Introduction

It is estimated that in 2006 about 41,000 women in the U.S. died from breast cancer and a disproportionate number were black [1]. This excess in mortality is a reflection of the markedly poorer survival experienced by black women, which has persisted for decades [2]. Major reasons for this disparity have been attributed to the increased number of black women diagnosed at later stage of disease and at an earlier age, where breast cancers tend to be more aggressive and result in poorer outcomes. Nationally, racial differences in survival continue even within stage and age groups [3–6]. Other factors which may contribute to this disparity include socioeconomic and demographic factors, cultural beliefs, health care access, co-morbid conditions, and tumor biology [7–19]. Another factor may be differences in delivery of cancer treatment. Adherence to established treatment guidelines may not be uniform across racial/ethnic groups and differences in treatment may be a major contributing factor to racial disparities in outcome [5, 14, 20–22].

Clinical practice guidelines for first course treatment of invasive breast cancer are well-established [23–25]. Some studies suggest that U.S. black women diagnosed with breast cancer are less likely to receive treatments concordant with established guidelines [26–28]. Several have reported racial differences in treatment delay, prescription, utilization, and response [29–32]. Black women may less frequently receive first course surgical and radiation treatment according to recommended guidelines [20, 27, 33, 34] and more frequently may not receive any surgery, even after accounting for a greater prevalence of inoperable cancers [21]. The benefits of adjuvant therapies to improving outcomes are clear [35, 36]. Appropriate use of adjuvant therapies reduce the risk of recurrence and improve chances for long-term survival. However, some studies suggest that black women are less likely to receive appropriate chemotherapy [37–42]. Whether black women are less likely to receive hormonal therapy remains equivocal [10, 39, 43]. It is established that black women have a lower incidence of Estrogen receptor positive and/or progesterone receptor positive (ER+/PR+) tumors and thus for a higher proportion of black women, hormonal therapy would not be indicated [44–46]. However, the prescription of hormonal therapy in published studies has not been studied separately among hormone receptor positive patients, who should be offered hormonal therapy, and negative patients, who should not. Treatment may also differ by health system providers [27, 28].

To assess racial differences in the initial treatment of invasive breast cancer, we used data from the Atlanta Surveillance, Epidemiology, and End Results (SEER) Registry. In this population-based study, we focused on the five metropolitan Atlanta SEER counties, an area with a sizeable black population and where large teaching hospitals are located.

Methods

Data collection and description

Data were obtained from the Atlanta SEER Registry of the National Cancer Institute, which has collected population-based cancer data in the metropolitan Atlanta area since 1975. All female invasive breast cancer cases (ICD-O site codes C50.0–C50.9, behavior code 3) [47] diagnosed among residents of the five Atlanta SEER counties (Fulton, Dekalb, Cobb, Clayton, and Gwinnett) from January 1, 2000 to December 31, 2001 were included. A small percentage of women who reported race as neither black nor white were excluded (2.1%). Patients diagnosed only by death certificate or autopsy were also excluded.

Treatment factors analyzed included type of first course treatment, delay in first course treatment, delivery of cancer directed surgery, type of surgery, and receipt of chemotherapy, radiotherapy, or hormonal therapy. Categories presented are based on treatment guidelines [23, 24, 48, 49]. The delay interval was calculated by subtracting the date of diagnosis from the earliest date of first course treatment; surgery, neoadjuvant chemotherapy, or initiation of chemotherapy or radiation therapy for distant metastatic disease. The delay groups were then established based on guidelines and literature review. Type of surgery was categorized into two groups, breast conserving surgery (BCS) or mastectomy. BCS included excisional biopsy/lumpectomy, wedge resection, quadrectomy, tylectomy, and segmented or sub-cutaneous mastectomy. Mastectomy included simple, modified, radical, and NOS (not otherwise specified), with or without removal of the uninvolved contralateral breast. Adjuvant treatments (radiotherapy, chemotherapy, and hormonal therapy) were dichotomized into receipt/no receipt categories.

Racial differences in treatment were analyzed according to their basis for treatment, based on stage at diagnosis, age at diagnosis, and specific tumor characteristics. Race in the SEER data is abstracted from the medical record and supplemented, where missing, from additional sources. Stage at diagnosis was presented according to the SEER Summary Staging guidelines using the broad categories of local, regional, or distant disease [50]. Age at diagnosis was categorized into three groups (18–50, 51–70, and > 70) according to age-specific treatment criteria. ER and PR status were categorized as positive (+) which included borderline, negative (−), or unknown. Tumor size, number of positive nodes, and nodal status were collected using SEER Extent of Disease guidelines [51] and grouped according to the AJCC Cancer Staging Manual [52] and treatment criteria [23, 24, 48, 49]. Grade represents the degree of tumor cell differentiation; defined as highly (Grade 1), moderately (Grade 2), or poorly (Grade 3) differentiated. In the absence of a stated differentiation, the Bloom–Richardson grading scheme was used [53].

Statistical analyses

Frequency distributions and χ2 tests of independence were used to describe the study population by race as well as to examine racial differences in treatment within stage groups. In order to examine racial differences in treatment according to recommended guidelines, the race–treatment relationships were then stratified on selected age groups, disease and tumor characteristics, and other treatments. Frequency distributions, χ2 tests of independence, and Cochran–Mantel–Haenszel statistics were utilized to analyze the racial differences in and across strata. Unknown categories for all variables were kept in the analyses. P-values reflect analyses with and without unknown categories included.

Results

Table 1 displays population and tumor characteristics by race. The study population was 26.5% black and 73.5% white. Most of the black cancer patients resided in the counties of Fulton and Dekalb (75.9%), while the majority of white patients were from Cobb, Fulton, and Gwinnett counties (75.5%). Black women were highly more likely than white women to be diagnosed at age 50 or younger (45.3% vs. 27.3%). The reverse was true for women over age 70 (15.3 vs. 24.7% respectively). Black women were also more likely to be diagnosed at later stage; 48.4% were diagnosed with regional or distant disease vs. 33.9% for white women. The tumors of black women were also much more frequently larger size, ER−, PR−, and high grade. The proportion of women diagnosed with four or more positive regional nodes was also higher, as was the percentage of women with nodes not examined (20.0% for black women vs. 11.8 % for white women). The status of a large proportion of the tumor characteristics was unknown, but these differences did not appear to differ significantly by race, with the exception of tumor size and grade. Within each of the 5 Atlanta SEER counties, the results were consistent with the overall results reported (data not shown). Within stage groups, most of the racial differences also remained (data not shown). The only exception was among those diagnosed with distant disease; differences in PR or grade status were not observed.

Table 1 Population and tumor characteristics by race for women diagnosed with invasive breast cancer, Atlanta, GA, 2000–2001
Full size table

Racial differences in first course treatment, within stage, are presented in Table 2. Between 21 and 27% of the women received treatment on the same day of diagnosis, primarily surgery. Black women were more likely than white women to experience delays greater than 30 days, regardless of stage at diagnosis. The delay for the majority of black women was greater than two weeks and black women were 4- to 5-fold more likely to experience delays greater than 60 days.

Table 2 First course treatment for invasive breast cancer (by race, within stage), Atlanta, GA, 2000–2001
Full size table

Most women diagnosed with local-regional disease received surgery as their first type of treatment within the spectrum of first course therapy, but racial differences were observed (Table 2). While chemotherapy was the first type of treatment received for a relatively small proportion of women, black women were more than twice as likely as white women to first receive chemotherapy at any stage of disease.

Greater than 90% of both black and white women diagnosed with either local or regional disease received surgery (Table 2), but surgery was significantly less often performed on black women (Table 2). However, of those who received surgery, there were no racial differences observed for type of surgery, about 65% of women with local disease received BCS and over 40% with regional disease. Racial differences in nodal scope of surgery and number of nodes examined were also present. Among women with local or regional disease, black women were more likely to not have any lymph nodes examined and less likely to have sentinel lymph node biopsies. However, among those who did have lymph nodes examined, more black women had greater than 10 nodes examined.

Tables 3 and 4 exhibit racial differences in surgical procedures and radiation, restricted to women diagnosed with local or regional disease. Black women were less likely than white women to receive surgery for tumors larger than 1.0 cm (Table 3). Black women were also less likely to receive surgery within each age group. Despite these differences in administration of surgery, racial differences in type of surgery by tumor size or age group were not observed.

Table 3 Racial differences in recipt of surgery, statified on tumor size and agea
Full size table
Table 4 Racial differences in receipt of radiation therapy by type of surgery, stratified on tumor size, nodal status, and agea
Full size table

Overall, about 62% of black women and 72% of white women received radiotherapy after BCS (P < 0.001) (Table 4). For patients diagnosed prior to age 71, black women less often received radiotherapy after BCS. When stratified by tumor size, a lower proportion of black women received BCS-radiation, but only for tumors < 2.0 cm. Among women with tumors larger than 5.0 cm treated with mastectomy, black women were also less likely to receive radiotherapy. Black women were also less likely to receive combined BCS—radiotherapy when less than 4 nodes were positive or mastectomy—radiotherapy when 4 or more nodes were positive.

Table 5 shows no racial differences in chemotherapy receipt among women diagnosed with node-positive local or regional disease (72.9% and 70.8% of black and white women received chemotherapy), but younger women were more likely than older women to receive chemotherapy. Black women were more likely than white women, however, to receive chemotherapy for node negative disease (34.6% vs. 26.3% respectively). Among node-negative patients, the majority of women with smaller tumors that were either higher grade or ER−PR− did not receive chemotherapy.

Table 5 Racial differences in receipt of chemotherapy for invasive breast cancer, stratified on tumor characteristics, nodal status, and agea, b
Full size table

A minority of women received hormone therapy for ER and/or PR positive disease, regardless of age (Table 6). Striking racial differences were also observed. Overall, black women were about 40% as likely to receive hormonal treatment for tumors that were ER and/or PR positive, and this was consistent within all age groups under 71.

Table 6 Racial differences in receipt of hormonal therapy for invasive breast cancer, stratified on hormone receptor status and agea
Full size table

Discussion

Although adherence to established treatment guidelines were relatively high for all women, black women were less likely to receive the recommended stage and tumor characteristic specific surgical, radiation, and hormonal treatments and were more likely to experience delays in treatment. No differences were observed for receipt of chemotherapy.

Delay

In our study, black women were more likely to experience treatment delays, particularly very long delays, within any given stage. Median time to treatment decreased as stage progressed, but significant racial differences remained. Several studies have reported that black women are more likely to experience delays in treatment; delays that are not explained away by demographic and clinical factors [29–31]. These delays appear to be especially prevalent among younger women [30, 31]. In our study although the racial differences in delay were equivalent across age strata (data not shown), almost 50% of black patients were under age 50. Regardless of stage and consistent with other studies [6, 14, 54], we also found that breast cancer in black women, particularly those under age 50, was more prone to unfavorable prognostic factors that bode a more aggressive disease. This combination of more aggressive tumors with less timely treatment could seriously impact outcome. Overall delay is associated with lower survival [55]. However, the influence of treatment delays are less certain as most studies fail to distinguish diagnosis delay from treatment delay [30, 55–57]. Results from the National Breast and Cervical Cancer Early Detection Program reported treatment delays of two weeks or less [29]. In our study, only 50% of white women and 38% of black women received treatment within a two week interval. Time to diagnosis overall, and between races, was less efficient in our study and reasons for this merit further research.

Surgery

A disproportionate number of black women did not receive cancer directed surgery for local-regional disease, 7.5% vs. 1.5% of white women. Of these, 56% of black women and 27% of white women did receive chemotherapy. Tumor response to the chemotherapy may have inadvertently led patients to forgo surgery, but reasons for surgery not being performed need investigation. Although surgery is not recommended for metastatic disease, about 40% of women with distant metastatic disease received surgery. It is appropriate to surgically treat patients in these circumstances for local control, quality of life, and palliative reasons. We do not know if this was the case. BCS is the preferred surgical choice for local-regional disease and smaller tumors [23, 48], and we found that black and white women received BCS equivalently. Similar to our findings, others have reported a lack of cancer-directed surgery among black women, but results regarding BCS have been conflicting [20, 27, 58, 59]. Surgery may be contraindicated for some women; however our findings do suggest surgical under-treatment for black women with local-regional disease, over-treatment of all women with metastatic disease, and that breast conservation for all women merits improvement.

Radiation

For local-regional disease, radiation therapy should be given after BCS, regardless of nodal status [23, 48]. Only 60–70% of our study patients received radiotherapy after BCS. Compared to white women, black women were consistently and significantly less likely to receive radiotherapy after BCS; including those with smaller tumors, node negative disease, < 4 positive regional nodes, or under age 70 years. Radiation should also be given after mastectomy if the tumor is greater than 5.0 cm or 4 or more nodes are positive [23, 24]. Black women, again, were less likely to receive radiation for either of these scenarios, although the difference for the former was not significant. Other population-based studies have indicated that black women are at significant risk of not receiving radiotherapy as recommended [15, 27, 33] and reasons for this may be related to extraneous factors such as transportation and social support.

Chemotherapy

In general, women were under-treated with adjuvant chemotherapy. Although black women were somewhat more likely to receive chemotherapy, the racial differences were not significant, regardless of nodal status, tumor size, or age. Of note, the proportion of women with combined small tumors and aggressive traits (higher grade, or ER−PR−) who should have received chemotherapy was very low (12–24%). We also investigated type of chemotherapy received (data not shown). Multiple agents are preferred over single agents and a vast majority of both black and white women received multiple agents (96.5% and 96.0% respectively).

Hormonal therapy

The benefits of hormonal therapy are clear. Adjuvant hormonal therapy should be recommended to all women with ER or PR positive tumors, regardless of age, menopausal status, tumor size, nodal status, or Her2/neu status [23, 24, 48]. Only a minority of women in our study who could have received hormonal therapy were documented as having done so; 41% of the women who were ER and/or PR positive. Receipt of hormonal therapy among black women was considerably less than for white women and did not appreciably vary whether we investigated ER+ and/or PR+, ER+, PR+, or ER+PR+ tumor status, or age groups. Consistent with other population-based findings [45], the prevalence of ER or PR positive tumors was significantly less for black women compared to white women. Thus, due to tumor status, and exacerbated by treatment differences, black women in our study were at a severe disadvantage in reaping the benefits of hormone therapy.

While the effects of racial differences in breast cancer treatments are uncertain [5, 10, 14, 20–22, 33], treatment recommendations are based on high level evidence that demonstrate their efficacy for improving outcomes. The Institute of Medicine recently published a document outlining quality measures for improving standards of breast cancer diagnosis and treatment in Georgia, based on established treatment guidelines [60]. Our findings from Atlanta SEER data, suggest these standards are not adequately implemented even in a large metropolitan area where teaching hospitals provide a sizeable proportion of breast cancer care. With respect to adjuvant therapies, a preponderance of women was under-treated. A SEER national study reported a decrease in appropriate use of adjuvant therapies from 1983–1995 [61] and a more recent SEER national study found that trends in the appropriate use of adjuvant hormonal therapy and chemotherapy have increased over time from 1987–2000, but 10–25% of the patients received neither treatment and the results were not reported within racial/ethnic groups [62]. The lack of quality care among black women in our study is of particular concern. Treatment differences can adversely affect outcome and reasons for the differences need to be explored. Efforts to reduce suboptimal treatment could improve outcomes for all women, and could significantly contribute to eliminating the poorer outcomes experienced by black women. The power of provision of high quality care is illustrated in the population-based study of Wojcik and colleagues [17]. In this US Department of Defense equal access system, black women diagnosed with breast cancer had a five-year mortality rate of 24.8% compared with 18.1% for whites. Equivalent rates in the US at that time were 34.2 and 18.2%. Their study suggests that improved access and utilization of prevention and treatment services can reduce survival disparity by as much as two-thirds.

Limitations

The under-treatment we report for adjuvant therapies may be inflated. The medical record is not as complete for adjuvant treatments and under-reporting may be considerably high in the SEER registry, especially for chemotherapy and hormonal therapy. This is partially attributable to adjuvant therapies being given outside of the diagnostic and treatment facilities from which the SEER data is abstracted, and potentially months after initial diagnosis. Notwithstanding, these differences should not affect women differentially, thus the observed racial differences need further investigation. In a sub-sample of patients whose data were re-abstracted as part of a patterns of care study, we observed that a higher proportion of patients were found to receive hormonal therapy, yet the racial differences persisted. Information on HER-2/neu status is not available from the SEER registry, thus we could not examine the equivalency of HER-2/neu testing among women or appropriate receipt of systemic therapy. We also did not have information on health status and co-morbidities which could have contraindicated receipt of some treatments. However, the data indicate that contraindications were not a major contributing factor.

Strengths

This is one of the first studies to report on racial differences in a comprehensive array of first course treatments, utilizing specific treatment guidelines as the standard for comparing differences. This study may be the first in a targeted population-based setting where outcomes are more readily translational to the community. In the national SEER data, black women comprise only 7–8% of the female breast cancer population. Our study represented a large proportion of black women, women for whom there has been a paucity of representation or treatment studies.

Conclusion

Our findings suggest treatment standards are not adequately or equivalently met among Black and White women, even in an area where teaching hospitals provide a substantial portion of breast cancer care. A considerable proportion of women received sub-optimal first course treatment for invasive breast cancer. Black women were more likely to not receive recommended surgical, radiation, or hormonal treatments and to experience delays in treatment. Treatment differences can adversely affect outcome and reasons for the differences need to be addressed. Efforts to translate these findings into improved treatment adherence could enhance outcomes for all women, particularly those most disadvantaged.