Quality of Life Following Receipt of Adjuvant Chemotherapy With and Without Bevacizumab in Patients With Lymph Node–Positive and High-Risk Lymph Node–Negative Breast Cancer

Key Points Question Who is at risk of diminished quality of life (QOL) in extended follow-up of breast cancer survivors treated with intensive systemic treatment? Findings Among 455 participants included in this secondary analysis of a phase 3 randomized clinical trial, adding bevacizumab to chemotherapy was not negatively associated with QOL at 18 months. A substantial proportion of patients reported problems related to pain or discomfort and anxiety or depression; receiving mastectomy with radiation as well as identifying as Asian, Black, or American Indian or Alaska Native were associated with lower QOL. Meaning These findings suggest that early intervention and referral to supportive resources is critical for survivors at risk of experiencing diminished QOL, especially among individuals from underrepresented racial and ethnic groups and those undergoing extensive local therapy.


Introduction
It is well established that breast cancer treatment can affect not only short-term health outcomes but also longer-term health-related and psychosocial quality of life (QOL). [1][2][3] While clinical trials routinely document adverse effects during treatment, incorporation of patient-reported outcomes (PROs) into trial assessments increases our understanding of the acute and sustained consequences of treatment on physical and psychosocial well-being.
The ECOG-ACRIN Cancer Research Group (ECOG-ACRIN) E5103 multisite trial randomly assigned patients to chemotherapy with and without bevacizumab between November 2007 and February 2011, and a large consecutively enrolled subset were surveyed regarding QOL and health status in follow-up. 4 The primary objective of the current analysis was to compare QOL and health status among treatment arms in the parent trial at 18 months following enrollment. As secondary objectives, we explored differences in QOL and health status by local therapy as well as demographic, clinical, and treatment factors associated with of worse QOL at 18 months after enrollment.
Collection of PROs beyond the completion of active treatment from patients, all of whom received chemotherapy and nearly 80% of whom received bevacizumab, allowed for the evaluation of QOL in the setting of targeted therapy, and provides important information regarding the experience of breast cancer survivors with higher-risk disease in extended follow-up.

Study Design
This study was a secondary analysis assessing quality of life objectives in ECOG-ACRIN E5103, a double-blind phase 3 trial open to enrollment between November 2007 and February 2011. 4 The trial protocol appears in Supplement 1. Patients were randomly assigned 1:2:2 to 1 of 3 study arms: adjuvant doxorubicin, cyclophosphamide followed by paclitaxel with placebo (group A); adjuvant doxorubicin, cyclophosphamide with bevacizumab followed by paclitaxel with bevacizumab (group B); or adjuvant doxorubicin, cyclophosphamide with bevacizumab followed by paclitaxel with At baseline, patients were provided a copy of the surveys to complete in clinic and were then asked to take them home. A research interviewer then called and collected the PROs over the telephone. Patients were mailed a copy of the follow-up surveys 7 to 10 days prior to their scheduled interview, with a request to complete them prior to the arranged time of their interview.

Overall QOL
Overall QOL was assessed with the Functional Assessment of Cancer Therapy Scale-Breast Cancer with arm subscale (FACT-B+4). The FACT-B includes the FACT-General (FACT-G), version 4.0, developed by Cella and colleagues 6 as an overall cancer-specific QOL measure, with a 9-item subscale concerning breast cancer-specific issues. 5 The FACT-G consists of a 27 item core QOL measure grouped into 4 subscale domains: physical well-being, social and/or family well-being, emotional well-being, and functional well-being. 6 Almost all items are rated on a 5-item Likert scale, from 0, indicating not at all, to 4, indicating very much, and ask individuals about how they have felt during the past 7 days. The breast cancer module consists primarily of physical symptoms, body image, and sexual issues. The arm subscale includes 4 additional items plus 1 item from the breast cancer module to assess the impact of arm morbidity on patients following surgery. FACT-B scores range from 0 to 148; scores for the physical, social, and functional subscales range from 0 to 28. For the emotional subscale, breast cancer-specific subscales, and arm subscales, scores range from 0 to 24, 0 to 40, and 0 to 20, respectively. Higher scores represent better QOL. A difference of 7 to 8 points on the FACT-B score has been determined to be a minimally important difference (MID), with MIDs for individual subscale scores ranging from 1 to 3 points. 7,8

Health Status
The EuroQol 5-Dimensions 3-Levels (EQ-5D-3L) is a measure of health status for use in evaluating health and health care. 9 Descriptively, the EQ-5D consists of 5 dimensions: mobility, self-care, usual activity, pain or discomfort, and anxiety or depression. Each dimension has 3 levels designated as no problem, some problem, or extreme problem, with patients selecting their current level of function for that day on each dimension. Three different health status metrics were evaluated: (1) health status evaluation, representing the proportion reporting no problems vs some or extreme problems within each of the 5 dimensions; (2) index score, derived from the 5 dimensions, ranging from −0.11 (death) to 1 (perfect health) with an MID of 0.06 points; and (3) visual analog score (VAS), ranging from 0 (worst imaginable) to 100 (best imaginable), with an MID of 7 points. 10

Other Covariates
Sociodemographic characteristics, such as marital and education status, were collected using a modified version of the Cancer and Leukemia Group B (CALGB) background information form. 11

Results
Of 4836   A total of 257 women (57%) underwent mastectomy; of these, approximately two-thirds (171

Study Participant Characteristics
[66%]) received postmastectomy radiation therapy (PMRT). Among the 195 women who had breast conserving surgery (BCS), 184 (94%) received whole breast radiation therapy; 3 women (2%) had partial breast irradiation and 8 (4%) did not receive any type of radiation. All women who had BCS were analyzed as a single group, as small numbers precluded any evaluation of women who had BCS without radiation separately from those who did. Regarding axillary surgery, 86 women (19%) had a sentinel lymph node biopsy (SLNB) only, 108 (24%) had an axillary lymph node dissection (ALND) only, and 261 (57%) had both an SLNB and an ALND.

Comparison of QOL and Health Status Between Local Therapy Strategies
FACT-B, EQ-5D-3L, and EQ-VAS scores by local therapy strategy are presented in Table 3.      with usual activities. There were no statistically significant differences by local therapy strategy within each of the 5 EQ-5D-3L health status dimensions.

Factors Associated With QOL
Modeling was used to assess the association of sociodemographic and clinical factors with QOL as measured by the FACT-B. Univariable and multivariable model results are presented in FACT-B score for women who had a mastectomy with PMRT was more than 6 points lower than the mean FACT-B score for women who had BCS in the univariate model (coefficient, −6.6; 95% CI, −11.1 to −2.1) and slightly attenuated in the multivariable model (coefficient, −5.5; 95% CI, −10.1 to −0.9).
The mean FACT-B score for women with ER-positive tumors was 5 points lower than the mean FACT-B score for women with ER-negative tumors in univariable analyses (coefficient, −5.1; 95% CI, −9.4 to −0.9), although this effect was modestly attenuated in multivariable analyses. Age, a known prognostic factor, although not statistically significant, was retained by the model selection criteria

Discussion
At 18 months post enrollment, the addition of bevacizumab to an adjuvant chemotherapy regimen was not negatively associated with QOL, including among those who received extended duration therapy. Additionally, by 18 months, QOL scores were similar to US general population references.
However, in post hoc analyses, while there did not appear to be differences between women who had BCS and women who had mastectomy without PMRT, women who had mastectomy and PMRT had lower FACT-B and EQ-VAS scores compared with women who had BCS. Additionally, among those who had mastectomy, those who received PMRT had lower FACT-B arm subscale scores than those who did not have PMRT, suggesting decreased QOL attributable to greater arm morbidity associated with the addition of radiation therapy.
Many women may not be monitored as closely by their clinicians when active treatment ends, yet we observed that a substantial symptom burden persists in the months following the completion of active treatment. Overall, in our study, 58% of all patients reported problems with pain and discomfort. Prior studies have documented substantial levels of chronic pain reported by women after breast cancer surgery, with 1 systematic review reporting a median frequency of persistent pain of 37.5%. [13][14][15] Established treatment-related risk factors for experiencing chronic postsurgical pain include radiation, receipt of ALND, and receipt of chemotherapy. 16 Given that all women in our study were treated with chemotherapy, most had either ALND or SLNB with ALND, and two-thirds of women who had mastectomy received PMRT, reports of persistent pain and discomfort at 18 months after diagnosis should underscore the need for effective pain amelioration strategies into longerterm survivorship, particularly among patient populations who undergo heavy treatment.
Additionally, we found that women who identified as Asian, Black,

Limitations
Our findings should be interpreted in the context of some limitations. Our study was not designed to analyze differences by surgery or radiation, thus comparisons between BCS and mastectomy (with and without radiation) were exploratory and were not adjusted for multiple comparisons. While we adjusted for potential factors known to be associated with QOL in the regression model, we cannot exclude the possibility that differences in preoperative factors, including QOL, may be associated with surgical preferences as well as postoperative QOL. Furthermore, we did not have data available regarding how many women underwent postmastectomy reconstruction which may also factor into longer-term QOL. Finally, although we observed an association between race and QOL, our study population was predominantly White (86%). Ensuring that Black patients as well as patients from other racial and ethnic backgrounds who historically have been underrepresented in clinical trials are offered the opportunity to participate in clinical trials should be a priority. These efforts should extend to trials that include QOL end points, where diversity in the patient-reported experience is essential to capture.

Conclusions
This secondary analysis of a randomized clinical trial found no significant differences in QOL between study groups, with a substantial proportion of patients reporting problems related to pain or discomfort and anxiety or depression. Attention to both physical and psychosocial well-being is essential both during and after completion of active breast cancer treatment. Many of the problems that were reported by survivors in our study are amenable to intervention, underscoring the need for timely referral to supportive resources, especially for underserved populations. Routine monitoring of symptoms and functional status is warranted to identify and ameliorate symptoms early that may adversely affect QOL as well as to improve pain and optimize and sustain psychological health and physical functioning in breast cancer survivors during treatment and into survivorship.