Impact of Race, Socioeconomic Status, and Clinicopathologic Features on Clinical Outcomes in Triple-Negative Breast Cancer in the ECOG-ACRIN EA1131 Trial

Abstract

A higher incidence of triple negative breast cancer (TNBC) and worse social determinants of health (SDOH) both contribute to racial survival disparities among Black women diagnosed with breast cancer. A post hoc analysis of the ECOG-ACRIN EA1131 study (adjuvant platinum versus capecitabine in stage II-III TNBC with residual disease after neoadjuvant chemotherapy) evaluated racial disparities in disease free survival (DFS) and overall survival (OS) among Black and White patients. Of 415 patients enrolled in EA1131, 376 were included in this analysis (308 White [82%] and 68 Black [18%]). Common characteristics included basal-subtype TNBC (77%), grade 3 disease (71%), residual stage II disease (49%), private insurance (70%), and a BMI ≥30 (49%). There were no racial differences in grade, clinical or pathological stage. Black patients were more likely to have basal-subtype TNBC (89% vs 75%; p=0.009), a BMI ≥30 (62% vs 46%; p=0.026), to reside within the lowest neighborhood socioeconomic index (nSES) quartile (39% vs 22%; p=0.008), and have Medicaid (32% vs 13%; p<0.001) compared to White patients. Despite these differences, there were no significant differences in DFS or OS by race (HR 0.99, 95% CI: 0.62-1.57 and HR 0.60, 95% CI: 0.32-1.12, respectively) among Black and White patients.

Introduction

While breast cancer mortality rates have significantly decreased, the rate of decline for Black patients has been slower, resulting in a 38% higher mortality rate compared to White patients.^1–7 Factors contributing to worse survival outcomes among Black breast cancer patients include a higher incidence of triple negative breast cancer (TNBC) and basal-like phenotype,^{6, 8–11} advanced stage at diagnosis,^1,2,11 more comorbidities,^1,6,12 and disparities in care and social determinants of heath (SDOH).^{6,11,13, 14} TNBC is an aggressive subtype associated with higher risk of recurrence and death.¹⁵

The ECOG-ACRIN EA1131 trial was a randomized phase III postoperative trial to determine whether invasive disease-free survival (IDFS) would be improved with adjuvant platinum-based therapy versus capecitabine in patients with residual TNBC following neoadjuvant chemotherapy (NAC).¹⁶ We performed a post-hoc analysis to evaluate whether racial disparities in clinicopathological features, basal-subtype, and SDOH impacted survival outcomes among patients enrolled in EA1131.¹⁶

Methods

The details of ECOG-ACRIN EA1131 (NCT02445391) trial protocol and results have been previously reported.¹⁶ We present a secondary analysis of Black and White patients in the EA1131 trial. Race was self-reported; Racial groups with <15 patients or unknown race were excluded from analysis (n=39) (Supplement 1). Baseline data included PAM50 intrinsic breast cancer subtype, tumor grade, pretreatment clinical and post-NAC pathologic stage, insurance type, body mass index (BMI), DFS, and overall survival (OS). The Agency for Healthcare Research Quality (AHRQ) neighborhood socioeconomic index (nSES) was calculated using residential zip codes linked to county level data on occupation, income, poverty, wealth, education, and crowding. Higher nSES signifies lower neighborhood deprivation. Our primary objective was to examine racial disparities in biological and non-biological factors and survival outcomes among patients in EA1131.

Statistical methods.

Binary and categorical data were analyzed with Fisher’s exact test to evaluate differences in baseline characteristics by race. Cox regression modeling was used to estimate hazard ratios (HR) with 95% confidence intervals (CI) for DFS and OS when adjusting for race, treatment received, intrinsic subtype, grade, clinical stage, pathologic stage, nSES, BMI, and insurance. DFS and OS were estimated with the Kaplan-Meier method. Log-rank test was used to test for group differences.

Results

Of the patients enrolled in EA1131, 376 were included (308 White [82%] and 68 Black [18%]) (Table 1). Common characteristics included basal subtype TNBC (77%), grade 3 disease (71%), private insurance (70%), and BMI ≥30 (49%). There were no racial differences in grade, pretreatment clinical stage or post-NAC pathologic stage (Table 1). Black patients were more likely to have basal subtype TNBC (89% vs 75%; p=0.009), BMI ≥30 (62% vs 46%; p=0.026), Medicaid insurance, a government-funded health insurance program in the United States, (32% vs 13%; p=<0.001), and to live in a neighborhood in the lowest nSES quartile (39% vs 22%; p=0.008). White patients were more likely to have private insurance (73% vs 57%; p=0.013).

Table 1.

Clinical Characteristics of study group by race

	Total	White – n (%)	Black – n (%)	p-value**
N = 376	376	308 (82%)	68 (18%)
Intrinsic subtype
Basal	282 (77%)	223 (75%)	59 (89%)	0.009
Non-basal	82 (23%)	75 (25%)	7 (11%)
Missing	12	10	2
Grade
1	6 (2%)	6 (2%)	0 (0%)	0.21
2	73 (20%)	65 (21%)	8 (12%)
3	266 (71%)	212 (69%)	54 (79%)
Unknown	29 (7%)	23 (8%)	6 (9%)
Missing	2	2	0
Clinical Stage*
II	247 (66%)	203 (66%)	44 (65%)	0.888
III	129 (34%)	105 (34%)	24 (35%)
Missing	0	0	0
Pathologic Stage
I	88 (23%)	71 (23%)	17 (25%)	0.694
II	181 (49%)	146 (48%)	35 (51%)
III	104 (28%)	88 (29%)	16 (24%)
Missing	3	3	0
Insurance Type
Private	264 (70%)	225 (73%)	39 (57%)+	0.003
Medicarea	21 (6%)	18 (6%)	3 (5%)
Medicaida	61 (16%)	39 (13%)	22 (32%)++
Others	30 (8%)	26 (8%)	4 (6%)
Missing	0	0	0
Socioeconomic Status Quartile
1	94 (25%)	68 (22%)	26 (39%)^	0.001
2	93 (25%)	73 (24%)	20 (30%)
3	94 (25%)	78 (25%)	16 (24%)
4	93 (25%)	88 (29%)	5 (7%)
Missing	2	1	1
Body Mass Index (BMI)
Mean	N/A	30.3 (SD = 7.6)	33.1 (SD = 7.1)	0.008
<25	78 (23%)	68 (24%)	10 (16%)
25-29	98 (28%)	84 (30%)	14 (22%)	0.067
≥30	171 (49%)	131 (46%)	40 (62%)&
Missing	29	25	4

^*EA1131 only enrolled patients with clinical stage II or III disease.

^**Denotes p-value comparing the clinical characteristic as a whole between White versus Black patients.

⁺P-value of difference between Black vs White patients with private insurance was 0.013.

⁺⁺P-value of difference between Black vs White patients with Medicaid was <0.001.

^{^}P-value of difference between Black and White patients in SES quartile 1 was 0.008.

^&P-value of difference between Black vs White patients with BMI ≥30 was 0.026.

^aMedicare and Medicaid are public government-funded health insurance programs in the United States

Multivariate analysis revealed patients with residual stage II and III disease had significantly worse DFS and OS compared with residual stage I. Patients with stage II and III disease were more likely to have disease progression (HR 1.94, 95% CI 1.09-3.45 and HR 3.87, 95% CI 2.11-7.10, respectively) (Table 2). Patients with residual stage II had numerically inferior OS compared to residual stage I, not statistically significant. Patients with residual stage III disease had worse OS when compared to residual stage I disease (HR 4.67; 95% CI 2.09-10.43, p=0.000).

Table 2.

Cox regression model for DFS and OS

	Disease free survival (DFS)		Overall survival (OS)
	Hazard Ratio (95% CI)	p-value	Hazard Ratio (95% CIl)	p-value
Race
White	1.00 [Reference]		1.00 [Reference]
Black	0.99 (0.62-1.57)	0.967	0.60 (0.32-1.12)	0.110
Treatment Type
Platinum agent	1.00 [Reference]		1.00 [Reference]
Capecitabine	1.05 (0.73-1.50)	0.799	1.25 (0.80-1.95)	0.326
Intrinsic Subtype
Non-basal	1.00 [Reference]		1.00 [Reference]
Basal	1.55 (0.94-2.57)	0.087	1.39 (0.75-2.59)	0.293
Grade
1	1.00 [Reference]		1.00 [Reference]
2	0.96 (0.22-4.23)	0.952	0.59 (0.13-2.76)	0.507
3	1.63 (0.39-6.89)	0.507	0.99 (0.23-4.32)	0.994
unknown	1.36 (0.28-6.52)	0.704	0.87 (0.16-4.72)	0.876
Clinical Stage
II	1.00 [Reference]		1.00 [Reference]
III	1.36 (0.90-2.06)	0.143	1.17 (0.68-2.01)	0.563
Pathologic Stage
I	1.00 [Reference]		1.00 [Reference]
II	1.94 (1.09-3.45)	0.024	1.92 (0.88-4.18)	0.100
III	3.87 (2.11-7.10)	0.000	4.67 (2.09-10.43)	0.000
BMI
<25	1.00 [Reference]		1.00 [Reference]
25-29	1.58 (0.93-2.70)	0.091	1.58 (0.80-3.13)	0.186
≥30	1.16 (0.70-1.90)	0.569	1.52 (0.81-2.85)	0.195
Unknown	1.92 (0.71-5.24)	0.201	1.26 (0.28-5.69)	0.764
Insurance Type
Private	1.00 [Reference]		1.00 [Reference]
Medicare	1.48 (0.70-3.16)	0.308	1.44 (0.58-3.60)	0.43
Medicaid	1.48 (0.92-2.38)	0.107	1.37 (0.76-2.47)	0.296
Other	1.22 (0.60-2.49)	0.586	1.23 (0.53-2.85)	0.632
Socioeconomic Status Quartile
1	1.00 [Reference]		1.00 [Reference]
2	0.89 (0.54-1.47)	0.644	0.92 (0.48-1.76)	0.806
3	0.88 (0.53-1.48)	0.634	1.16 (0.61-2.19)	0.651
4	1.07 (0.64-1.80)	0.795	0.99 (0.52-1.88)	0.972

The median DFS for White patients was 42.6 months versus 25.1 months for Black patients which was not significant (p=0.24) (Figure 1A). An ad hoc analysis showed a significant difference in DFS by race after 18 months (HR 3.72; 95% CI: 1.34-10.33; p=0.012), while there was no difference in DFS within the first 18 months (HR 0.81; 95% CI: 0.49-1.35; p=0.427). There was no significant difference (p=0.91) in median OS between White patients (54 months) and Black patients (not yet reached) (Figure 1B). There were also no significant differences in DFS or OS by race (HR 0.99, 95% CI: 0.62-1.57 and HR 0.60, 95% CI: 0.32-1.12, respectively) (Table 2).

Figure 1.

Kaplan-Meier curve of (A) disease-free survival (DFS) relative to race and (B) overall survival (OS) relative to race.

Discussion

In the EA1131 trial, Black patients with residual TNBC were more likely to have basal-subtype disease, higher BMI, be under- or uninsured, and to live in a neighborhood with a low nSES compared to White patients. Despite this racial disparity in baseline biologic and non-biologic factors, there was no significant difference in DFS and OS by race. Our results were consistent with others demonstrating a racial difference in breast cancer intrinsic subtypes in TNBC.^{5, 17–19} Similar to prior studies,^{10, 20–22} pathologic stage after NAC was directly correlated to survival. We did not observe racial disparity in extent of residual disease or a differential impact of pathologic stage on DFS which has been previously reported.²³ The lack of racial disparity we found in survival may in part be attributed to enrollment in a clinical trial which is usually associated with less variability in care and improved survival.^24,25 Although Black patients represent approximately 12% of new breast cancer diagnoses in the United States,^26,27 they comprise only about 3% of participants in breast cancer clinical trials supporting FDA approvals,^28,29 reflecting the lowest enrollment rates among racial and ethnic groups, even in TNBC-focused studies.^28,30 Given the importance of racial diversity for developing effective therapies across diverse populations, targeted strategies – including broadened eligibility criteria, representative enrollment targets, expanded trial access in community settings, culturally tailored education, and diversification of clinical trial teams – are essential to improve participation among underrepresented populations.^28,31,32

As has been shown,³³ Black patients in EA1131 were more likely to be in a lower nSES quartile than White patients, however there was no association between nSES and mortality. A retrospective study demonstrated patients within the lowest nSES group were least likely to receive guideline-concordant breast cancer care and more likely to die from their disease.³⁴ Taken together, these findings suggest enrollment in a clinical trial may mitigate the effects of SDOH on outcomes. Underinsured or government-insured patients often present with more advanced stage disease³⁵ and have worse cancer-specific mortality.³⁶ Conversely, private insurance status is associated with chemotherapy completion,³³ achieving pCR, and improved mortality.¹⁰ Similar to prior studies, Black patients were more likely to have Medicaid and White patients were more likely to have private insurance,^15,36 however this had no association with survival outcomes. Obesity has been shown to be associated with worse outcomes in patients with aggressive intrinsic subtypes.¹⁸ We observed a higher incidence of obesity among Black patients, as have others^9,19,37,38 but did not demonstrate an association between BMI and inferior outcomes.^39,40 Our lack of association between BMI and inferior outcomes may have been limited due to our small sample size.^39,40

Since publication of the primary results of ECOG-ACRIN EA1131, the standard of care for early-stage, high-risk TNBC has shifted to neoadjuvant chemoimmunotherapy based on the benefit demonstrated in KEYNOTE-522.⁴¹ The addition of neoadjuvant pembrolizumab to chemotherapy significantly improves pathologic complete response (pCR) rates and event-free survival, though at the cost of increased toxicity and financial burden. However, real-world data highlight persistent disparities. National Cancer Database analyses show that Black and Hispanic patients and those with public or no insurance are less likely to receive neoadjuvant chemoimmunotherapy,⁴² and Black patients experience worse overall survival, likely reflecting clinicopathologic factors and social determinants of health rather than treatment efficacy.⁴³ These findings emphasize the need for equitable access to novel therapies and continued real-world evaluation to ensure that the benefits of neoadjuvant chemoimmunotherapy are realized across all eligible patient populations.

Strengths and Limitations

To our knowledge, this is the first study evaluating racial disparities of clinicopathological variables and their impact on survival outcomes in a prospective trial of high-risk TNBC patients with residual disease. This study included a more generalizable representation of Black patients (18%) in the US population compared to several trials.⁴⁴ However, our study was limited by the small sample size of Black patients (68), which limits the statistical power to detect meaningful differences in outcomes by race. We performed an ad hoc analysis that is purely exploratory in nature to investigate the splitting of the DFS curve around 18 months which did not detect a significant difference possibly due to the small sample size. Furthermore, these results may not be generalizable outside the context of a clinical trial, where there are established racial disparities in care¹⁵ and patients with more comorbidities.⁴⁵ As an unplanned post hoc secondary analysis, this study is inherently susceptible to bias, multiple testing, and residual confounding and was not powered for the outcomes evaluated. Accordingly, the findings should be interpreted as exploratory and hypothesis-generating rather than definitive.

Conclusion

Black patients with residual TNBC enrolled in EA1131 were more likely to have aggressive basal-subtype tumors, to be obese, and to have lower socioeconomic status. Despite these factors, no significant differences in survival by race were observed. While this analysis was limited by sample size, one possibility is enrollment in a clinical trial, by minimizing differences in treatment received, has the potential to mitigate racial inequities in care. Larger prospective studies are necessary for confirmation.