group and breast cancer subtype with the 95% confidence intervals (CIs). Cox proportional hazards regression models were used to examine the relationship between race/ethnicity and survival outcomes, adjusted for breast cancer subtype, stage, and grade and patient factors, including age, smoking history, and BMI. Treatment information was not included in the Cox regression model, because it was not available. The proportional hazards assumption was examined using Schoenfeld residual plots, and the sensitivity of the results to departures from this assumption was assessed using models stratified by any terms of concern. All data analysis was performed using SAS, version 9.3, and the R statistical language, version 2.15. The University of BC/BCCA research ethics board approved the present study.
Dante Wan et al Results Patient Population A total of 2222 patients with invasive breast cancer newly diagnosed during 2006 were identified in the BCOU database. The following 393 patients (18%) were excluded (Figure 1): 150 without available HAF data, 135 with missing self-reported race/ethnicity on the HAF, and 108 with the information on ER, PR, and HER2 status required to assign a breast cancer subtype missing. The final analysis included 1829 patients. The distribution of race/ethnicity in descending order was white (n ¼ 1435; 73%), East Asian (n ¼ 155; 8%), Aboriginal (n ¼ 71; 4%), South Asian (n ¼ 65; 3%), and Southeast Asian (n ¼ 51; 3%). A total of 52 patients (3%) had a selfreported race/ethnicity that was either “other” or an uncommon choice with too few occurrences to be analyzed as a separate group (eg, African American; n ¼ 5).
Clinical and Demographic Variables The patient characteristics were unevenly distributed across the groups (Table 1). The median age at diagnosis was 58 years (interquartile range, 49-69); however, the age varied considerably by ethnicity, with white women presenting at a typically older age and Southeast Asian and other patients presenting with a lower median age at diagnosis. The proportion of patients aged < 40 years was not markedly different across the groups; these proportions were determined from very small counts for most groups (n < 10 for all groups except white). A greater proportion of Southeast Asian
Figure 1 Patient Identification and Exclusions
2222 pts with IBC idenƟfied in BCOU database 150 charts either unretrievable or lacking HAF forms 2072 medical charts with HAF forms for review 135 forms with missing race/ ethnicity data 1937 HAF forms with complete race/ ethnicity data
women were premenopausal (63%), more than twice the proportion for white women (29%). No significant difference was found in the stage distribution among the groups, although South Asian patients had a greater proportion of high-grade disease. Aboriginal women had the greatest rates of smoking and elevated BMI. East Asian women had the lowest BMI. South Asian patients had the lowest current and ever smoking rates.
Subtype Distribution by Ethnicity The ERþ/HER2 subtype was the most common breast cancer subtype (n ¼ 1320; 72%), followed by the HERþ (n ¼ 287; 16%) and ER/HER2 subtypes (n ¼ 222; 12%; Table 2). The white and East Asian patients had the greatest proportions of ERþ/HER2 breast cancer, and the Southeast Asian, South Asian, and Aboriginal patients had the greatest proportions of HER2þ breast cancer. South Asian and Aboriginal patients had a greater proportion of ER/HER2 disease, but this breast cancer subtype was very uncommon in East Asian patients. Overall, white and East Asian women had the most favorable subtype distribution, with a relatively greater proportion of ERþ/HER2 tumors and lower HER2þ and ER/HER2þ tumors. In contrast, South Asian women had relatively greater frequencies of the less favorable breast cancer subtype.
Survival Outcomes With a median follow-up of 5.5 years (range, 5-6) for the living patients, the 5-year BCSS for the entire cohort was 90% (95% CI, 89-92), with no significant difference among groups (P ¼ .136; Figure 2). The 5-year OS for the entire cohort was 87% (95% CI, 85-88), with a significant difference among groups likely related to the difference in age at diagnosis (P ¼ .030; Figure 3). The 5-year BCSS and OS rates for the cohort stratified by both race/ethnicity and breast cancer subtype are listed in Table 3. Cox proportional hazards regression analysis for the BCSS endpoint showed no significant difference in outcomes across the groups after adjustment (likelihood ratio test [LRT], P ¼ .131). This result was also consistent across the sensitivity analyses that stratified models by key prognostic variables. Models for the OS endpoint demonstrated a persistent significant difference in outcome across groups (LRT, P ¼ .053) after adjustment for other prognostic variables. This difference was driven largely by differences between the white and East Asian groups (hazard ratio, 0.40; 95% CI, 0.18-0.88; P ¼ .022). This result was consistent in sensitivity analyses that stratified models by key prognostic factors.
Discussion 108 pts with missing subtype informaƟon
1829 pts with known race/ethnicity and subtype data for final analysis
Abbreviations: BCOU ¼ Breast Cancer Outcomes Unit; HAF ¼ Health Assessment Form; IBC ¼ Invasive Breast Cancer; pts ¼ patients.
In our specific population, the distribution of breast cancer subtypes varied significantly by race/ethnicity. Consistent with previous studies, the ERþ/HER2 subtype was the most common subtype across all ethnicities, with greater frequencies in white and East Asian women (w70%) and lowest in Southeast Asian and South Asian women (w60%). These results suggest that race/ ethnicity plays a significant role in the biology of invasive breast cancer. To our knowledge, ours is the first study examining breast cancer subtypes in Aboriginal women (n ¼ 71). Consistent with findings from the Surveillance, Epidemiology, and End Results program, we observed that South Asian women had a greater incidence of ER/PR breast cancers than did white
Clinical Breast Cancer Month 2015
-3
Distribution of Breast Cancer Subtypes by Race/Ethnicity Table 1 Patient Characteristics at Diagnosis According to Ethnicity Race/Ethnicity Variable
White (n [ 1435)
East Asian (n [ 155)
Aboriginal (n [ 71)
South Asian (n [ 65)
SE Asian (n [ 51)
Other (n [ 52)
60.0
54.0
55.0
54.0
50.0
49.0
50.0-70.0
47.0-68.0
47.0-65.0
47.0-64.0
44.0-61.0
43.5-64.5
<.0001
Age at diagnosis (years) Median IQR
<.0001
Age group (years) <40
68 (4.7)
8 (5.2)
6 (8.5)
4 (6.2)
4 (7.8)
7 (13.5)
40-59
647 (45.1)
91 (58.7)
39 (54.9)
36 (55.4)
34 (66.7)
29 (55.8)
60
720 (50.2)
56 (36.1)
26 (36.6)
25 (38.5)
13 (25.5)
16 (30.8) <.0001
Menopausal status Premenopausal
420 (29.3)
66 (42.6)
27 (38)
27 (41.5)
32 (62.7)
27 (51.9)
Postmenopausal
994 (69.3)
89 (57.4)
44 (62)
38 (58.5)
19 (37.3)
25 (48.1)
Unknown
21 (1.5)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
Stage at diagnosis
.2267
I
590 (41.1)
65 (41.9)
24 (33.8)
15 (23.1)
20 (39.2)
21 (40.4)
II
563 (39.3)
58 (37.4)
29 (40.8)
37 (56.9)
22 (43.1)
21 (40.4)
III
219 (15.3)
28 (18.1)
15 (21.1)
11 (16.9)
7 (13.7)
8 (15.4)
IV
59 (4.1)
3 (1.9)
3 (4.2)
1 (1.5)
1 (2)
2 (3.8)
3 (0.2)
1 (0.6)
0 (0)
1 (1.5)
1 (2)
0 (0)
1
357 (24.9)
23 (14.8)
8 (11.3)
10 (15.4)
6 (11.8)
11 (21.2)
2
553 (38.5)
70 (45.2)
30 (42.3)
14 (21.5)
17 (33.3)
19 (36.5)
3
496 (34.6)
59 (38.1)
32 (45.1)
39 (60)
28 (54.9)
20 (38.5)
3 (1.9)
1 (1.4)
2 (3.1)
85 (54.8)
33 (46.5)
31 (47.7)
30 (58.8)
30 (57.7)
Unknown
<.0001
Grade
Unknown
29 (2)
0 (0)
2 (3.8)
Nodal status Negative
.4182 795 (55.4)
Positive
478 (33.3)
60 (38.7)
26 (36.6)
29 (44.6)
16 (31.4)
15 (28.8)
Unknown
162 (11.3)
10 (6.5)
12 (16.9)
5 (7.7)
5 (9.8)
7 (13.5)
Current
175 (12.2)
2 (1.3)
16 (22.5)
2 (3.1)
0 (0)
4 (7.7)
Former
646 (45)
13 (8.4)
34 (47.9)
2 (3.1)
10 (19.6)
13 (25)
Never
600 (41.8)
138 (89)
20 (28.2)
61 (93.8)
39 (76.5)
34 (65.4)
2 (3.9)
1 (1.9)
<.0001
Smoking status
Unknown
14 (1)
2 (1.3)
1 (1.4)
0 (0)
26.6
23.2
28.9
27.1
25.6
26.5
23.4-30.7
21.3-25.4
25.5-31.6
24.5-29.0
22.8-28.8
23.8-30.9
<.0001
BMI Median IQR
P Value
Data presented as n (%), unless noted otherwise. Abbreviations: BMI ¼ body mass index; IQR ¼ interquartile range; SE ¼ Southeast.
4
-
women.16 The rate of HER2þ cancer among South Asian, Southeast Asian, and Aboriginal women was significantly greater (w25%) but was much lower in white women. These findings are concordant with data from the California Cancer registry, in which Asian (Korean, Filipino, Vietnamese, and Chinese) women had greater rates of HER2þ breast cancer compared with non-Hispanic white women.4 Our results have important implications for resource allocation and clinical care in multiethnic settings with universal healthcare. Although we did not specifically examine the trend of breast cancer subtype distribution by race/ethnicity over time, it is likely that with the ever-increasing proportions of visible immigrants, a gradual
Clinical Breast Cancer Month 2015
increase in the cost associated with the treatment of certain breast cancer subtypes, especially the HER2þ subtype, would occur. This is particularly relevant in regions with a high rate of immigration. It is also important in terms of health education and the role of prevention and screening. Understanding patient age at presentation and the subtype prevalence could aid in targeting immigrant groups with rational educational materials. Approximately 1,162,900 foreign-born people arrived in Canada between 2006 and 2011.17 These recent immigrants accounted for 17% of the total foreign-born population in Canada and 3.5% of the total population. Of the Group 8 countries, in 2010, Canada
Dante Wan et al Table 2 Breast Cancer Subtype According to Ethnicity Race/Ethnicity Cancer Subtype
White
East Asian
Aboriginal
South Asian
SE Asian
P Value
Other
Hormone status
.0003
ERþ or PRþ, HER2
1058 (73.7)
120 (77.4)
45 (63.4)
35 (53.8)
31 (60.8)
31 (59.6)
ERþ or PRþ, HER2þ
130 (9.1)
19 (12.3)
12 (16.9)
12 (18.5)
10 (19.6)
11 (21.2)
ER, PR, HER2þ
67 (4.7)
9 (5.8)
5 (7)
6 (9.2)
4 (7.8)
2 (3.8)
ER, PR, HER2
180 (12.5)
7 (4.5)
9 (12.7)
12 (18.5)
6 (11.8)
8 (15.4)
1188 (82.8)
139 (89.7)
57 (80.3)
47 (72.3)
41 (80.4)
42 (80.8)
.0548
197 (13.7)
28 (18.1)
17 (23.9)
18 (27.7)
14 (27.5)
13 (25)
.0002
ERþ and/or HERþ ERþ HER2þ
Data presented as n (%). Abbreviations: ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; PR ¼ progesterone receptor; SE ¼ Southeast.
higher BMI, which could affect the outcomes.18,19 Further studies of Aboriginal women should add more information. Although we included a diverse population, this particular composition might not necessarily mirror that of other areas within North America or the rest of the world, limiting its generalizability. Our population was multiethnic but had a predominance of persons of either European or Asian origin. We did not consider finer subgroups of Europeans nor did we have a sizable population with African roots. Also, 243 of 2222 women (11%) were excluded from the analysis because of an unknown race/ethnicity or breast cancer subtype, as were 52 of 1829 women (3%) with self-reported “other” race/ ethnicity. Despite the large sample size, most of our defined race/ ethnicity categories contained a small number of patients; thus, our observed breast cancer subtype distributions within these groups were imprecise from a statistical viewpoint. Just as with a number of other studies, the use of IHC receptor profiles served only as a surrogate for the molecular breast cancer subtype as determined by gene expression profiling.
0.9 0.7
0.8
Caucasian East Asian Aboriginal South Asian SE Asian Other
P=0.136
0.6
Breast Cancer-Specific Survival
1.0
Figure 2 Breast Cancer-Specific Survival According to Ethnicity
0.5
had the greatest proportion of foreign-born peoples (21%), followed by Germany (13%) and the United States (13%). In the 2011 Canadian census, nearly 6,264,800 people identified themselves as a member of the visible minority population, accounting for 19% of the country’s total population, an increase from 16% in the 2006 census. The 3 largest visible minority groups—South Asians, Chinese, and blacks—accounted for 61% of the visible minority population.17 Knowledge of the racial/ethnic differences in cancer biology could, therefore, be potentially helpful in planning healthcare resources, including screening and treatment. Also, economic planning might be facilitated by knowing the rates of different subtypes in the population, especially because costly targeted therapies are determined by subtype. Many reports have suggested that a differential distribution of breast cancer subtypes across geographic areas or within specific races/ ethnicities is present. However, these studies were limited by nonuniversal access to healthcare, which might have led to the underrepresentation of large subgroups of the general population. Our study was performed in a Canadian province with universal healthcare and a centralized cancer system, reducing this potential bias. Although the referral rate for newly diagnosed invasive breast cancer to the BCCA is not absolute, our study captured 85% to 90% of all invasive breast cancer cases in BC. However, residual confounding secondary to referral bias would be unlikely to alter the study findings. The present study provided a sufficient sample size and follow-up length to evaluate the 5-year outcomes, especially in the curative setting. This is the first study to evaluate the distribution of breast cancer subtypes in a diverse population that includes North American Aboriginal women. Additionally, our system, with universal access and screening mammography program, might provide a more comprehensive patient population than could similar reports. Despite these strengths, visible minorities have continued to be underrepresented in most breast cancer clinical trials. Further understanding of the biology, response to treatments, and toxicity profile stratified by race/ethnicity might better inform the optimal management strategies for these particular subpopulations. Our results have shown that Aboriginal women present with the highest median BMI and are more likely to be current or recent smokers at diagnosis. This is of interest, in particular, because other recent studies have suggested that other races/ethnicities, specifically African-American women, might have a
0
1
2
3
4
5
Years since diagnosis
Abbreviation: SE ¼ Southeast.
Clinical Breast Cancer Month 2015
-5
Distribution of Breast Cancer Subtypes by Race/Ethnicity We chose to use self-reported race/ethnicity as the exposure of interest, realizing that no reference standard exists for defining a particular individual. Although the HAF explicitly queried for selfreported ethnicity, we realize that in multicultural settings, some people might identify with > 1 racial or ethnic category and that others might not be able to make a clear distinction between their own race and ethnicity. Although self-identified ethnicity might not be consistent with “biologic ethnicity,” “genetic ethnicity,” or race, it is possible that many patients used these terms interchangeably when completing the HAF. Because of the overlap between race and ethnicity, and because the distinction between Hispanic and nonHispanic ethnicity is not relevant in Canada, we chose to use the term “race/ethnicity,” consistent with other recent, large cohorts evaluating a similar research question.21,22
0.7
0.8
Caucasian East Asian Aboriginal South Asian SE Asian Other
P=0.030
0.5
0.6
Overall Survival
0.9
1.0
Figure 3 Overall Survival According to Ethnicity
0
1
2
3
4
5
Years since diagnosis
Abbreviation: SE ¼ Southeast.
Although different racial/ethnic groups might present with different biologic cancer factors, other factors might affect the outcome. Recent data have suggested that body composition and obesity vary across race/ethnicity, which could, in turn, affect the outcomes, independently of breast cancer subtype distributions.20 In our cohort, we found that our Aboriginal cohort presented with the highest median BMI and were more likely to be current or recent smokers at diagnosis. How these factors might affect outcomes requires study. Also, owing to the very small number of study subjects in some racial/ethnic subsets, overfitting of the statistical model and the generalizability of this research are potential limitations of our research. Furthermore, residual confounding might be present because treatment, socioeconomic status, and comorbidity data were not collected.
Table 3 Five-Year Breast Cancer-Specific and Overall Survival According to Ethnicity and Breast Cancer Subtype Variable
5-year BCSS (%)
Clinical Practice Points Breast cancer subtypes occur differentially across different racial/
ethnic groups. The ERþ/HER2 was the most common subtype, with a
White
89.5 (87.9-91.1)
85.5 (83.7-87.3)
East Asian
96.1 (93.1-99.2)
94.2 (90.6-97.9)
Aboriginal
92.8 (86.9-99.1)
88.7 (81.7-96.4)
South Asian
90.4 (83.3-98.0)
84.6 (76.3-93.9)
SE Asian
94.1 (87.9-100.0)
94.1 (87.9-100.0)
Other
92.3 (85.3-99.8)
90.4 (82.7-98.8)
Breast cancer subtype
-
Although the ethnicity subsets were small, we found differences in the rates of ER and HER2 positivity by ethnicity, with ER disease more frequent in Aboriginal and South Asian women. These groups and Southeast Asians also had a greater incidence of HER2þ disease. These results are provocative and suggest further studies and validation on larger populations are necessary. Studies that also include genomic analysis on both blood and tumor samples might provide more robust epidemiologic information.
5-year OS (%)
Race/ethnicity
6
Conclusion
ERþ or PRþ, HER2
93.8 (92.4-95.1)
90.1 (88.5-91.7)
ERþ or PRþ, HER2þ
88.8 (84.4-93.4)
83.0 (77.9-88.4)
ER, PR, HER2þ
78.5 (70.6-87.3)
78.5 (70.6-87.3)
ER, PR, HER2
77.0 (71.6-82.8)
73.4 (67.8-79.5)
Data in parentheses are 95% confidence intervals. Abbreviations: BCSS ¼ breast cancer-specific survival; ER ¼ estrogen receptor; HER2 ¼ human epidermal growth factor receptor 2; PR ¼ progesterone receptor; SE ¼ Southeast.
Clinical Breast Cancer Month 2015
greater frequency in white and East Asian women (w70%) and lowest in Southeast Asian and South Asian women (w60%). Significantly greater rates of HER2þ cancer was found among South Asian, Southeast Asian, and Aboriginal women (w25%), with much lower rates in white women. Race/ethnicity has been associated with differential breast cancer outcomes in prospective and retrospective studies, including the present study. Differences in clinical outcomes among ethnicities are likely multifactorial, including socioeconomic factors, access to screening programs and treatment, inherent racial biologic factors, and variable distributions of breast cancer subtypes. Our results have important implications for resource allocation and clinical care in multiethnic settings with universal healthcare. In areas with high immigration rates, a gradual increase in the cost might be associated with treatment of certain breast cancer subtypes, especially HER2þ. Understanding patient age at presentation and the subtype prevalence could also aid in targeting immigrant groups with rational educational materials.
Disclosure The authors have stated that they have no conflicts of interest.
References 1. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature 2000; 406:747-52.
Dante Wan et al 2. Kwan ML, Kushi LH, Weltzien E, et al. Epidemiology of breast cancer subtypes in two prospective cohort studies of breast cancer survivors. Breast Cancer Res 2009; 11:R31. 3. Parise CA, Bauer KR, Brown MM, Caggiano V. Breast cancer subtypes as defined by the estrogen receptor (ER), progesterone receptor (PR), and the human epidermal growth factor receptor 2 (HER2) among women with invasive breast cancer in California, 1999-2004. Breast J 2009; 15:593-602. 4. Telli ML, Chang ET, Kurian AW, et al. Asian ethnicity and breast cancer subtypes: a study from the California Cancer Registry. Breast Cancer Res Treat 2011; 127: 471-8. 5. Huo D, Ikpatt F, Khramtsov A, et al. Population differences in breast cancer: survey in indigenous African women reveals over-representation of triple-negative breast cancer. J Clin Oncol 2009; 27:4515-21. 6. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA 2006; 295:2492-502. 7. Bauer KR, Brown M, Cress RD, Parise CA, Caggiano V. Descriptive analysis of estrogen receptor (ER)-negative, progesterone receptor (PR)-negative, and HER2negative invasive breast cancer, the so-called triple-negative phenotype: a populationbased study from the California Cancer Registry. Cancer 2007; 109:1721-8. 8. Lund MJ, Trivers KF, Porter PL, et al. Race and triple negative threats to breast cancer survival: a population-based study in Atlanta, GA. Breast Cancer Res Treat 2009; 113:357-70. 9. Carvalho LV, Pereira EM, Frappart L, et al. Molecular characterization of breast cancer in young Brazilian women. Rev Assoc Med Bras 2010; 56:278-87. 10. Furberg H, Millikan R, Dressler L, Newman B, Geradts J. Tumor characteristics in African American and white women. Breast Cancer Res Treat 2001; 68:33-43. 11. Eley JW, Hill HA, Chen VW, et al. Racial differences in survival from breast cancer: results of the National Cancer Institute Black/White Cancer Survival Study. JAMA 1994; 272:947-54. 12. Porter PL, Lund MJ, Lin MG, et al. Racial differences in the expression of cell cycle-regulatory proteins in breast carcinoma. Cancer 2004; 100:2533-42.
13. Clegg LX, Li FP, Hankey BF, Chu K, Edwards BK. Cancer survival among US whites and minorities: a SEER (Surveillance, Epidemiology, and End Results) program population-based study. Arch Intern Med 2002; 162:1985-93. 14. Shavers VL, Brown ML. Racial and ethnic disparities in the receipt of cancer treatment. J Natl Cancer Inst 2002; 94:334-57. 15. Surveillance, Epidemiology, and End Results Program. SEER stat database: mortalityeALL COD, public-use with state, total U.S. for expanded races/Hispanics (1975-2010). Bethesda, MD: National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, Cancer Statistics Branch. Available at: https://www.seer.cancer.gov. Accessed: May 1, 2015. 16. Kakarala M, Rozek L, Cote M, Liyanage S, Brenner DE. Breast cancer histology and receptor status characterization in Asian Indian and Pakistani women in the U. S.—a SEER analysis. BMC Cancer 2010; 10:191. 17. Statistics Canada. 2011 Census of Population. Available at: http://www12.statcan. gc.ca/census-recensement/index-eng.cfm. Accessed: May 1, 2015. 18. Kwan ML, John EM, Caan BJ, et al. Obesity and mortality after breast cancer by race/ethnicity: the California Breast Cancer Survivorship Consortium. Am J Epidemiol 2014; 179:95-111. 19. Lu Y, Ma H, Malone KE, et al. Obesity and survival among black women and white women 35 to 64 years of age at diagnosis with invasive breast cancer. J Clin Oncol 2011; 29:3358-65. 20. Sparano JA, Gray RJ, Zujewski JA, et al. Recurrence score and clinicopathologic characteristics of TAILORx participants by race and ethnicity. J Clin Oncol 2014; 32(suppl 26), abstract 36. 21. Warner ET, Tamimi RM, Hughes ME, et al. Racial and ethnic differences in breast cancer survival: mediating effect of tumor characteristics and sociodemographic and treatment factors. J Clin Oncol 2015; 33:2254-61. 22. Iqbal J, Ginsburg O, Rochon PA, Sun P, Narod SA. Differences in breast cancer stage at diagnosis and cancer-specific survival by race and ethnicity in the United States. JAMA 2015; 313:165-73.
Clinical Breast Cancer Month 2015
-7