Patient-Reported Outcomes after Irradiation of Tissue Expander vs. Permanent Implant in Breast Reconstruction: A Multicenter Prospective Study

Abstract

Background:

Whether to radiate the tissue expander prior to implant exchange or to defer radiation until after exchange in immediate, two-staged expander/implant (E/I) reconstruction remains uncertain. In this study, we evaluated the effects of radiation timing on complication rates and patient-reported outcomes (PROs) in patients undergoing immediate E/I reconstruction.

Methods:

Immediate E/I reconstruction patients undergoing post-mastectomy radiation therapy (PMRT) at 11 Mastectomy Reconstruction Outcomes Consortium (MROC) sites with demographic, clinical, and complication data were analyzed. PROs were assessed with BREAST-Q, PROMIS, and EORTC QLQ-BR23 surveys preoperatively and two years post-operatively. Survey scores and complication rates were analyzed using bivariate comparison and multivariable regressions.

Results:

Three hundred and seventeen patients met inclusion criteria. 237 patients underwent PMRT before expander-implant exchange (TE-XRT), and 80 patients after exchange (Implant-XRT). Controlling for covariates, timing of radiation had no significant effect on risks of overall complications (odds ratio(OR)=1.25, p=0.46), major complications (OR=1.18, p=0.62) or reconstructive failure (OR=0.72, p=0.49). Similarly, radiation timing had no significant effect on two-year PROs measured by the BREAST-Q or EORTC. Outcomes measured by PROMIS showed less anxiety, fatigue, and depression in the Implant-XRT group. Compared with preoperative assessments, two-year PROs significantly declined in both cohorts for satisfaction with breast, physical well-being, sexual well-being, but improved for anxiety and depression.

Conclusion:

Radiation timing (before or after exchange) had no significant effect on complication risks or on most PROs in immediate E/I reconstruction. Although lower levels of anxiety, depression, fatigue were observed in the Implant-XRT group, these differences may not be clinically significant.

INTRODUCTION

In light of results from studies like the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) meta-analysis showing reduction in breast cancer mortality and both loco-regional and distant recurrence with post-mastectomy radiation therapy (PMRT), U.S. and European breast cancer therapy guidelines have recommended strong consideration of PMRT for patients with one to three positive lymph nodes, broadening the indications for radiation therapy in breast cancer patients. ^1-5 Meanwhile, it has been well-established that improvements in body image after breast reconstruction positively impact health-related quality of life in women after mastectomy.^6-9 Unfortunately, studies have also shown poorer aesthetic outcomes, lower overall satisfaction, and higher complication rates in patients who undergo breast reconstruction with PMRT, compared with women receiving reconstruction without PMRT.^{10, 11} According to the American Society of Plastic Surgeons, the number of tissue expander-implant (E/I) reconstructions have increased by 18 percent from 2012 to 2016.¹² Given these trends, a growing number of patients choosing implant-based reconstruction will also require PMRT.

Because immediate expander/implant reconstruction requires a second operation to exchange the tissue expander for an implant, the optimal timing of radiation relative to the exchange procedure has been a long-debated subject. Many plastic surgeons remain uncertain whether it is preferable to radiate the tissue expander prior to implant exchange or to defer radiation until after the exchange. Previous studies have addressed PMRT timing in implant-based reconstruction patients with varied conclusions. Some have reported that PMRT prior to E/I exchange (after tissue expander placement) is associated with higher rates of failure^13-16, while others have concluded that timing of PMRT (before or after exchange) has no effect on the risks of reconstruction.^17-22 Many of these studies have been limited by their small sample size, single center designs and lack of patient-reported outcome (PRO) assessments. Using a multi-center prospective cohort design, the current study sought to compare complication rates and PROs in immediate E/I patients receiving PMRT before or after their E/I exchange procedures.

METHODS

Study Population

Patients in this study were recruited as part of the Mastectomy Reconstruction Outcomes Consortium (MROC), a prospectively designed study that recruited patients from 11 academic and private practices in the United States and Canada from 2012 to 2015, funded by the National Cancer Institute in 2011 (1R01CA152192). With institutional review board approval from all sites, MROC recruited women aged 18 years or older undergoing first time reconstruction after mastectomy for breast cancer treatment or prophylaxis. This analysis included women undergoing immediate, two staged E/I reconstruction and PMRT, and with two-year follow-up data. Patients were categorized as those receiving PMRT to their tissue expanders (TE-XRT—i.e., prior to exchange) and those undergoing PMRT to the subsequent reconstructive implant (Implant-XRT—i.e., following exchange).

Women electing to undergo breast reconstruction after a failed reconstruction or previous breast procedures such as breast augmentation, mastopexy, or reduction were excluded as part of the MROC cohort. Women receiving direct-to-implant reconstruction or any form of autologous reconstruction (with or without implant) were also excluded from this analysis. Lastly, any patients who had reconstructive failure (implant removal due to a complication) at any stage were excluded from the PRO analysis, since a full PRO survey panel was not administered in these cases. However, failures were still included in the complication analysis. Patients without complete two-year follow-up PRO data were also excluded.

Preoperative and postoperative demographic and clinical information was collected from the participants’ electronic medical records (EMRs). Variables included age, race/ethnicity, body mass index (BMI), laterality (bilateral vs. unilateral), acellular dermal matrix (ADM) use, chemotherapy, mastectomy type (nipple sparing, simple, mixed), smoking status, and diabetic status. One and two years post-operatively, EMRs were reviewed to collect complication data. A complication was defined as any adverse postoperative event secondary to the procedure that required additional treatment for the patient. A major complication was defined as an adverse event that required re-admission or re-operation, while reconstructive failure was defined as any implant that necessitated explantation.

Surveys, including the BREAST-Q, PROMIS-29, and EORTC QLQ-BR23, were administered preoperatively and at one and two years postoperatively to assess PROs. The BREAST-Q is a validated PRO instrument that has been widely implemented to assess health-related quality of life and patient satisfaction after breast reconstruction.²³ The BREAST-Q subscales used in this analysis were satisfaction with breast, psychosocial well-being, physical well-being, and sexual well-being. Developed in conjunction with the National Institutes of Health, the Patient-Reported Outcomes Measurement Information System (PROMIS) survey is a validated patient health outcomes measure assessing a wide range of PROs.²⁴ The PROMIS subscales adopted in this analysis included physical function, anxiety, depression, fatigue, sleep disturbance, satisfaction with participation in social roles, and pain interference. For negative constructs such as anxiety, depression, fatigue and sleep disturbance, higher scores correspond to worse outcomes. The EORTC QLQ-BR23 is a breast cancer patient-specific quality of life measurement tool developed from the European Organization for Research and Treatment of Cancer.²⁵ Subscales employed from the EORTC QLQ-BR23 were body image and sexual functioning.

Statistical Analysis

Baseline patient characteristics were compared by timing of radiation (TE-XRT vs. Implant-XRT) using Student’s t-test for continuous variables and chi-square test or Fisher’s exact test for categorical factors. For comparison of each two-year complication outcomes between TE-XRT and Implant-XRT patients, separate mixed-effects logistics regression models were used. Each model included an indicator for Implant-XRT and was adjusted for baseline patient characteristics. The model also included random intercepts for each hospital/center to account for between-center variability. For comparison of PRO scores, separate linear mixed-effects regression models were used. Similarly, each model included an indicator for Implant-XRT, patient characteristics as covariates including the baseline values of the outcome variable, and random intercepts for each hospital/center. To reduce potential bias, multiple imputations with chained equations were applied to create 10 complete imputed datasets. The regression models specified above were fit for each imputed data set. The results were then combined using Rubin’s rule. We reported adjusted odds ratios (ORs) for complications and beta coefficients for PROs, with 95% confidence intervals (CI) and corresponding p-values. Assuming an effect size of 0.5 and sample size of 190 and 72 in each group with two-year PRO data yielded an 89% power to detect a Cohen’s effect size (d) of 0.5 using a two-sided 0.05 level test. Statistical analysis was performed with SAS 9.4 (SAS Institute, Cary, NC) with statistical significance set at 0.05.

RESULTS

Among a total of 3120 patients in the MROC study, 317 patients met inclusion criteria. Of those, 237 patients received PMRT after tissue expander placement (TE-XRT), and 80 patients received PMRT after E/I exchange (Implant-XRT). The demographic characteristics of both the TE-XRT and Implant-XRT cohorts are summarized in Table 1. Most demographic variables including age, BMI, laterality, extent of disease, smoking, and diabetic status were not statistically significantly different between the two cohorts; however, there were significant differences in acellular dermal matrix (ADM) use and chemotherapy. More TE-XRT patients underwent surgery with ADM than Implant-XRT patients (59.1% vs. 32.5%; p < 0.001), while Implant-XRT patients were more likely to undergo more chemotherapy than TE-XRT patients (92.5% vs. 60.8%; p < 0.001). The distribution of patients by center are illustrated in Table 2.

Table 1:

Baseline Patient Characteristics

Variable	TE-XRT	Implant-XRT	p-value
No. patients	237	80
Age, mean (SD)	47.4 (10.4)	45.3 (10.1)	0.108
BMI, mean (SD)	26.3 (5.2)	25.2 (5.1)	0.104
Laterality
Unilateral	101 (42.6%)	31 (38.8%)	0.544
Bilateral	136 (57.4%)	49 (61.3%)
ADM usage
ADM used	140 (59.1%)	26 (32.5%)	< 0.001
ADM not used	97 (40.9%)	54 (67.5%)
Chemotherapy
During or after recon	144 (60.8%)	74 (92.5%)	< 0.001
Not during or after recon	93 (39.2%)	6 (7.5%)
Extent of disease
Local	49 (20.7%)	11 (13.8%)	0.172
Regional	188 (79.3%)	69 (86.3%)
Mastectomy type
NS	20 (8.4%)	6 (7.5%)	1.000†
SMR	214 (90.3%)	73 (91.3%)
Mixed	3 (1.3%)	1 (1.3%)
Race
White	204 (87.6%)	73 (93.6%)	0.207†
Other	29 (12.4%)	5 (6.4%)
Smoking
None smoker	157 (67.4%)	55 (68.8%)	0.821
Previous or current smoker	76 (32.6%)	25 (31.3%)
Diabetes
Yes	10 (4.2%)	2 (2.5%)	0.737†
No	227 (95.8%)	78 (97.5%)

^†Fisher's Exact Test

Table 2:

Patient Distribution by Center

Site	TE-XRT	Implant-XRT	Total
Brigham and Women's Hospital	63 (26.6%)	3 (3.75%)	66
Georgia Institute for Plastic Surgery	0 (0.0%)	1 (1.25%)	1
Georgetown University Hospital	15 (6.3%)	6 (7.5%)	21
University of Manitoba	3 (1.3%)	0 (0.0%)	3
MD Anderson Cancer Center	32 (13.5%)	1 (1.3%)	33
University of Michigan	4 (1.7%)	1 (1.3%)	5
Northwestern Memorial Hospital	62 (26.2%)	3 (3.8%)	65
Memorial Sloan-Kettering Cancer Center	46 (19.4%)	54 (67.5%)	100
Ohio State University	5 (2.1%)	11 (13.8%)	16
St. Joseph Mercy Ann Arbor	7 (3.0%)	0 (0.0%)	7
TOTAL	237	80	317

At two years postoperatively, the unadjusted rates of any complication or major complications did not differ significantly between the TE-XRT and Implant-XRT groups (Table 3). Without controlling for potential confounders, the reconstruction failure rate was almost twice as high in the TE-XRT group, compared to the Implant-XRT group (19.8% vs. 10%, p = 0.04, Table 3). However, after controlling for clinical covariates, the odds ratios for overall complications, major complication, and reconstruction failure did not differ significantly between the two cohorts. (Table 4)

Table 3:

Radiated TE and Permanent Implant Complications

	TE-XRT (Total 237 Patients)	Implant-XRT (Total 80 Patients)	p-value
Any Complication	95 (40.1%)	33 (41.3%)	0.85
Major Complication	82 (34.6%)	26 (32.5%)	0.73
Failure	47 (19.8%)	8 (10.0%)	0.04

Table 4:

Multivariable Regression for Complications

Variable	Overall complication			Major complication			Failure
	OR1	95% CI2	p-value	OR1	95% CI2	p-value	OR1	95% CI2	p-value
Implant XRT (vs. TE XRT)	1.25	(0.69, 2.25)	0.459	1.18	(0.62, 2.22)	0.615	0.72	(0.28, 1.83)	0.488
ADM used	1.02	(0.63, 1.67)	0.931	1.15	(0.67, 1.97)	0.625	0.92	(0.45, 1.86)	0.814
Chemo during/after reconstruction	0.96	(0.56, 1.63)	0.870	1.11	(0.63, 1.94)	0.719	1.20	(0.59, 2.41)	0.618
Bilateral reconstruction	1.76	(1.06, 2.92)	0.030	1.76	(1.04, 2.99)	0.037	1.19	(0.62, 2.29)	0.607
Age	1.03	(1.00, 1.05)	0.028	1.03	(1.00, 1.05)	0.036	1.03	(1.00, 1.06)	0.057
BMI	1.08	(1.03, 1.13)	0.001	1.07	(1.02, 1.13)	0.003	1.05	(0.99, 1.11)	0.084

¹OR: Odds Ratio

²CI: Confidence Interval

Of the 237 TE-XRT and 80 Implant-XRT patients, two-year PROs were missing in 55 (17%) patients (47 patients dropped out, and 8 patients had reconstruction failures). The remaining 262 patients (190 TE-XRT patients and 72 Implant-XRT patients) were included in the PRO analysis. Mean preoperative and two-year postoperative scores are summarized in Table 5. For all three survey instruments, preoperative PROs were not significantly different across the two cohorts, except the Implant-XRT group scored higher pre-operatively in both PROMIS anxiety (p=0.002) and depression (p=0.002) subscales compared to the TE-XRT group. Overall, most PRO subscales showed declines at two years postoperatively, compared to preoperative levels in both the Implant-XRT and TE-XRT cohorts, including the satisfaction with breast, physical well-being, and sexual well-being subscales of BREAST-Q. However, both cohorts showed improvement from baseline in anxiety and depression at two years (Table 5).

Table 5:

Unadjusted TE-XRT and Implant-XRT Patient-Reported Outcomes (PRO)

PRO measure	Time	TE-XRT (n=190) Mean (SD)	p-value1	Implant XRT (n=72) Mean (SD)	p-value1
BREAST-Q: Satisfaction with breast	Pre-op	64.4 (21.5)		65.4 (23)
	Two year post-op	55.7 (17.6)	0.009	51 (18)	<0.001
BREAST-Q: Psychosocial well-being	Pre-op	72 (18.1)		70.8 (17.3)
	Two year post-op	67.8 (17.7)	0.089	66.3 (19)	0.158
BREAST-Q: Physical well-being	Pre-op	80.1 (14.2)		77.1 (15.6)
	Two year post-op	71.6 (12.7)	<0.001	71 (16.4)	0.004
BREAST-Q: Sexual well-being	Pre-op	59.4 (18.1)		62 (17)
	Two year post-op	49.1 (14.2)	<0.001	47.6 (16.5)	<0.001
PROMIS: Physical function	Pre-op	52.2 (7)		53.4 (6.3)
	Two year post-op	51.9 (6.1)	0.239	51.7 (6.1)	0.142
PROMIS: Anxiety	Pre-op	58.4 (8.1)		61.9 (8.6)
	Two year post-op	51.9 (8.6)	<0.001	48.3 (8.1)	<0.001
PROMIS: Depression	Pre-op	49.4 (7.9)		53 (8.7)
	Two year post-op	47.1 (7.6)	0.018	45.9 (6.4)	0.002
PROMIS: Fatigue	Pre-op	49.5 (9)		49.7 (10.3)
	Two year post-op	49.8 (9.1)	0.693	45.3 (9.4)	0.173
PROMIS: Sleep disturbance	Pre-op	52 (3)		52.1 (3)
	Two year post-op	51.8 (3.1)	0.450	51.3 (2.7)	0.248
PROMIS: Satisfaction with participation in social roles	Pre-op	51.3 (8.8)		51.5 (9.3)
	Two year post-op	52.5 (9)	0.312	53.5 (7.6)	0.572
PROMIS: Pain interference	Pre-op	46.3 (7.4)		46.2 (8.4)
	Two year post-op	46.5 (7.5)	0.339	48.1 (9.6)	0.102
EORTC: Body image	Pre-op	74.8 (25.2)		76.2 (27.9)
	Two year post-op	69.4 (25.5)	0.029	68.5 (30.3)	0.575
EORTC: Sexual functioning	Pre-op	31.3 (22.9)		33.6 (25.1)
	Two year post-op	35.4 (24.9)	0.868	30 (17.7)	0.596

Note: Preop comparisons between cohort were not significant for any measures, except Anxiety (p=0.002) and Depression (p=0.002).

¹From testing for changes from baseline to 2 years

At two years, after adjusting for demographic and clinical covariates, the Implant-XRT cohort reported significantly lower levels of anxiety (beta=−4.8, p=0.01), depression (beta=−2.9, p=0.05), and fatigue (beta=−4.4, p=0.03), compared with the TE-XRT group on the PROMIS survey (Table 6). However, no significant cohort differences were observed at two years in any of the other PRO subscales of the BREAST-Q, PROMIS, and EORTC.

Table 6:

Multivariable Regression of Implant-XRT vs. TE-XRT: Two-Year Postoperative Patient-Reported Outcomes (PRO)

PRO measure	Beta1	95% CI2	p-value
BREAST-Q: Satisfaction with breast	−3.89	(−11.00, 3.23)	0.28
BREAST-Q: Psychosocial well-being	0.48	(−7.72, 8.68)	0.91
BREAST-Q: Physical well-being	−0.64	(−7.19, 5.90)	0.84
BREAST-Q: Sexual well-being	−1.00	(−8.41, 6.40)	0.78
PROMIS: Physical function	−0.04	(−2.40, 2.32)	0.97
PROMIS: Anxiety	−4.77	(−8.00, −1.53)	0.01
PROMIS: Depression	−2.92	(−5.85, 0.02)	0.05
PROMIS: Fatigue	−4.42	(−8.33, −0.50)	0.03
PROMIS: Sleep disturbance	−0.36	(−1.95, 1.24)	0.65
PROMIS: Satisfaction with participation in social roles	2.13	(−1.31, 5.56)	0.22
PROMIS: Pain interference	2.86	(−1.05, 6.77)	0.14
EORTC: Body image	4.22	(−5.57, 14.00)	0.39
EORTC: Sexual functioning	−1.40	(−8.58, 5.77)	0.70

¹Beta coefficient represents the adjusted difference in Implant-XRT PRO compared to TE-XRT PRO (i.e. Implant-XRT scored 4.77 points lower on PROMIS anxiety subscale compared to TE-XRT, adjusting for baseline covariates)

²CI: Confidence Interval

DISCUSSION

Post-mastectomy radiation therapy has been shown to decrease locoregional and distant cancer recurrence and increase overall breast cancer survival rates in patients with one to three positive nodes.^{1, 2, 26, 27} Meanwhile, multiple studies have shown an increase in immediate implant-expander reconstruction after mastectomy over the past two decades.^{12, 28} As a consequence of these trends, a corresponding rise in the number of patients who undergo radiation after reconstruction is inevitable. It has been well-documented that radiation may compromise wound healing and lead to capsular contracture, pain, and reconstruction failure compared to non-radiated implant-reconstructed breasts.^29-33 A recent study comparing radiation of autologous-based reconstruction to implant-based reconstruction in the MROC cohort found 2.6 times higher odds of complications in radiated implants when compared to non-radiated implants at two years.³⁴ Meanwhile, irradiation in autologous reconstructions was not associated with a significantly higher risk of complications at two years.³⁴ Despite studies showing the superiority of autologous reconstruction over implant-based reconstruction in the setting of PMRT, a growing number of women are choosing or being guided towards implant-based reconstruction.³⁵ Given the increasing utilization of both PMRT and implant-based reconstruction, understanding the effects of PMRT timing on the outcomes of immediate implant-based reconstruction has become essential for surgical decision-making.

Previous studies investigating the optimal timing of radiation in association with immediate E/I reconstruction have returned mixed results, with many limited by their small, single-center patient populations and their retrospective designs.^13-22 A report by Ogita and coworkers compared 32 TE-XRT patients with 49 Implant-XRT patients, with a median follow-up of 32 months.¹⁷ The authors found no significant differences in rates of reconstruction failure, re-operation, or infection with respect to timing of radiation.¹⁷ Among the largest single-center studies exploring PMRT timing in E/I procedures is an analysis by Cordeiro and colleagues comparing 94 TE-XRT and 210 Implant-XRT patients.¹⁴ This study demonstrated higher rates of Baker classification Grade III or IV capsular contractures in the Implant-XRT group (44.6%), compared with the TE-XRT group (15.9%). The authors postulated that this finding may have been due to aggressive capsulotomies performed at the time of exchange to a permanent implant. In addition, the paper noted that TE-XRT group had a higher reconstructive failure rate (18.1%), compared with the Implant-XRT cohort (12.4%), although this was not a statistically significant difference. Kaplan-Meier analysis predicted a 32 percent TE-XRT failure rate at six years, compared with a 16.4 percent Implant-XRT failure rate, (p < 0.01).¹⁴ Similarly, a recent meta-analysis of 20 studies including 2348 patients concluded that the risk of reconstructive failure was 2.33 times higher in the TE-XRT group, compared with the Implant-XRT group. However, incidence of capsular contracture was lower in the TE-XRT cohort, compared to Implant-XRT patients.¹⁵ In contrast with these earlier reports, more recent studies have noted no significant differences in failure rates when irradiating a tissue expander or a permanent implant.^18-22

In our study, most baseline demographic variables were not statistically different between the TE-XRT and Implant-XRT cohorts. One of the two variables that did vary significantly was ADM use. Acellular dermal matrix was used more commonly in the TE-XRT cohort than in the Implant-XRT group (59.1% vs. 32.5% respectively, p < 0.001). Although some studies have reported higher rates of postoperative seroma and infection with ADM³⁶, more recent analyses have not found significant ADM effects on either major or minor complications³⁷. It is uncertain why ADM use varied across the current study’s cohorts. However, given the potential effects of this material on complications and PROs, we chose to control for its use in our analyses. Likewise, receipt of chemotherapy varied significantly across our cohorts. Although most previous research has not reported significant chemotherapy effects on breast reconstruction outcomes³⁸, this variable was also included in our regressions.

Both Implant-XRT and TE-XRT cohorts demonstrated relatively high rates of complications, approximately 40%. This may be due to variations across studies in how complications are identified and recorded. In our study, a complication was defined as any adverse postsurgical event requiring additional treatment. Complication data were gathered by a trained site coordinator at each study center and centrally reviewed, ensuring that adverse events were counted as accurately as possible for two years following mastectomy and tissue expander placement. In addition, this observation may be attributable to the negative effects of radiation therapy on wound healing and implant-based reconstruction. The higher unadjusted reconstruction failure rate in the TE-XRT group compared to the Implant-XRT group (19.8 % vs. 10.0%, p = 0.04) may be explained by differences in age and BMI between the two cohorts. Both have been identified as risk factors for postoperative complications in previous studies.^{17, 39, 40} Although the differences were not statistically significant, the TE-XRT cohort was older (mean age 47.4 vs. 45.3, p = 0.11) and had higher BMIs (26.3 vs. 25.2, p = 0.10), which may have contributed to the higher unadjusted failure rates observed in the TE-XRT cohort. Adjusting for these variables, the odds of overall complication, major complication, and failure were not significantly different between the two cohorts. These findings suggest that timing of radiation may not have a significant impact on surgical risks, and that radiation should be administered when most oncologically beneficial in patients undergoing two-staged expander/implant reconstruction. However, the potential for increased complications after radiation irrespective of the timing of radiation should be discussed preoperatively with patients.

Among the study cohorts, two-year postoperative PRO measures were not significantly different for any of the BREAST-Q and EORTC subscales. However, the PROMIS anxiety, depression, and fatigue subscale scores were significantly lower (better) in the Implant-XRT group compared to TE-XRT patients. The BREAST-Q and PROMIS evaluate somewhat different domains of psychosocial well-being and quality of life, with PROMIS being more focused on physical symptoms and functional status. Thus, the variability in the findings from these instruments is not surprising. In addition, the reported minimal clinically important difference (MCID) of PROMIS is between 3.0 and 4.5⁴¹, and the beta coefficient between Implant-XRT and TE-XRT was at the cusp of that range; therefore, whether this statistical difference is clinically significant is debatable. The lower anxiety, depression and fatigue seen in Implant-XRT group compared with TE-XRT group, however, is notable in several aspects. It is in spite of the likely less time in Implant-XRT group to recover from radiation before the two-year assessment due to postponement of PMRT until after expander/implant exchange; and it is in spite of the higher pre-operative anxiety and depression scores in Implant-XRT patients than TE-XRT patients. One possible explanation for this finding is that the likely higher psychosocial satisfaction in the Implant-XRT patients resulting from nearing the completion of cancer treatment and reconstruction had a protective effect against the negative psychological effects of radiation therapy.

The significant decline in pre to postoperative BREAST-Q scores in both the TE-XRT and Implant-XRT cohorts were an unexpected finding. We have previously shown that breast reconstruction after mastectomy, whether immediate or delayed, preserves or restores pre-mastectomy PRO levels.⁴² This decrease in PROs may be explained by the relatively higher postoperative complication rates after PMRT compared to breast reconstruction patients who did not undergo PMRT. Alternatively, radiation therapy may affect physical and psychosocial well-being in manners that cannot be explained by the current study. However, given that this was not a primary outcome of this study, further research is warranted to elucidate the effects of PMRT on PROs in patients with breast cancer.

This analysis had a number of strengths. Our study is one of the first to evaluate the effects of PMRT timing on both complication rates and PROs in immediate expander/implant reconstruction. Our large, multicenter patient population enabled us to control for various potential confounding variables in our analyses. We used previously validated instruments to prospectively assess PROs after controlling for baseline scores. Because study centers vary in their protocols for PMRT, we also adjusted for study site in our analyses. The adverse effects of radiation may vary by differences in myriad factors, including targets, total dose and fractionation, boost administration, bolus application, use of techniques such as intensity modulation, and, time between radiation and surgery.

Despite its strengths, our study did have some limitations. As a prospective cohort study and not a randomized, controlled trial, MROC patients were not randomized for PMRT timing. Therefore, it is possible that unrecognized variables, not controlled for in our analysis, could confound our findings. As noted above, this study also did not collect details of specific radiation protocol variations among centers. Nine of the 11 sites in this study were academic centers, with all sites located in North America; therefore, the findings in this study may not be generalizable to all practice settings or geographical regions. Lastly, radiation effects on implant-based reconstruction are known to evolve over many years, and this study’s focus on two-year complication rates and PROs did not evaluate longer-term differences secondary to radiation timing.

CONCLUSION

The current multi-center, prospective study suggests that timing of post-mastectomy radiation therapy has no significant effect on complication rates or on most patient-reported outcomes in patients receiving immediate expander/implant reconstruction and PMRT. Thus, timing of PMRT relative to expander/implant exchange should remain based on oncologic, not reconstructive considerations.