Racial/Ethnic Differences in Pathologic Complete Response to Neoadjuvant Chemotherapy and Survival Among Early-Stage, Erb-B2 Receptor Tyrosine Kinase 2-Low Breast Cancer Patients

Abstract

Racial/ethnic differences in pathologic complete response (pCR), and in overall survival (OS) by pCR status, among early-stage, erb-b2 receptor tyrosine kinase 2 (ERBB2)-low breast cancer patients after neoadjuvant chemotherapy (NACT) are unknown. Data were from the 2010–2020 National Cancer Database that included Asian/Pacific Islander (API), American Indian/Alaska Native/Other (AIANO), Black, Hispanic, and White patients. pCR and OS were modeled using logistic regression and Cox regression, respectively. Of 25,577 patients, Black patients achieved a 17.4% pCR rate, Hispanic 16.0%, White 14.7%, API 13.5%, and AIANO 10.9%. AIANO patients had lower odds of pCR than White patients (adjusted odds ratio, 0.66; 95% confidence interval [CI], 0.48–0.91). Among patients without pCR, API (adjusted hazard ratio [aHR], 0.62; 95% CI, 0.51–0.76) and Hispanic (aHR, 0.77; 95% CI, 0.67–0.89) patients had lower mortality risks than White patients. Among patients with pCR, similar OS rates were observed between Hispanic (aHR, 1.08; 95% CI, 0.66–1.78), Black (aHR, 0.77; 95% CI, 0.55–1.09), API (aHR, 0.41; 95% CI, 0.15–1.12), or AIANO (aHR, 0.35; 95% CI, 0.05–2.50) and White patients. Post-NACT pCR rates were similar across racial/ethnic groups of early-stage, ERBB2-low breast cancer patients. Among patients without pCR, API and Hispanic patients had better OS; among patients with pCR, there was no differential OS by race/ethnicity. Our findings suggest the need for longitudinal studies of OS differences in this patient population.

INTRODUCTION

Human epidermal growth factor receptor-2 (HER2, also known as erb-b2 receptor tyrosine kinase 2; ERBB2) is an important biomarker for breast cancer prognosis and survival [1]. ERBB2-low is defined as an immunohistochemical (IHC) score of 1+ or 2+ with a negative fluorescent in situ hybridization (FISH) test [2] and while it was traditionally classified as ERBB2-0 (i.e., ERBB2-negative), accounting for 50%–60% of ERBB2-negative tumors [3]. While ERBB2-targeted therapies have hitherto been ineffective at improving clinical outcomes of breast cancer patients with low ERBB2 expression [4], a recent randomized controlled trial demonstrated that trastuzumab deruxtecan, an antibody-drug conjugate, is efficacious in ERBB2-low metastatic breast cancer patients [5]. This offers a unique opportunity to extend ERBB2-targeted treatment to a larger ERBB2-low breast cancer patient population in neoadjuvant and adjuvant settings.

Recent studies have reported that early-stage, ERBB2-low breast cancer patients achieve a pathologic complete response (pCR) rate of 13.0%–36.3% after neoadjuvant chemotherapy (NACT); however, they have found contradicting results when comparing invasive disease-free, progression-free, and overall survival (OS) between patients with ERBB2-low tumors and those with HER-negative tumors [6,7,8,9,10,11,12]. Moreover, these study sample sizes are small and include less racially/ethnically diverse cohorts of breast cancer patients. Existing literature has documented racial/ethnic disparities in mortality rates among patients diagnosed with breast cancer by age group, stage, and hormone receptor status, however, ERBB2-low was not assessed [13]. Furthermore, differences in post-NACT pCR between racial/ethnic groups, and in OS by pCR status, among early-stage, ERBB2-low breast cancer patients are currently unknown.

Given the increasing use of NACT in clinical practice [13], it is imperative to explore how pCR rates following NACT and subsequently OS differ across racial/ethnic populations. To address these gaps in the literature, we sought to assess racial/ethnic differences in post-NACT pCR, and in OS by pCR status, among early-stage, ERBB2-low breast cancer patients using a large real-world clinical oncology database.

METHODS

Study design and population

This was a retrospective study of early-stage, ERBB2-low breast cancer patients in the 2010–2020 National Cancer Database (NCDB), a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society and an oncology registry capturing approximately 70% of new U.S. cancer diagnoses annually [14]. Eligibility criteria included ages ≥ 18 years, American Joint Committee on Cancer (AJCC) stage I, II, and III disease, ERBB2-low expression, and record of NACT. Per the NCDB Data Use Agreement, we suppressed frequencies of < 10 to protect the confidentiality of patients. This study was exempt from the University of Chicago Institutional Review Board’s oversight because the NCDB data are de-identified.

Measures

ERBB2-low was defined as an IHC score of 1+ or 2+, with non-amplification by FISH. pCR, measured as the absence of tumor cells in the primary site and associated draining lymph node(s) during histopathologic examination (i.e., ypT0/Tis ypN0), was dichotomized as “achieved” and “did not achieve.” The survival event was vital status, classified as dead or alive. OS was measured from the date of surgery to the date of death or the last patient contact. Note that the vital status information was not available for patients who were diagnosed in 2020 because of limited patient follow-up. Therefore, only the OS outcomes of patients from 2010–2019 were analyzed.

Racial/ethnic groups included Asian/Pacific Islander (API), American Indian/Alaska Native/Other (AIANO), Black, Hispanic, and White. Demographic and clinicopathologic characteristics included age at diagnosis, year of initial diagnosis, type of health insurance, median household income quartile, Charlson-Deyo comorbidity score, AJCC stage group, histologic type, estrogen receptor/progesterone receptor (ER/PR) status, AJCC clinical T/N stage, and tumor grade.

Statistical analysis

Patient characteristics were summarized using descriptive statistics, with distributions compared using analysis of variance, Kruskal-Wallis, or Pearson’s χ² tests as appropriate. Multivariable logistic regression was used to compute adjusted odds ratios (aORs) and 95% confidence intervals (CIs) and to assess racial/ethnic differences in pCR rates after NACT. Potential confounders, including age, year of initial diagnosis, type of insurance, Charlson-Deyo comorbidity score, AJCC stage, histology, ER/PR status, and tumor grade, were adjusted for in the multivariable model. We calculated Kaplan-Meier (K-M) statistics to obtain median survival time (in months) and 95% CI and compared survival functions using log-rank tests. We also reported the estimated probabilities of 5- and 10-year OS. Multivariable Cox proportional hazards regression models were fit to obtain adjusted hazard ratios (aHRs) and 95% CIs and to examine differential OS by race/ethnicity and pCR status. On multivariable analysis, we adjusted for age, Charlson-Deyo comorbidity score, AJCC stage, histology, ER/PR status. tumor grad, and clinical T/N stage. These p-values of < 0.05 were considered statistically significant. All statistical tests were two-sided and were performed using StataSE 17 (StataCorp, College Station, USA).

RESULTS

Patient characteristics

A total of 25,577 early-stage, ERBB2-low breast cancer patients were identified. The mean age was 54.1 years (standard deviation, 12.4). Most (66.0%) were White, followed by 18.0% Black, 9.1% Hispanic, 5.1% API, and 1.8% AIANO. Compared with White patients, other racial/ethnic groups were diagnosed at a younger age and were more likely to be uninsured or on Medicaid. Black patients were more likely to have locally-advanced diseases and high-grade tumors than other racial/ethnic groups (Supplementary Table 1).

pCR

The overall post-NACT pCR rate was 15.2%. Black patients achieved a higher pCR rate (17.4%) than Hispanic (16.0%), White (14.7%), API (13.5%), and AIANO (10.9%) patients. Significant differences in demographic and clinicopathologic characteristics between patients who achieved pCR and those who did not achieve (Supplementary Table 2). In the adjusted logistic regression model, AIANO patients had lower odds of pCR than White patients (aOR, 0.66; 95% CI, 0.48–0.91), whereas there were no significant differences in odds of pCR between Hispanic and White (aOR, 1.12; 95% CI, 0.98–1.29), between Black and White (aOR, 1.01; 95% CI, 0.92–1.12), or between API and White (aOR, 0.95; 95% CI, 0.79–1.14) patients (Table 1).

Table 1. Racial/ethnic differences in pathologic complete response among early-stage erb-b2 receptor tyrosine kinase 2-low breast cancer patients.

Race/Ethnicity		Pathologic complete response (n = 25,577)		Logistic regression
		Did not achieve	Achieved
		Number (row %)	Number (row %)	Crude OR (95% CI)	p-value	aOR* (95% CI)	p-value
Total		21,687 (84.8)	3,890 (15.2)	-	-	-	-
	White	14,408 (85.3)	2,493 (14.7)	1.00 (reference)		1.00 (reference)
	Black	3,797 (82.6)	801 (17.4)	1.22 (1.12–1.33)	< 0.001	1.01 (0.92–1.12)	0.796
	API	1,125 (86.5)	175 (13.5)	0.90 (0.76–1.06)	0.206	0.95 (0.79–1.14)	0.576
	Hispanic	1,957 (84.0)	372 (16.0)	1.10 (0.98–1.24)	0.121	1.12 (0.98–1.29)	0.096
	AIANO	400 (89.1)	49 (10.9)	0.71 (0.52–0.96)	0.024	0.66 (0.48–0.91)	0.012

Note: Bold values represent statistically significant results.

OR = odds ratio; CI = confidence interval; aOR = adjusted odds ratio; API = Asian/Pacific Islander; AIANO = American Indian/Alaska Native/Other.

^*Adjusted for age, year of initial diagnosis, type of insurance, Charlson-Deyo comorbidity score, stage group, histologic type, estrogen receptor status, progesterone status, tumor grade, and nodal involvement.

OS

The K-M OS curves by race/ethnicity and pCR status are presented in Figure 1. Patients who achieved pCR had greater estimated probabilities of 5- and 10-year OS than those who did not achieve pCR (Table 2). On Cox regression analyses, among patients who did not achieve pCR, API (aHR, 0.62; 95% CI, 0.51–0.76) and Hispanic (aHR, 0.77; 95% CI, 0.67–0.89) patients had lower mortality risks than White patients, while there were no significant differences in OS between Black and White (aHR, 1.07; 95% CI, 0.99–1.17) or between AIANO and White (aHR, 0.88; 95% CI, 0.65–1.18) patients. Among patients who achieved pCR, no significant differential OS between Hispanic and White (aHR, 1.08; 95% CI, 0.66–1.78), between Black and White (aHR, 0.77; 95% CI, 0.55–1.09), between API and White (aHR, 0.41; 95% CI, 0.15–1.12), or between AIANO and White (aHR, 0.35; 95% CI, 0.05–2.50) patients were observed (Table 2).

Figure 1. Kaplan-Meier curves of overall survival in early-stage erb-b2 receptor tyrosine kinase 2-low breast cancer patients by race/ethnicity and pathologic complete response status.

pCR = pathologic complete response; API = Asian/Pacific Islander; AIANO = American Indian/Alaska Native/Other.

Table 2. Racial/ethnic differences in overall survival among early-stage erb-b2 receptor tyrosine kinase 2-low breast cancer patients by pathologic complete response status.

Race/Ethnicity		Median survival time in mo* (95% CI)	p-value†	5-yr OS* (95% CI)	10-yr OS* (95% CI)	Crude HR (95% CI)	aHR‡ (95% CI)
Among patients who did not achieve pCR
	White	NR	< 0.001	0.81 (0.80–0.82)	0.66 (0.65–0.68)	1.00 (reference)	1.00 (reference)
	Black	NR		0.76 (0.74–0.78)	0.61 (0.57–0.64)	1.26 (1.16–1.36)∥	1.07 (0.99–1.17)§
	API	NR		0.89 (0.87–0.91)	0.77 (0.70–0.82)	0.57 (0.47–0.69)∥	0.62 (0.51–0.76)∥
	Hispanic	NR		0.85 (0.83–0.87)	0.76 (0.71–0.80)	0.73 (0.64–0.84)∥	0.77 (0.67–0.89)∥
	AIANO	NR		0.84 (0.79–0.89)	0.70 (0.57–0.80)	0.76 (0.57–1.02)	0.88 (0.65–1.18)
Among patients who achieved pCR
	White	NR	0.167	0.92 (0.90–0.93)	0.84 (0.81–0.88)	1.00 (reference)	1.00 (reference)
	Black	NR		0.93 (0.91–0.95)	0.87 (0.81–0.92)	0.84 (0.60–1.16)	0.77 (0.55–1.09)
	API	NR		0.99 (0.95–0.99)	0.82 (0.39–0.96)	0.38 (0.14–1.03)	0.41 (0.15–1.12)
	Hispanic	NR		0.93 (0.88–0.96)	0.89 (0.81–0.94)	0.83 (0.51–1.33)	1.08 (0.66–1.78)
	AIANO	NR		0.97 (0.78–0.99)	0.97 (0.78–0.99)	0.30 (0.04–2.18)	0.35 (0.05–2.50)

Note: Bold values represent statistically significant results.

OS = overall survival; CI = confidence interval; HR = hazard ratio; aHR = adjusted hazard ratio; pCR = pathologic complete response; API = Asian/Pacific Islander; AIANO = American Indian/Alaska Native/Other; NR = not reached.

^*Median OS time and 5-year and 10-year OS were estimated using the Kaplan-Meier method.

^†The p-value was calculated using the log-rank test.

^‡Adjusted for age, Charlson-Deyo comorbidity score, American Joint Committee on Cancer stage group, histologic type, estrogen status, progesterone status, tumor grade, clinical T stage, and nodal involvement.

^§p < 0.1; ^∥p < 0.001.

DISCUSSION

In this study, we utilized a national cohort of early-stage, ERBB2-low breast cancer patients to examine whether post-NACT pCR rates and OS by pCR status differed across racial/ethnic groups. Contrasting from previous studies with smaller sample sizes containing racially/ethnically homogenous cohorts [6,7,8,9,10,11,12], our study had greater power to assess differences in prognoses between ERBB2-low patients of different racial/ethnic groups. Additionally, to our knowledge, the current study is the first and largest to describe these differential prognoses in this patient population from a real-world clinical oncology registry.

We found an overall pCR rate of 15.2% in early-stage, ERBB2-low breast cancer patients having achieved pCR, consistent with previous studies [6,9,10,12]. However, these studies did not assess racial/ethnic differences in post-NACT pCR rates, whereas our study revealed that more than one in 6 Black patients achieved pCR, followed by more than one in 7 Hispanic or White patients, more than one in 8 API patients, and approximately one in 9 AIANO patients did. The data suggested that racial/ethnic minority patients achieved similar pCR rates as their White counterparts. However, after adjusting for potential confounders, compared with White patients, only AIANO patients had significantly lower odds of having achieved pCR. One possible contributing factor to this disparity is that since AIANO women tend to be diagnosed with breast cancer at a younger age compared with women of other races/ethnicities [15] and with White women in our study cohort (Supplementary Table 1), AIANO patients presenting may be at more advanced stages, thereby leading to a lower pCR rate.

Furthermore, we discovered that among early-stage, ERBB2-low breast cancer patients who did not achieve pCR following NACT, API and Hispanic patients had better OS compared to their White counterparts, which is consistent with prior research on racial/ethnic disparities in survival outcomes among patients with early-stage disease or triple-negative breast cancer [16,17,18]. Among patients who did achieve pCR, no differential OS between racial/ethnic minority patients and White patients was found. Our findings suggest that racial/ethnic disparities exist among patients not achieving pCR after NACT in the neoadjuvant setting. While our study is exploratory in nature, these findings warrant future research into examining other important factors that are not collected by the NCDB, such as access to health care, behavioral measures, and additional socioeconomic status indicators, that potentially contribute to these differences in OS across racial/ethnic groups in this patient population.

Several limitations should be noted. First, this analysis may have encountered unidentified biases and the NCDB is a clinical database likely being subject to reporting errors. Therefore, prospective cohort studies are needed to confirm our results. Second, the NCDB does not collect data on the specific regimen of NACT administered. Thus, we could not evaluate the differential effects of various NACT regimens. Third, OS is a reliable measure of survival but is not specific to breast cancer. Because the NCDB does not collect cause-specific survival information, survival endpoints associated with breast cancer diagnoses were not assessed. Lastly, although the NCDB is a national registry, it is not population-based, i.e., breast cancer patients may not be representative of all U.S. breast cancer patients, which limits the generalizability of our findings.

In conclusion, in this national cohort of early-stage, ERBB2-low breast cancer patients, racial/ethnic minority patients had similar post-NACT pCR rates as White patients, though relatively higher in Black patients. AIANO patients showed significantly lower odds of pCR. Among patients who did not achieve pCR, API and Hispanic patients had better OS; among patients who achieved pCR, there was no differential OS by race/ethnicity. Our findings suggest further investigation of possible causes of racial/ethnic OS differences after NACT and the need for longitudinal studies of OS in this patient population.

SUPPLEMENTARY MATERIALS

Supplementary Table 1

Demographic and clinicopathologic characteristics of early-stage erb-b2 receptor tyrosine kinase 2-low breast cancer patients by race/ethnicity

Supplementary Table 2

Demographic and clinicopathologic characteristics of early-stage erb-b2 receptor tyrosine kinase 2-low breast cancer patients by pathologic complete response status