Patient-Reported Quality of Life after Breast Conserving Surgery with Radiotherapy versus Mastectomy and Reconstruction

Structured Abstract

Objective:

To compare long-term QOL outcomes in breast cancer survivors who received breast conserving surgery with radiotherapy (BCS+RT) with those who received mastectomy and reconstructive surgery (Mast+Recon) without RT and identify other important factors.

Summary Background Data:

The long-term differences in patient-reported QOL outcomes following BCS+RT and Mast+Recon are not well understood.

Methods:

We identified patients from the Texas Cancer Registry with stage 0-II breast cancer diagnosed 2009–2014 after BCS+RT or Mast+Recon without RT. Sampling was stratified by age and race and ethnicity. A paper survey was sent to 4,800 patients which included validated BREAST-Q and PROMIS modules. Multivariable linear regression models were implemented for each outcome. Minimal clinically important difference for BREAST-Q and PROMIS modules, respectively, was 4 points and 2 points.

Results:

Of 1,215 respondents (25.3% response rate), 631 received BCS+RT and 584 received Mast+Recon. The median interval from diagnosis to survey completion was 9 years. In adjusted analysis Mast+Recon was associated with worse BREAST-Q psychosocial well-being (effect size −3.80, p=0.04) and sexual well-being (effect size −5.41, p=0.02), but better PROMIS physical function (effect size 0.54, p=0.03) and similar BREAST-Q satisfaction with breasts, physical well-being, and PROMIS upper extremity function (p>0.05) compared to BCS+RT. Only the difference in sexual well-being reached clinical significance. Older (≥65) patients receiving BCS+RT and younger (<50) patients receiving autologous Mast+Recon typically reported higher QOL scores. Receipt of chemotherapy was associated with detriments to multiple QOL domains.

Conclusions:

Patients who underwent Mast+Recon reported worse long-term sexual well-being compared to BCS+RT. Older patients derived a greater benefit from BCS+RT while younger patients derived a greater benefit from Mast+Recon. These data inform preference-sensitive decision making for women with early-stage breast cancer.

Mini Abstract:

In a population-based study of long-term patient-reported quality of life (QOL) among breast cancer survivors, mastectomy and reconstruction (Mast+Recon) was associated with worse sexual well-being compared to breast conserving surgery with radiotherapy (BCS+RT). Older (≥65) patients who received BCS+RT and younger patients who received autologous Mast+Recon typically reported the highest QOL scores. These data inform preference-sensitive decision making for women with early-stage breast cancer.

Introduction

Breast cancer is the most common non-skin cancer diagnosed among women in the United States (US).¹ Local treatment for early-stage breast cancer consists of either breast conserving therapy, most frequently involving breast conserving surgery followed by post-operative radiotherapy (BCS+RT), or mastectomy, often followed by reconstructive surgery, typically without radiotherapy (Mast+Recon).² While the oncologic equivalency of the two treatment strategies is well-established^3–6, they involve markedly different technical approaches, with BCS+RT consisting of a smaller surgery but exposure to therapeutic radiation, and Mast+Recon typically consisting of multiple larger surgeries but with omission of therapeutic radiation. The differences in these treatments may have implications for cosmesis and physical, psychological, and sexual health. Patient-reported quality of life (PR-QOL) data are increasingly recognized as valid and important outcomes in cancer patients^7–10 and can provide key comparative information in situations where there is equipoise of traditional clinical outcomes but where significant differences in side effects and quality of life may result.^11,12

Existing PR-QOL studies, however, are heterogeneous and include patients treated with or without adjuvant RT or reconstruction; as a result, the long-term ramifications of each treatment strategy remain poorly understood.^13–17 Prior work published by our group found clinically meaningful increases in psychosocial and sexual well-being with breast conservation compared to mastectomy and reconstruction, but confirmation of these findings in a larger, more contemporary cohort with additional relevant PR-QOL outcomes is needed.¹⁸ Currently, the utilization of Mast+Recon is rising in the US^19–21 despite evidence of increased rates of surgical complications and cost.^22,23 By focusing on a direct comparison of BCS+RT and Mast+Recon without RT, the typical expected treatment pathways at consultation, the long-term PR-QOL data presented help clarify the decision between BCS+RT and Mast+Recon for the numerous early-stage breast cancer patients faced with the dilemma each year. In addition, age and race and ethnicity are known to be important factors influencing both local treatment choice and outcome^14,24, and we hypothesized that stratifying by age and race and ethnicity a priori would allow for estimates of QOL outcomes personalized by these categories.

Methods

Cohort Selection and Sampling Strategy

We identified patients from the Texas Cancer Registry (TCR) with stage 0-II breast cancer diagnosed between 2009–2014 who underwent either BCS+RT (n=32,240) or Mast+Recon without RT (n=17,494), did not have multiple breast primary cancers, had a mailing address, and were alive as of 2018. The cohort sampling was stratified by TCR-coded race and ethnicity and age (<50, 50–64, ≥65 years). TCR race and ethnicity groups included Asian American-Pacific Islander (AAPI), non-Hispanic Black (hereafter, Black), Hispanic of any race or ethnicity (hereafter, Hispanic), and non-Hispanic White (hereafter, White). We chose to enrich the cohort for racial and ethnic minority groups. Race and ethnicity data were supplemented with data from self-report and resulted in inclusion of American Indian/Alaska Native (AIAN) as a separate stratum. Patients were classified as having received BCS+RT or Mast+Recon without RT based on TCR characteristics prior to survey mailing with confirmation by self-report. The final sample included 4,800 patients (2,770 BCS+RT and 2,030 Mast+Recon). Details regarding patient selection and sampling are provided in Supplemental Tables 1 and 2.

Study Measures

Under institutional review board approval, sampled individuals were mailed a study invitation letter, $10 gift card, and paper survey which included demographic and treatment questions, BREAST-Q modules²⁵ (satisfaction with breasts, psychosocial well-being, physical well-being, sexual well-being), and the Patient-Reported Outcomes Measurement Information System (PROMIS) upper extremity and physical function modules²⁶. These measures were selected based on qualitative interviews conducted with 20 breast cancer survivors to determine which domains were most relevant to long-term QOL. The surveys were mailed from November 2020 through December 2021 with methods developed by Dillman et al. implemented to enhance response rates.²⁷ Survey collection was closed in April 2022. BREAST-Q raw scores underwent Rasch transformation to yield a final score from 0 to 100, with a higher score indicating a better outcome.²⁵ PROMIS raw scores underwent transformation to yield a final T-score with a mean of 50 and a standard deviation of 10, with a higher score indicating a better outcome.²⁸ Based on published data, we considered a minimal clinically important difference for BREAST-Q modules to be 4 points²⁹, and for PROMIS upper extremity and physical function to be 2 points³⁰.

Statistical Analysis

We compared baseline patient and clinical characteristics by treatment group using the Chi-square test. Multivariable linear regression models were used to evaluate the association of candidate covariables with each outcome. Candidate covariables included those from self-report (race and ethnicity, current marital status, bra cup size, educational attainment, and household income; baseline body mass index (BMI), smoking status [>100 lifetime cigarettes or current use in the past 7 days]; history of chemotherapy and hormone therapy receipt) and TCR data (age at diagnosis, stage, tumor size, grade, node positivity, number of sampled nodes, estrogen receptor status, HER2 status).

Backward selection was used to retain variables with p<0.1. Treatment group and stratification variables (age, race and ethnicity) were retained in the models regardless of significance. For the primary outcomes, we performed an additional analysis considering the type of reconstructive surgery and tested interactions of age or race and ethnicity by treatment group. All multivariable analyses were adjusted for weights from sampling strata and non-response rates. Statistical analyses were performed in SAS^® version 9.4 (SAS Institute, Cary, NC) using two-sided tests with p<0.05 considered statistically significant.

Results

Analytic Cohort

Of 1,215 patients who responded (25.3%), 631 received BCS+RT and 584 received Mast+Recon. We excluded BCS+RT cohort patients who denied history of breast cancer (n=5) or RT (n=4) and Mast+Recon cohort patients who denied history of breast cancer (n=3) or reconstruction (n=14) and those who reported receiving postmastectomy RT (n=66), yielding a total of 1,123 patients (n=622 BCS+RT and n=501 Mast+Recon) who comprised the analytic cohort. Response rates varied by treatment group, age, race and ethnicity, year of diagnosis, tumor size, and nodal status, with patients in the Mast+Recon group more likely to respond (28.8% in Mast+Recon vs. 22.8% in BCS+RT, p<0.001) (Supplemental Table 3).

Baseline Characteristics

The median age was 54 years (interquartile range [IQR], 47 to 64 years) with 56.6% (n=688) of patients identifying with racial and ethnic minority groups including 24.0% Hispanic (n=292), 22.2% Black (n=270), 8.6% AAPI (n=105), and 1.7% AIAN (n=21). Most patients were married, overweight or obese, lifetime non-smokers, residents of a metro area, with a household income of ≤$80,000, node negative, with a tumor size of 0 to 2 cm, hormone receptor positive, and HER2 receptor negative. The median interval from diagnosis to survey completion was 9 years (IQR, 7 to 10 years).

There were multiple significant differences in baseline characteristics between BCS+RT and Mast+Recon patients with Mast+Recon patients more likely to be young, White or Hispanic, married, and non-smokers; have a normal BMI, a larger bra cup size, higher household income and educational attainment, negative lymph nodes, tumor size >2 cm, bilateral breast cancer, hormone receptor negative disease, and HER2 receptor positive disease; and have received chemotherapy. Full details can be found in Supplemental Table 4.

BREAST-Q Scores

Of the 1,123 respondents who were confirmed to have received either BCS+RT or Mast+Recon without RT according to their survey responses, BREAST-Q outcomes were scoreable in 1,024 (91.2%) for satisfaction with breasts, 1,074 (95.6%) for psychosocial well-being, 1,076 (95.8%) for physical well-being, and 1,011 (90.0%) for sexual well-being.

The unadjusted mean BREAST-Q satisfaction with breasts, psychosocial, physical, and sexual well-being scores for BCS+RT and Mast+Recon, respectively, were 61.8 vs. 63.4 (p=0.30), 78.1 vs. 74.7 (p=0.02), 77.0 vs. 78.4 (p=0.31), and 58.0 vs. 51.1 (p<0.001). In multivariable linear regression, as compared to BCS+RT, Mast+Recon was associated with statistically worse psychosocial well-being (effect size −3.80, 95% CI −7.33 to −0.27; p=0.04) and sexual well-being (−5.41, 95% CI −9.78 to −1.04; p=0.02), but similar satisfaction with breasts (3.19, 95% CI −0.71 to 7.08; p=0.11) and physical well-being (1.96, 95% CI −1.14 to 5.07; p=0.22) (Supplemental Table 5).

Satisfaction with breasts was higher among patients with household income >$80,000 (8.15, 95% CI 3.74 to 12.55, p<0.001).

Psychosocial well-being was higher among patients age 50–64 (5.02, 95% CI 1.21 to 8.82, p=0.01) and ≥65 (12.94, 95% CI 8.13 to 17.74, p<0.001) compared to age <50 and with household income >$80,000 (9.31, 95% CI 5.21 to 13.41, p<0.001). Psychosocial well-being was worse among obese patients (BMI ≥30 kg/m², −5.29, 95% CI −9.79 to −0.79, p=0.02) and those who were unmarried (−6.44, 95% CI −10.28 to −2.6, p=0.001).

Physical well-being was higher among patients age ≥65 (6.38, 95% CI 2.37 to 10.40, p=0.002) compared to age <50 and with a college degree or more education (7.52, 95% CI 4.31 to 10.73, p<0.001). Physical well-being was worse among AIAN (−11.69, 95% CI −16.61 to −6.78, p<0.001), Black (−4.21, 95% CI −7.84 to −0.57, p=0.02), and AAPI (−10.68, 95% CI −21.22 to −0.13, p=0.047) patients with White as referent and among those with a history of chemotherapy receipt (−4.34, 95% CI −7.80 to −0.89, p=0.01).

Sexual well-being was higher among patients age ≥65 (11.5, 95% CI 5.11 to 17.89, p<0.001) compared to age <50 and with household income >$80,000. Sexual well-being was worse among obese patients (−9.90, 95% CI −15.34 to −4.45, p<0.001) and patients with a history of smoking (−5.4, 95% CI −10.41 to −0.39, p=0.04) or a history of chemotherapy receipt (−7.73, 95% CI −12.24 to −3.32, p<0.001).

PROMIS Scores

Of the 1,123 respondents who were confirmed to have received either BCS+RT or Mast+Recon without RT according to their survey responses, PROMIS upper extremity was scoreable in 1,085 (96.6%) and PROMIS physical function in 1,097 (97.7%).

The unadjusted mean PROMIS upper extremity and physical function scores for BCS+RT and Mast+Recon, respectively, were 46.0 vs. 49.0 (p<0.001) and 34.7 vs. 35.9 (p<0.001). In multivariable linear regression, as compared to BCS+RT, Mast+Recon was associated with statistically better PROMIS physical function (effect size 0.54, 95% CI 0.05 to 1.03; p=0.03) but similar PROMIS upper extremity (0.93, 95% CI −0.51 to 2.37; p=0.21) (Supplemental Table 6).

Upper extremity function was higher among patients with C bra cup size compared to A/B (3.46, 95% CI 1.53 to 5.39, p<0.001) and with household income >$80,000 (4.14, 95% CI 2.52 to 5.77, p<0.001). Upper extremity function was worse for patients age ≥65 (−2.88, 95% CI −4.93 to −0.82, p=0.006) compared to age <50, for AIAN patients with White as referent (−6.27, 95% CI −11.25 to −1.29, p=0.01), for patients who were obese (−2.71, 95% CI −4.52 to −0.91, p=0.003), for patients with a history of smoking (−2.11, 95% CI −3.70 to −0.53, p=0.01), and for patients with a history of chemotherapy receipt (−2.87, 95% CI −4.41 to −1.33, p<0.001).

Physical function was higher among patients with household income >$80,000 (1.36, 95% CI 0.71 to 2.00, p<0.001). Physical function was worse for patients age ≥65 (−1.51, 95% CI −2.37 to −0.65, p<0.001) compared to age <50, for AIAN patients with White as referent (−3.30, 95% CI −5.80 to −0.80, p=0.01), for patients who were obese (−1.09, 95% CI −1.73 to −0.45, p<0.001), and for patients with a history of chemotherapy receipt (−0.88, 95% CI −1.42 to −0.35, p=0.001).

Subgroup Analysis and Interactions

Among patients who underwent Mast+Recon, 306 (52%) received implant reconstruction, 187 (32%) received autologous reconstruction, and 91 (16%) were unknown while 414 (71%) had both breasts removed and 455 (78%) had both breasts reconstructed.

When patients who underwent Mast+Recon were divided into two groups by the type of reconstructive surgery (autologous, implant), mean adjusted BREAST-Q satisfaction with breasts and physical well-being was significantly higher for autologous reconstruction compared to implant reconstruction or BCS+RT (Figure 1). Conversely, mean adjusted BREAST-Q sexual well-being was significantly lower for implant reconstruction compared to autologous reconstruction or BCS+RT.

Figure 1.

Adjusted BREAST-Q and PROMIS scores for mastectomy and autologous reconstruction (Autologous), implant reconstruction (Implant), or breast conserving surgery and radiotherapy (BCS+RT). Error bars represent 95% confidence intervals compared to referent group BCS+RT. * indicates a statistically and clinically significant difference compared to BCS+RT.

When analyzing patients who underwent Mast+Recon based on whether they received unilateral or bilateral reconstruction, mean adjusted BREAST-Q satisfaction with breasts was significantly higher for bilateral reconstruction compared to BCS+RT whereas mean adjusted psychosocial and sexual well-being were both significantly lower for unilateral reconstruction compared to BCS+RT (Figure 2).

Figure 2.

Adjusted BREAST-Q and PROMIS scores for bilateral reconstruction (Two Reconstructed breasts), unilateral reconstruction (One Reconstructed Breast), or breast conserving surgery and radiotherapy (BCS+RT). Error bars represent 95% confidence intervals compared to referent group BCS+RT. * indicates a statistically and clinically significant difference compared to BCS+RT.

There were significant interactions between treatment and age for BREAST-Q satisfaction with breasts (p=0.05), sexual well-being (p=0.06), and physical well-being (p=0.05) (Table 1). Subgroup analysis of treatment and age group for each of these outcomes was performed.

Table 1.

Interaction of Treatment, Age, and Race and Ethnicity in Multivariable Linear Regression Weighted Models of BREAST-Q and PROMIS Patient-Reported Outcomes

Outcome	P for interaction (treatment and age)	P for interaction (treatment and race and ethnicity)
BREAST-Q
Satisfaction with Breasts	0.05	0.40
Psychosocial Well-Being	0.20	0.95
Sexual Well-Being	0.06	0.32
Physical Well-Being	0.05	0.67
PROMIS
Upper Extremity	0.68	0.98
Physical Function	0.29	0.96

Patients ≥65 years old who underwent BCS+RT and patients <50 and 50–64 who underwent autologous Mast+Recon had significantly higher satisfaction with breasts compared to patients <50 who underwent BCS+RT (Table 2). Similarly, patients ≥50 who underwent BCS+RT, patients of all ages who underwent autologous Mast+Recon, and patients <50 who underwent implant Mast+Recon had higher physical well-being compared to patients <50 who underwent BCS+RT. Finally, patients ≥65 who received BCS+RT had higher sexual well-being but patients <50 who underwent implant Mast+Recon had lower sexual well-being compared to patients <50 who received BCS+RT.

Table 2.

Adjusted BREAST-Q Scores in Groups Defined by Treatment and Age

		Satisfaction with Breast1 (n=1024)		Physical Well-Being2 (n=1076)		Sexual Well-Being3 (n=1011)
Treatment and Age		Estimates	P	Estimates	P	Estimates	P
BCS+RT	<50	55.31	Referent	74.34	Referent	61.45	Referent
	50–64	57.12	0.51	80.03	0.02	62.27	0.72
	65+	63.42	0.02	84.22	<0.001	73.92	0.001
Mast+Recon (Implant)	<50	59.92	0.11	80.73	0.03	54.82	0.04
	50–64	61.22	0.11	79.05	0.10	58.05	0.40
	65+	58.53	0.42	76.96	0.52	60.91	0.91
Mast+Recon (Autologous)	<50	65.46	0.004	84.02	0.003	54.71	0.12
	50–64	65.03	0.01	87.45	<0.001	62.69	0.78
	65+	57.93	0.72	89.86	0.002	55.22	0.47

BCS, breast conserving surgery; Mast, mastectomy; Recon, reconstruction; RT, radiation therapy

Adjusted covariables include

¹race and ethnicity and household income,

²race and ethnicity, smoking history, education, and chemotherapy receipt, and

³race and ethnicity, marital status, body mass index, smoking history, household income, and chemotherapy receipt

Discussion

The initial discussion of treatment for early-stage breast cancer is typically centered around either breast conserving surgery followed by postoperative radiotherapy (BCS+RT) or mastectomy and reconstruction (Mast+Recon) without RT. Since these treatments result in similar survival and rates of local recurrence, this decision is considered highly preference-sensitive and could be enhanced through incorporation of patient-reported quality-of-life (PR-QOL) data.³¹ In our population-based study of 1,123 Texas breast cancer survivors, there was 1) similar satisfaction with breasts, physical well-being, and upper extremity function, 2) worse psychosocial well-being and sexual well-being, and 3) better physical function among patients who underwent Mast+Recon without RT compared to those who received BCS+RT at a median of 9 years after cancer diagnosis. Age, household income, BMI, race and ethnicity, and chemotherapy were also consistently found to be important factors associated with multiple PR-QOL domains.

While the effect sizes for differences in psychosocial well-being, sexual well-being, and physical function were statistically significant, they were small. The minimal clinically important difference (MCID) for BREAST-Q psychosocial well-being and sexual well-being has been suggested to be 4 points based on a distribution-based analysis by Voineskos et al²⁹, and the MCID for PROMIS upper extremity and physical function has been estimated at 2 points.³⁰ Therefore, only the difference in sexual well-being constituted a clinically important difference based on these criteria (effect size 5.41 favoring BCS+RT) with a marginal difference in psychosocial well-being (effect size 3.80 favoring BCS+RT). Overall, this suggests that patients may have relatively similar long-term experiences after the two treatment options despite minor differences in PR-QOL.

Prior work by our group demonstrated larger statistically and clinically meaningful improvements in psychosocial and sexual well-being with BCS+RT compared to Mast+Recon in a smaller cohort of Texas breast cancer survivors with a median time from diagnosis to survey of 10.3 years.¹⁸ In addition to being significantly larger, the present study cohort also represents a more contemporary cohort with additional important PROMIS QOL outcomes captured. An important distinction is that PROMIS physical function measures global physical function including activities of daily living, mobility, and upper extremity function²⁶ whereas BREAST-Q physical well-being focuses on function, pain, and discomfort of the upper extremity and breast.²⁵ The similar PROMIS outcomes with BCS+RT or Mast+Recon show that more global function including mobility and ability to carry out activities of daily living are not significantly affected by local treatment strategy. However, the reason for the difference in effect size estimates in psychosocial and sexual well-being is not immediately clear as the sampling strategy, demographics, and follow up time of the two cohorts are very similar, apart from the diagnosis years sampled (2006–2008 for the prior study, and 2009–2014 for the present study). One hypothesis is that reconstructive techniques, volume, and/or referral patterns improved over this time period.³² Alternatively, random variation, or response bias could have contributed, and highlight the importance of multiple high-quality studies to provide confirmation of findings. Nonetheless, the concordance of both studies in finding improvements in psychosocial well-being and sexual well-being, but similar satisfaction with breasts and physical well-being with BCS+RT compared to Mast+Recon, strengthens the validity of these findings.

There has been a marked increase in the proportion of BCS-eligible patients undergoing mastectomy in recent years^19–21, with the greatest increase in patients with node-negative and in-situ disease. These trends appear to be driven by higher rates of contralateral prophylactic mastectomies (CPM), and may in part be due to significant knowledge deficits regarding treatment options, expectations of improved breast symmetry and cosmesis with CPM and reconstruction, and an overestimation of the risk of developing a contralateral recurrence.^20,31,33,34 The data presented here, demonstrating similar clinically meaningful long-term QOL outcomes between BCS+RT and Mast+Recon without RT apart from sexual well-being favoring BCS+RT, suggest that while the decision between the two treatment strategies remains personal and situation-dependent, the tradeoff for many women might favor breast conservation due to the increased operative and recovery times, rates of surgical complications, and cost associated with mastectomy with reconstruction without a clear QOL benefit.^22,23

However, nuance is required as we found in subgroup analysis that patients undergoing autologous reconstruction reported more favorable cosmetic and physical well-being outcomes and patients undergoing bilateral breast reconstruction reported better cosmetic outcomes. In addition, older (≥65) patients derived the greatest benefit from a BCS+RT strategy while younger (<50) patients derived a greater benefit from Mast+Recon (apart from sexual well-being). This could be related to either increased biologic sensitivity to radiotherapy³⁵ or greater bother by late radiation effects such as skin fibrosis, induration, thickening, pigmentation, and telangiectasias, and muscle atrophy and tightness among younger patients. Conversely, larger and additional surgeries may have more frequent complications, delayed wound healing, and greater negative impact on long-term QOL in older patients.³⁶

We chose to focus specifically on patients who received BCS+RT and Mast+Recon without RT, rather than include patients who underwent BCS alone, mastectomy alone, or mastectomy and reconstruction with RT to provide a clean comparison of the two treatment options that would be most discussed at initial consultation for a patient with early-stage breast cancer, lead to similar cancer outcomes, and have a high chance for favorable cosmesis. While prior studies have found significantly improved QOL in domains of satisfaction with breasts, physical well-being, psychosocial well-being, sexual well-being, and body image and sexuality after breast conservation compared to mastectomy, they may have overestimated the benefit of breast conservation due to the inclusion of patients who received BCS without RT as well as mastectomy patients who did not undergo reconstruction and/or received adjuvant RT.^{13,14,37–39}

The potential impact of RT, reconstruction, and type of reconstruction on QOL among mastectomy patients is demonstrated by Jagsi et al., who examined cosmetic outcome in a longitudinal study of breast cancer survivors and found similar cosmetic outcome score for mastectomy with reconstruction compared to breast conserving therapy, but worse scores for mastectomy without reconstruction.¹⁵ Among those who underwent mastectomy, the highest satisfaction with reconstruction was for autologous reconstruction without RT, followed by autologous reconstruction with RT, implant reconstruction without RT, and markedly lower scores for implant reconstruction with RT. Indeed, numerous studies have found autologous reconstruction to be associated with superior long-term QOL in multiple domains compared to implant-based reconstruction⁴⁰; therefore, the anticipated type of reconstruction should be considered when initially counseling a patient on treatment options.

Chemotherapy has been associated with detriments in QOL among breast cancer survivors and recent data suggest the existence of distinct QOL trajectory clusters that could be mediated by factors such as age, comorbidities, income, and behaviors such as tobacco use and physical activity.^41–43 However, the impact of chemotherapy on long-term QOL is still poorly understood, with some work suggesting a mostly transient impact on QOL.⁴⁴ Our data now demonstrate a strong and enduring association of chemotherapy with detriments in multiple QOL domains including physical well-being, sexual well-being, upper extremity function, and physical function out to approximately 9 years after cancer diagnosis. This broad impact on QOL could be mediated by chemotherapy-induced weight gain, premature menopause, neuropathy, cardiovascular disease, fatigue, and/or cognitive changes and underscores the need for a better understanding of and appreciation for the lasting side effects of chemotherapy.^44,45 There may be additional patients who do not derive benefit from chemotherapy, including those who are clinically node-positive, that can be identified from genomic risk profiling and patients should be carefully selected by their providers to optimize long-term QOL.⁴⁶

Another strength of our study was the racial diversity of the cohort, with over half of respondents identifying with a minority racial or ethnic group. American Indian Alaskan Native (AIAN) patients reported worse physical well-being, upper extremity function, and physical function while non-Hispanic Black patients reported worse physical well-being. The finding of worse QOL outcomes among AIAN patients replicates data from our group’s prior work and merits additional consideration and study.¹⁸ Although we did not find evidence of pervasive racial and ethnic differences in QOL outcomes, this may partially have been masked by the inclusion of only mastectomy patients who underwent reconstruction as there is a history of racial and ethnic disparities in use and access to postmastectomy reconstruction in the US.^24,47,48 Nevertheless, it is reassuring that when patients of different racial and ethnic backgrounds receive standardized treatment, long-term QOL outcomes are similar.

There are some important limitations of this analysis. Radiotherapy technique including fractionation, use of partial breast irradiation or tumor bed boost, type of external radiation or brachytherapy technique and surgical technique including oncoplastic reconstruction, symmetry procedures, muscle-sparing reconstruction, or skin-sparing mastectomy were not captured. As a cross-sectional study, changes in QOL over time could not be tracked and it is unknown whether patients had similar baseline PROs prior to undergoing treatment. In addition, there were significant differences in baseline characteristics among patients who underwent BCS+RT compared to Mast+Recon which may not have been fully accounted for in multivariable analysis. While the population-based sampling strategy of the study is a strength, it was performed in Texas and may not be generalizable to other regions or countries.

In this large, racially diverse population-based cohort of breast cancer survivors, those who underwent BCS+RT reported statistically better long-term psychosocial well-being and sexual well-being, but worse physical function compared to Mast+Recon without RT. However, these differences were small, with only the difference in sexual well-being meeting the threshold for clinical relevance. Autologous reconstruction was associated with better QOL outcomes while older (≥65) patients receiving BCS+RT and younger (<50) patients receiving Mast+Recon also typically reported better QOL outcomes. Among other prognostic factors identified, the use of chemotherapy was associated with broad long-term detriments to physical well-being, sexual well-being, upper extremity function, and physical function. These data inform preference-sensitive decision making for women with early-stage breast cancer.