Characteristics and treatment modalities for African American women diagnosed with stage III breast cancer†
Presented in part at the Society of Surgical Oncology, March 13–16, 2007, Washington, DC and American Society of Clinical Oncology, June 1–5, 2007, Chicago, Illinois.
Stage III breast cancers account for about 6% to 7% of all invasive breast cancers diagnosed annually in the United States. In African American (AA) women, the incidence of stage III breast cancers is almost double that in Caucasian women. The aim of this study was to correlate age, receptor status, nuclear grade, and differences in treatment modalities for stage III breast cancer in an inner-city hospital serving a large AA population.
A retrospective review was performed for all stage III primary breast cancers diagnosed and or treated from 2000 to 2006.
Of 840 primary invasive breast cancers, the authors identified 107 as stage III, 40.2% IIIA, 32.7% IIIB, 16.8% T4D, and 10.3% IIIC. The majority of the patients were AA (n = 93, 86.9%). Stage IIIC patients were younger (P < .05). Triple negative tumors (TNT) accounted for 29.0%. TNT were more likely among the inflammatory breast cancers (50.0%) compared with the other 3 groups (P < .05). Twenty-two patients (20.5%) refused chemotherapy, and 24 of the 91 patients (26.3%) who should have received chest wall radiation refused. There was no difference in race, marital status, religion, or age in the patients that refused chemotherapy or radiation therapy versus the majority of patients in this series who received standard care.
Stage III breast cancers in AA women have distinct clinical characteristics. A high number of these patients refused chemotherapy and radiation therapy. Reasons for refusal need to be better defined so strategies can be implemented to improve compliance for these advanced stage patients. Cancer 2009. © 2009 American Cancer Society.
Stage III breast carcinomas account for about 6% to 7% of all invasive breast cancers diagnosed annually in the United States, with inflammatory breast cancer accounting for 1% to 2% of the cases. Inflammatory breast cancer (stages IIIB, T4d) is now recognized as a distinct clinical pathologic entity, characterized by different prognostic profiles.1
Locally advanced breast cancers evolved as a unique diagnostic entity when Bell in 1814 described the seriousness of a breast mass presenting with pain and overlying skin discoloration. Clinically, locally advanced cancers are characterized by skin ulceration, satellite nodules, and tumor fixation to the chest wall. Inflammatory breast carcinoma is characterized clinically by mottled erythema over at least ⅓ of the breast, warmth, peau d'orange, lymphedema, and ridges dividing areas of induration and normal areas of skin.
The incidence of stage III breast cancer among African American (AA) women is twice as prevalent as observed in Caucasian women.2, 3 Multivariate analyses of breast cancer survival rates have shown that AA women are more likely to present with advanced stage breast cancer compared with Caucasian women in equal-access healthcare systems.4, 5 These findings demonstrate that later stage breast cancer presentation is a major contributor to the lower survival rates described in AA women.
Another contributing factor to the lower survival rates for African Americans is racial differences in treatment. Socioeconomic disadvantages and decreased access to healthcare are likely to result in delayed diagnosis and treatment, as well as treatment noncompliance.3 In addition, AA women with breast cancer are at increased risk for developing high-grade, estrogen receptor-negative disease, and they face an increased risk of being diagnosed at younger age with breast carcinoma.6
The aim of this study was to correlate age, receptor status, nuclear grade, and differences in treatment modalities for stage III breast cancer in an inner-city cancer center serving a largely AA population.
MATERIALS AND METHODS
A retrospective chart review was performed on all stage III primary breast cancer patients diagnosed and/or treated among female patients from 2000 to 2006 at a single inner-city cancer center in Atlanta, Georgia serving a predominantly AA population. Data concerning disease stage at diagnosis were collected according to the Surveillance, Epidemiology, and End Results hospital tumor registry database and supplemented by patient chart review. The stage at diagnosis was based on the 6th edition of the American Joint Committee on Cancer Classification and Staging of Tumors (stages I-IV). This study was approved by the Emory University Institutional Review Board and Grady Hospital Research Oversight Committee.
Stage III was subclassified into 4 groups; IIIA, IIIB, T4D, and IIIC. The independent variables, including age, race, estrogen receptor (ER) status, progesterone receptor (PR) status, HER-2/neu (HER2) expression, triple negative status, nuclear grade, and differences in surgical and systemic treatment, were categorized as shown in Tables 1 and 2. Triple negative tumor (TNT) classification was indicated by negative expression of ER, PR, and HER2.
Table 1. Patients and Tumor Characteristics by Stage III Subgroups
|Age at diagnosis, y|| || || || || || || || || || || |
| Median||51.0||NA||54.0||NA||53.0||NA||50.5||NA||51.0||NA|| |
| Range||27-93||NA||27-81||NA||32-93||NA||33-80||NA||36-78||NA|| |
| <50 y||46||42.9||17||39.5||16||45.7||5||45.5||8||44.4||.68|
| ≥50 y||61||57.1||26||60.5||19||54.3||6||54.5||10||55.6|| |
|Race|| || || || || || || || || || || |
| Non-AA||14||13.1||4||9.3||4||11.4||3||27.3||3||16.7|| |
|ER status|| || || || || || || || || || || |
| Negative||47||43.9||17||39.5||15||42.8||11||61.1||4||36.3|| |
| Unknown||5||4.7||1||2.3||2||5.7||1||5.5||1||9.2|| |
|PR status|| || || || || || || || || || || |
| Negative||50||46.7||18||41.8||16||45.7||5||45.4||11||61.1|| |
| Unknown||8||7.6||2||4.8||3||8.6||2||11.9||1||9.5|| |
|HER2 status|| || || || || || || || || || || |
| Negative||82||76.6||35||81.4||24||68.5||15||83.3||8||72.7|| |
| Unknown||13||12.2||4||9.3||7||20.1||1||5.6||1||9.2|| |
|Triple negative§|| || || || || || || || || || || |
| No||68||63.6||30||69.8||24||68.6||8||44.4||6||54.6|| |
| Unknown||8||7.4||2||4.6||4||11.4||1||5.6||1||9.0|| |
|Grade|| || || || || || || || || || || |
| Intermediate||29||27.1||15||34.8||7||20.2||5||27.7||2||18.1|| |
| High||43||40.1||17||39.7||12||34.2||8||44.5||6||54.5|| |
| Unknown||24||22.4||5||11.6||13||37.1||5||27.7||1||9.4|| |
Table 2. Treatment by Stage III Subgroups
|Chemotherapy|| || || || || || || || || || |
|Radiotherapy|| || || || || || || || || || |
| Not eligible||16||15.0||6||13.9||7||20.0||2||11.3||1||9.3|
These marker studies were performed on formalin-fixed, paraffin-embedded tissue. Antibody clones 1D5 and PgR636 (Dako, Glostrup, Denmark) were used in the ER and PR assays, respectively. Polyclonal HER2 (Dako) was used for the HER2 assay.
ER and PR scores were based on nuclear staining, and positive staining was considered in cases with >5% of the nuclear positivity. HER2 scoring was based on cytoplasmic membrane staining and follows the American Society of Clinical Oncology and College of Anatomical Pathology guidelines and scoring system.7
Frequency distributions and Fisher exact tests (t test for age) were used to evaluate differences between the stage III subgroups by race, age, tumor characteristics, and treatment modalities.
Of the 840 primary female breast cancers, 107 (12.7%) patients presented with stage III disease. Of the 107, 43 (40.2%) patients were stage IIIA, 35 (32.7%) stage IIIB, 18 (16.8%) inflammatory breast cancer, and 11 (10.3 %) stage IIIC (Table 1). The vast majority of our population included AA patients (n = 93, 86.9%); the remaining 14 (13.1%) were Caucasian, Hispanic, and other ethnicities. About 43% were aged <50 years when diagnosed, with mean/median diagnosis age 53.4/51.0 years. ER, PR, and HER2 status was unknown in 5, 8, and 13 cases, respectively. Stage IIIC patients were younger (mean = 50.8) than stage IIIA and stage IIIB patients (P < .05). TNTs (defined as ER−, PR−, HER2−) comprised 29.0% of the stage III tumors and were more likely among inflammatory cancer patients (50.0%) than stage IIIA (25.5%), IIIB (20.0%), or IIIC (36.3%) patients (P < .05). High nuclear grade was most prevalent among IIIC patients (54.5%) and least prevalent among IIIB patients (34.2%).
Differences in chemotherapy and radiation therapy for all stage III patients are reported in Table 2. All patients were offered a combination of chemotherapy, surgery, and radiation therapy. All patients underwent surgical excision. Twenty-two patients refused chemotherapy (20.6%). All patients (n = 91) with primary or residual tumors ≥5.0 cm or ≥3 positive axillary nodes were considered for chest wall and axillary radiation therapy. Of those, 60 patients were treated, and 24 (22.4%) refused radiation therapy. There was no difference in marital status, religious background, or age of the patients who refused either chemotherapy or radiation therapy versus the majority of patients in this series who received standard of care therapy (data not shown).
For this group of stage III patients, there were no differences in age, tumor, and treatment characteristics between AA and non-AA women, with the exception of PR− (P = .04) and higher grade tumors (P = .04) being more prevalent among AA women (Table 3).
Table 3. Population, Tumor, and Treatment Characteristics by Race
|Stage|| || || || || |
| IIIB||31||33.3||4||28.5|| |
| T4d||15||16.1||3||21.5|| |
| IIIC||8||8.7||3||21.5|| |
|Age at diagnosis, y|| || || || || |
| Median||51.5||NA||51||NA|| |
| Range||32-80||NA||27-93||NA|| |
| ≥50||53||56.9||8||57.2|| |
|ER status|| || || || || |
| Negative||43||46.2||4||28.6|| |
| Unknown||5||5.6||0||0.0|| |
|PR status|| || || || || |
| Negative||47||50.5||3||21.5|| |
| Unknown||7||7.6||1||7.1|| |
|HER2 status|| || || || || |
| Negative||71||76.4||11||78.5|| |
| Unknown||11||11.8||2||14.4|| |
|Triple negative§|| || || || || |
| No||57||61.3||11||78.6|| |
| Unknown||8||8.6||0|| || |
|Grade|| || || || || |
| Intermediate||22||23.6||7||50.2|| |
| High||41||44.3||2||14.2|| |
| Unknown||21||22.5||3||21.4|| |
|Chemotherapy|| || || || || |
| Adjuvant||29||31.3||2||14.4|| |
| Refused||19||20.4||3||21.4|| |
| Unknown||9||9.6||0||0.0|| |
|Radiation therapy|| || || || || |
| Not eligible||15||16.1||1||7.3|| |
| Refused||22||23.6||2||14.2|| |
| Unknown||7||7.7||0||0.0|| |
Stage III breast cancer represents 12.7% of our patients in this case series, and almost ¼ of these patients refused chemotherapy and radiation therapy.
Identifying the reasons for this disparity has been previously reported and continues to be a challenge. In this study, we chose to focus on stage III breast cancer, because this is a group of patients who require multimodal therapy that is essential to enhance survival.
Low socioeconomic status and/or lack of screening may account for late-stage disease and larger tumor size at presentation seen in AA women compared with their Caucasian counterparts.6 Poverty rates are higher in the AA community, where 23.6% live below the US poverty level and 22% lack health insurance, compared with 7.7% and 12% of Caucasian women, respectively according to the US Census. However, socioeconomic disparities do not explain fully the unique patterns of breast cancer observed in AA women. The lower survival in AA women in equal access healthcare systems such as the military, even after adjusting for stage and age, indicates that biologically aggressive disease may have a substantial impact on outcome as well.3
In a recent study of 3314 women6 diagnosed with breast cancer in a large metropolitan area, cancer treatment standards were not adequately met for both the 26.5% African American and the 76.5% Caucasian women. Of particular note, AA women experienced longer treatment delays, regardless of the stage of diagnosis. AA women did not receive appropriate surgical treatment in 7.5% versus 1.5% of Caucasian women. Only 61% of AA versus 72% of Caucasian women received radiation therapy after breast-conserving surgery, and AA women eligible for hormonal therapy were less likely to receive it.
AA women have a younger age distribution compared with Caucasian women at any stage.8 In our study, we found that stage IIIC patients were younger at the time of diagnosis than stage IIIA and IIIB patients.
Race is also a statistically significant predictor of tumor aggressiveness as measured by histologic grade within each stage of breast cancer.9 Many other studies have shown that AA patients with breast carcinoma are more likely to have lymph node positivity, ER− and PR− tumors, and higher grade tumors.10-13 In population-based studies, TNTs have been demonstrated to be most prevalent among younger AA women6, 14 and premenopausal AA women (39%) compared with postmenopausal AA women (14%).15 In the aforementioned studies, the odds of presenting with a TNT were 2-fold or greater for AA women, after adjustment for age and stage. In this study, we showed that almost 30% of these stage III patients had TNTs, and up to 50% of the inflammatory breast cancer patients were triple negative, but overall we did not find a correlation with age or menopausal status.
Research interests about racial/ethnic differences in biological behavior have turned to molecular biology investigations. In recent years, the BP1 gene has been found to be differently expressed in different races. Initially cloned from a leukemia cell line, the BP1 gene is expressed in 89% of breast cancers from AA and only in 57% of cancers in Caucasian women.16, 17 Similarly, p53 mutation or overexpression15, 18 is also more prevalent among AA breast cancers. The challenge in this genetic category of causes for disparity is to develop target therapies for treatment.
In locally advanced breast cancers, the benefit of chemotherapy, surgery, and radiation therapy in improving survival and decreasing recurrence rate are well established.19 Unfortunately, in our cohort nearly ¼ of the patients were noncompliant. It is well known that AA women do not receive adequate chemotherapy dosing because of underlying neutropenia and morbid obesity.20 Approximately a fourth of our patients refused adjuvant/neoadjuvant chemotherapy and chest wall radiation therapy. A higher number of AA subjects refused adjuvant radiation therapy. The reasons for refusal of chemotherapy and lack of radiotherapy need further investigation. We did not find any correlation with age of diagnosis, marital status, and religious background. In a subanalysis by year, we did find that chemotherapy refusal decreased over time, but we did not observe improvements in radiation therapy refusals. Many other factors may be involved in the refusal of standard of care modalities, including socioeconomics, demographics, cultural beliefs, healthcare access, comorbid conditions, and patient choice.
At our institution, to overcome this high refusal rate, we have implemented a community outreach and internal navigational program to assure adherence to standard multimodalities therapy.21 The major role of these combined programs is to dispel fears of treatment and encourage compliance. In a recent audit of the National Quality Forum breast cancer indicators at our institution, we found that patients receiving multimodal therapy (surgery, chemotherapy/hormonal, and radiation treatment) are at risk for not completing all treatments regardless of stage. A recently funded grant is addressing intensive patient navigation for this group of women, ensuring that completion of all modalities occurs. Since January 2007, we have a dedicated nurse practitioner and social worker who carefully follow all patients during chemotherapy and ensure appropriate follow-up with surgical or radiation oncology after completion of medical oncology treatment. We strongly believe that these prospectively implemented interventions based on this and other studies at our center can significantly improve outcome in these advanced breast cancers patients.
The strength of this study is its large number of relatively rare clinicopathologic entities, such as stage III cancers and triple negative cancers, in a single institution. The advantage of a single institution series is the systematic standardization of clinical and pathologic criteria assigned by individual surgeons and pathologists in the classification of locally advanced breast cancer or inflammatory breast cancer.
The major limitations of this study include the lack of long-term follow-up to evaluate overall long-term survival differences. As with many retrospective analyses using tumor registry data and charts instead of prospective data collection, it is possible that staging errors occurred, especially in inflammatory breast cancers.
This study suggests that in AA women there are distinct clinical behaviors associated with stage III breast cancers. The high rate of refusal for important lifesaving therapies that we identified needs further study after implementation of our planned initiatives, so all women are offered and will receive appropriate care based on stage of diagnosis.
Conflict of Interest Disclosures
This work is funded in part by the Avon Foundation (M.R., M.J.L.).