Differences in breast tumor response to neoadjuvant chemotherapy by race – Is obesity the key?

Abstract

Introduction: Breast cancer treatment includes neoadjuvant chemotherapy offered to patients with aggressive subtypes and locally advanced breast cancer. Black women receive neoadjuvant chemotherapy more frequently as they present more advanced stage tumors and triple negative subtype. Non-Hispanic Black women have the greatest prevalence of obesity in most states. Patients with higher body mass index (BMI) have lower rates of chemotherapy response. Whether obesity is a confounding factor requires investigation.

Methods: A retrospective cohort study of patients diagnosed with non-metastatic breast cancer who completed neoadjuvant chemotherapy and had surgery at The Ohio State University Comprehensive Cancer Center between January 1 2005 and December 31 2019, was conducted. Demographic characteristics were compared using Pearson’s chi-square test, or Wilcoxon rank-sum test. Logistic regression models were used to assess associations between pathologic response and age, race, BMI, menopausal status, and insurance status.

Results: More Black women had high grade tumors, triple negative advanced stage disease, and higher BMI. Race and BMI were not significant predictors of poor pathologic complete response, however patients aged <40 years had higher odds of pathologic complete response.

Discussion: Further exploration evaluating differences in pathologic response by BMI could lead to a better understanding of the association between obesity and treatment response.

KEYWORDS:

• Breast cancer
• neoadjuvant chemotherapy
• obesity
• racial disparities

Introduction

Among women, breast cancer is the most commonly diagnosed cancer [1,2]. Higher incidence rates are observed in Non-Hispanic White women (NHW) relative to Non-Hispanic Black (NHB) women for most age groups [2]. However, NHW women have a lower mortality rate compared to NHB women and are diagnosed at a less advanced stage of disease [3,4]. Some of the contributing factors to these reported differences include late stage at diagnosis, more aggressive tumors, and obstacles to timely diagnosis and treatment for NHB women [4]. These factors are also associated with lower use of neoadjuvant chemotherapy (NAC).

Patients with advanced localized breast cancer and those who might benefit from tumor size reduction prior to conservation therapy are recommended NAC [5–7]. NAC is more likely to be administered to women with triple-negative breast cancer and those with HER-2 overexpressing tumors [5,6]. Pathological complete response (pCR) is used by researchers and clinicians to assess therapy response to NAC, determine prognosis, and make decisions on surgical therapy [8]. This allows for the stoppage of ineffective treatment on a group level in research and gives researchers the chance to evaluate the efficacy of novel therapeutic treatment [5,7]. NAC is more frequently administered to Asian, Hispanic, and NHB women than to NHW women. The difference is due to a higher percentage of these women's tumors being triple negative [7,9]. However, pCR for HER2-positive and triple-negative breast cancer is less likely for NHB women. There are few research studies that have looked at how obesity influences the effectiveness of NAC in treating primary breast tumors. In a retrospective study conducted in Saudi Arabia, researchers conducted univariate and multivariate analyses of BMI and pCR and found a negative association between the categories overweight and obese and pCR. These patients experienced poorer outcomes compared to those in the normal weight category [10]. Compared to NHW women, NHB women were not as likely to achieve pCR for triple negative and HER2 + tumors [11]. NHB women who do not have pCR following NAC, regardless of subtype, are at a greater risk for mortality [12]. Considering the disparities in NAC outcomes presented in the literature, it is important to investigate whether this racial difference is associated with clinical outcomes including tumor response.

This study investigated racial differences in pCR to NAC in breast cancer. We evaluated whether Black women with breast cancer have lower pCR in comparison with White women and whether obesity is associated with chemotherapy response.

Materials & methods

We conducted a retrospective review utilizing medical records of patients diagnosed with non-metastatic breast cancer who completed NAC and had surgery at Ohio State University James Comprehensive Cancer Center between January 1 2005, and December 31 2019. The study was approved by the Ohio State University Cancer Review Board. Data for qualified patients were downloaded into REDCap from The Ohio State University Information Warehouse [11, 13]. Each patient's electronic medical record was manually reviewed to fill any missing gaps from the initial query. Clinical stage (1–3) was calculated based on tumor size and nodal status. Estrogen receptor (ER), progesterone receptor (PR) and HER2 protein expression are evaluated by manual quantitative immunohistochemistry on formalin-fixed, paraffin-embedded tissue. For ER, the percentage of positive tumor cell nuclei is scored as <1% negative, 1–10% low positive or >10% positive and the overall staining intensity is categorized as weak, moderate or strong. For PR, the percentage of positive tumor cell nuclei is scored as <1% negative or >/ = 1% positive. HER2 protein expression is evaluated using FDA approved kit and scored according to the College of American Pathologists (CAP) criteria. Results of the operative treatment received was recorded to determine pathologic stage and chemotherapy response. The study endpoint, pCR, was assessed after definitive surgery based on the final pathology report. Response to NAC is categorized as non-response, partial response, and complete response. The definition of complete pathologic response used is the absence of invasive cancer in the breast and axillary lymph (ypT0 ypN0) BMI categories were defined based on World Health Organization classification where obesity is defined as 30 kg/m² [12]. For the data analysis, we included self-reported Black and White women, excluding patients classified as ‘Other’ race.

Summary statistics were used to describe patient and clinical characteristics including race, ethnicity, age, menopausal status, employment, insurance status, initial stage at diagnosis, tumor grade, estrogen receptor status (ER) and progesterone receptor status (PR), human epidermal growth factor (HER2) status, surgeries for breast cancer, radiation history, adjuvant chemotherapy, hormonal therapy, and targeted therapy. Differences between race and demographic characteristics were compared using Pearson’s chi-square test for categorical variables and t-test or Wilcoxon rank-sum test for continuous variables, where appropriate. Univariate and multivariable logistic regression models were used to assess associations between pCR and age, race, BMI, menopausal status, insurance status and employment status. Data was analyzed using SAS 9.4.

Results

The final sample of 882 patients included 11.7% Black women and 88.3% White women (Table 1). The majority of the participants (99%) identified as non-Hispanic and 1% identified as Hispanic. The median age of diagnosis among both Black and White women was 51 years. The median BMI of the study sample was 28.6 kg/m². Overall, the median follow-up duration was 147.4 months. Seventy two percent of women had private insurance, 24% had public insurance, and 4% were uninsured. Overall, 64% of the women had clinical stage 2 disease, 22% were triple negative, 62% HER-2 positive. Treatment types included surgery, hormonal therapy, and radiation. Overall, 52% of women received a mastectomy, 29% received a lumpectomy, 18% had reconstruction surgery, and 1% had no or unknown surgery type. Just over half of the total (52%) did not receive adjuvant hormonal therapy.

Table 1. Demographic and clinicopathological characteristics summary by race.

Variable	White (n = 779)	Black (n = 103)	Total (n = 882)	P-value
Age at diagnosis, median [IQR]	51 [41, 59]	51 [42, 60]	51 [41, 59]	0.850
Ethnicity				0.415
Hispanic	5 (1%)	0 (0%)	5 (1%)
Non Hispanic	774 (99%)	103 (100%)	877 (99%)
BMI, Median [IQR]	28.4 [24.3, 34]	31.5 [26.7, 35.2]	28.6 [24.5, 34.4]	0.001*
BMI				0.002*
BMI <30	454 (58%)	43 (42%)	497 (56%)
BMI ≥30	325 (42%)	60 (58%)	385 (44%)
Employment status				0.024
Missing	1 (0%)	0 (0%)	1 (0%)
Employed	645 (83%)	76 (74%)	721 (82%)
Unemployed	133 (17%)	27 (26%)	160 (18%)
Insurance status				<0.001*
Missing	1 (0%)	0 (0%)	1 (0%)
Private	595 (76%)	40 (39%)	635 (72%)
Public	163 (21%)	47 (46%)	210 (24%)
Uninsured	20 (3%)	16 (16%)	36 (4%)
Menopausal status				0.676
Premenopausal	351 (45%)	42 (41%)	393 (45%)
Postmenopausal	380 (49%)	55 (53%)	435 (49%)
Unknown	48 (6%)	6 (6%)	54 (6%)
Grade				0.189
High grade	456 (59%)	69 (67%)	525 (60%)
Intermediate grade	241 (31%)	27 (26%)	268 (30%)
Low grade	37 (5%)	1 (1%)	38 (4%)
Unknown	45 (6%)	6 (6%)	51 (6%)
Stage				0.399
1	109 (14%)	13 (13%)	122 (14%)
2	504 (65%)	62 (60%)	566 (64%)
3	166 (21%)	28 (27%)	194 (22%)
ER status				0.086
Negative	376 (48%)	59 (57%)	435 (49%)
Positive	403 (52%)	44 (43%)	447 (51%)
PR status				0.101
Negative	464 (60%)	70 (68%)	534 (61%)
Positive	315 (40%)	33 (32%)	348 (39%)
HER2 status				0.070
Negative	291 (37%)	48 (47%)	339 (38%)
Positive	488 (63%)	55 (53%)	543 (62%)
Subtype				0.061
ER+/PR+/HER2- & ER+/PR-/HER2-	122 (16%)	17 (17%)	139 (15%)
ER-/PR-/HER2-	164 (21%)	31 (30%)	195 (22%)
HER2+	493 (63%)	55 (53%)	548 (62%)
Pathological complete response				0.962
No	520 (67%)	69 (67%)	589 (67%)
Yes	259 (33%)	34 (33%)	293 (33%)

Abbreviations: BMI, body mass index; ER status, estrogen receptor status; PR status, progesterone receptor status; HER2, human epidermal growth factor receptor 2.

*Indicates statistical significance at p < 0.05.

Variables found to be associated with race included employment status, insurance status and BMI (Table 1). A higher proportion of White women were employed and had private insurance compared to Black women (83% vs. 74%, p- = 0.024; 76% vs. 39%, p < 0.001) respectively, while the median BMI was higher among Black women (31.5 kg/m²) than White women (28.6 kg/m²) (p = 0.001).

Only age was significantly associated with pCR. The odds of pCR in patients 40 years old was 1.6 times higher than those ≥40 years, both univariately [OR 95% CI 1.1–2.2] and after adjusting for the other variables in the model [OR 95% CI 1.1–2.4], p = 0.011 and p = 0.012, respectively. No associations were detected between pCR and race, BMI, menopausal status, insurance status, or employment status (Table 2).

Table 2. Univariable and multivariable predictors for pathological complete response.

Patient characteristic	Univariable analysis			Multivariable analysis
	OR	95% CI	P value	OR	95% CI	P value
Age at diagnosis
< 40 years old vs ≥ 40 years old (ref)	1.6	(1.1, 2.2)	0.011*	1.6	(1.1, 2.4)	0.012*
Race
White vs Black (ref)	1.0	(0.7, 1.6)	0.961	1.0	(0.6, 1.6)	0.921
BMI
BMI < 30 vs BMI ≥ 30 (ref)	1.1	(0.9, 1.5)	0.395	1.1	(0.8, 1.5)	0.522
Menopausal status
Postmenopausal vs Premenopausal (ref)	0.9	(0.7, 1.2)	0.608	1.1	(0.8, 1.5)	0.563
Insurance status
Public vs Private (ref)	0.9	(0.7, 1.3)	0.601	1.0	(0.7, 1.4)	0.956
Uninsured vs Private (ref)	1.3	(0.6, 2.5)	0.498	1.4	(0.7, 2.8)	0.413
Employment status
Employed vs Unemployed (ref)	1.2	(0.9, 1.8)	0.264	1.3	(0.9, 1.9)	0.225

Abbreviations: OR, odds ratio; BMI, body mass index.

*Indicates statistical significance at p < 0.05.

Discussion

In this retrospective analysis of institutional data, we found no difference in treatment response, measured by pCR, between NHW and NHB women who received NAC. However, we found significant differences in socioeconomic and demographic variables among between NHW and NHB patients who received NAC. These included NHB patients more likely to be unemployed, less likely to have private insurance, higher median BMI, and a higher proportion of women in the obesity range.

NHB women have the highest prevalence for obesity compared to NHW women [2]. Obesity is a potentially adverse prognostic factor for breast cancer. Patients with a higher BMI have been shown to have lower rates of chemotherapy response [10]. NHB women are treated with NAC more often as they frequently present with further developed stage tumors and the triple negative subtype [9]. It is important to understand the link between race, breast cancer subtype, and response to NAC to understand long-term survival effects of racial disparities in pCR and reduce the growing mortality gap between NHW and NHB women [14].

Chemotherapy response in difference racial groups

Differences in chemotherapy response have been linked to biological tumor response differences and failure to complete treatment [9,11,12]. From a clinical standpoint, failure to complete treatment is higher in Black patients and is associated with psychosocial and socioeconomic barriers, as well as high rates of medical comorbidities. Black breast cancer patients were less likely to complete NAC compared to White breast cancer patients [11]. Unlike prior reports, we also did not observe a difference in response by different breast cancer subtypes in this sample. A prospective analysis completed at the University of North Carolina Lineberger Comprehensive Cancer Center examined women with stage 2–3 non-metastasized breast cancer treated with NAC from 1998 to 2011. Overall survival rate and the pCR rate by race were not significantly different [15]. There were no differences in chemotherapy sensitivity between ‘African American and non-African American women’ [15]. Other studies, however, show contrasting results. A study completed within the National Cancer Data Base analyzed differences in pCR rates by rates among women with stage 1–3 breast cancer diagnosed between 2010 and 2011. Researchers found that compared to White women, Black women had lower rates of pCR for HER2-positive and triple-negative tumors [9]. It is unclear if this occurs because of treatment-related physiologic differences, chemosensitivity, or the inability to adjust for differences in socioeconomic factors [6,9].

Association between obesity and chemotherapy response

In this study, we found that Black patients had higher BMI compared to White women, and there was no association between BMI and pCR. Obesity has been shown to negatively affect outcomes after breast cancer diagnosis, and obesity, measured using BMI, is viewed as a poor prognostic factor [16,17]. There are few studies on how obesity impacts NAC response for primary breast tumor treatment. In a cohort study of patients treated with NAC at the University of Texas M.D. Anderson Cancer Center analyzing the relationship between BMI and pCR, researchers found patients with higher BMI had a lower likelihood of pCR after treatment. They found that obese patients had a lower overall survival rate compared to underweight or normal weight patients, which is consistent with existing literature [18,19]. Researchers are uncertain of the biological mechanism obesity may contribute to breast cancer prognosis. Obesity may modify medications and impact patient response to chemotherapy [18,20].

A meta-analysis reviewing the impact of BMI on pCR reported obese and overweight patients have lower pCR rate with NAC compared to underweight and normal weight patients [21].

Existing data regarding obesity status and chemotherapy completion reflects that the efficacy of cancer treatment may be lower in obese breast cancer survivors compared to non-obese patients upon treatment completion This poses a challenge in disease management and patient care among obese patients [16]. Understanding NAC tumor response may help us understand how obesity may influence pCR and breast cancer prognosis.

Differences in employment status and insurance status

Prior research indicates insurance status has a significant impact on timely treatment. Insurance status is linked to employment status as many private insurance holders are covered via employee sponsored private health insurance. In a retrospective review conducted from 2006 to 2018 at Loyola University Medical Center, researchers found that compared to patients with private insurance, patients with public insurance (Medicare) were observed to have lower odds of pCR or any positive response to NAC [22]. Several publications have reported reduced overall survival and reduced cancer-specific survival among women with Medicaid or no insurance [22,23]. These women often present more advanced stage of disease compared to women with private insurance. Patients without adequate insurance may not have access to supplementary medical services such as nutrition courses, psychological therapy, or exercise programs which have been documented to improve BC outcomes. Health insurance may serve as a proxy variable for socioeconomic factors such as access to preventive care and treatment, income, and nutritional status [22–24]. Lower socioeconomic status is linked with late-stage diagnosis, timely treatment, type of treatment received, and mortality [25]. The lower rate of chemotherapy completion is associated with these social determinants of health which can impact outcome including pCR and potentially survival.

Limitations

This study has several limitations due to the retrospective design and small sample of Black patients. Inherent in the nature of the study is bias associated with abstracting data from electronic medical records. Also, racial and ethnic diversity at our institution is modest with only ∼10% of patients having self-identified as Black. Hispanic ethnicity is self-reported in less than 2% of patients. Furthermore, we are unable to include data on chemotherapy treatment regimens and completion rates. Lastly, we were only able to collect a limited number of medical comorbidities and other sociodemographic variables were limited based on the high percentage of missing data.

Conclusions

Race and BMI were not significant predictors of pCR in this limited retrospective analysis. However, further exploration with a larger sample evaluating differences in pCR by BMI could lead to a better understanding of the association between obesity and pCR. Further exploration of the racial differences in pCR is still warranted considering the significant socioeconomic disparities and differences in obesity rates by race. These findings could potentially guide risk-based treatment choices after NAC. Future NAC studies must focus on the potential racial and ethnic treatment response in order to improve treatment for all women affected by breast cancer.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available on request from the corresponding author, [RR]. The data are not publicly available due to their containing information that could compromise the privacy of research participants.

Additional information

Funding

This work was supported by the Susan G. Komen Foundation under Grant TREND21677967 (RR).