Effect of radiotherapy after breast-conserving surgery in older patients with early breast cancer and breast ductal carcinoma in situ: a meta-analysis

Background There are no consistent agreements on whether radiotherapy after breast-conserving surgery (BCS) could provide local control and survival benefit for older patients with early breast cancer or breast ductal carcinoma in situ (DCIS). The present study aimed to evaluate the efficacy of radiotherapy after BCS in older patients with early breast cancer or DCIS. Results Radiotherapy could reduce the risk of local relapse in older patients with early breast cancer. The 5-year AR of local relapse was 2.2% and 6.2% for radiotherapy and non-radiotherapy group, respectively, with low 5-year ARD of 4.0% and high NNT of 25. The 10-year AR of local relapse was 5.3% and 10.5% for radiotherapy and non-radiotherapy group, respectively, with the 10-year ARD of 5.2% and NNT of 20. However, radiotherapy could not improve survival benefits, including overall survival, cancer-specific survival, breast-cancer-specific survival, and distant relapse. Moreover, radiotherapy could reduce the risk of ipsilateral breast events in older patients with DCIS. Materials and Methods PubMed and Embase database were searched for relevant studies. Hazard ratios (HRs), risk ratios (RRs), absolute risk (AR), absolute risk difference (ARD), and number needed to treat (NNT) were used as effect measures to evaluate the efficacy of radiotherapy in older patients. Conclusions Our study indicates that radiotherapy could slightly reduce the risk of local relapse in older patients with favorable early breast cancer. However, radiotherapy cannot translate into significant survival benefits.


INTRODUCTION
Breast cancer is a global health issue in women worldwide, and the incidence of breast cancer has risen in women with older age [1][2][3]. Although postoperative breast radiotherapy is the standard treatment following breast-conserving surgery (BCS) in early breast cancer, the treatment recommendations for postoperative radiotherapy do not consider age as a treatment factor and there is a lack of high-level evidence for the efficacy of postoperative radiotherapy after BCS in older patients with early breast cancer [4]. Moreover, it is inappropriate to extrapolate the results of studies in younger patients to older patients, because older patients may have high comorbidity and thus are frequently excluded from studies [5,6].
Older patients may have less aggressive tumor biology and the postmenopausal status contributes to a lower risk of tumor relapse [7,8]. Besides, the high risks of comorbidity in older patients are associated with increased complications and decreased tolerability of treatment and short life expectancy. Several studies have showed that the use of postoperative radiotherapy after

Research Paper
Oncotarget 28216 www.impactjournals.com/oncotarget BCS appears to have declined in older patients with early breast cancer [9,10]. Thus, it is difficult and important to explore the efficacy of postoperative radiotherapy in older patients with early breast cancer [11,12]. However, it remains controversial whether an omission of radiation therapy can be considered for older patients with early breast cancer.
Thus, the purpose of this study was to evaluate the efficacy of postoperative radiotherapy after BCS in older patients with early breast cancer and whether the efficacy of postoperative radiotherapy differed according to factors such as age and tumor characteristics. Moreover, we also assessed the efficacy of postoperative radiotherapy for older patients with breast ductal carcinoma in situ (DCIS).

Radiotherapy and IBTR
The results of IBTR indicated a lower risk of IBTR in patients who received radiotherapy compared with patients who did not receive radiotherapy (HR = 0.23, 95% CI = 0.09-0.60, I 2 = 54.3%; RR = 0.25, 95% CI = 0.12-0.50, I 2 = 67.8%). Similar results were obtained in the sensitivity analyses according to age, study design, tumor characteristics, and tamoxifen.

Radiotherapy and distant relapse
Our result indicated that radiotherapy after BCS could not reduce the risk of distant relapse (HR = 1.03, 95% CI = 0.77-1.39, I 2 = 0.0%; RR = 1.20, 95% CI = 0.92-1.57, I 2 = 0.0%; Figure 3). Moreover, we performed Oncotarget 28217 www.impactjournals.com/oncotarget sensitivity analyses on the basis of favorable tumor characteristics, advanced age, RCT, and tamoxifen. The results still showed that the risk of distant relapse was not associated with radiotherapy after BCS.

Radiotherapy and other secondary outcomes in early breast cancer
Our study also evaluated the associations between radiotherapy and subsequent mastectomy and contralateral breast cancer. Our results indicated that the risk of contralateral breast cancer was similar between the patients who received and did not receive radiotherapy (RR = 1.64, 95% CI = 0.84-3.23, I 2 = 0.0%), with similar AR. In addition, radiotherapy reduced the frequency of subsequent mastectomy for local relapse in ipsilateral side (RR = 0.29, 95% CI = 0.12-0.73, I 2 = 0.0%).

Impact of radiotherapy in DCIS
There were only two RCT providing subgroup analyses comprising older patients with DCIS, with follow up of approximately 8 years [16,25]. Radiotherapy could reduce the risk of ipsilateral breast events, including new DCIS or invasive cancer (radiotherapy group: AR = 9.2%, 95% CI = 6.3%-12.1%; non-radiotherapy group: AR = 26%, 95% CI = 21.8%-30.3%), with ARD of 16.8% (95% CI = 11.7-22.0%) and NNT of 6. However, our study could not obtain the results for the association between radiotherapy and survival, such as OS, CSS and breast-CSS, due to lack of data.

DISCUSSION
Postoperative radiotherapy is an integral part of treatment to reduce the local relapse in patients with early breast cancer [4]. However, the treatment recommendations for early breast cancer are mainly based on clinical studies that frequently include young and exclude older patients, and the use of radiotherapy after BCS obviously declines with advancing age [10]. Thus, there are no consistent agreements about whether radiotherapy could provide substantial local control and survival benefit for older patients with early breast cancer. Our meta-analysis of 14 studies provides evidence of an association between radiotherapy use and high local control in older patients with early breast cancer. However, radiotherapy use cannot translate into significant survival benefits.
Our study indicated that radiotherapy after BCS could reduce the risk of local relapse in older patients with early breast cancer. The AR of local relapse was 2.2% and 6.2% for patients who received and did not receive radiotherapy after approximately 5 years, respectively, corresponding with ARD of 4.0% and NNT of 25. Besides, the ARD for local relapse was 5.2% after a longer duration of follow up of approximately 10 years. However, radiotherapy could not improve survival benefits, including OS, CSS, breast-CSS, non-breast-CSS, non-CSS and other cancer-CSS. The results were confirmed by indepth sensitivity analyses. Moreover, radiotherapy could reduce the risk of ipsilateral breast events in older patients with DCIS.
Clinically, age is an important and necessary factor to be considered in the use of radiotherapy in patients with early breast cancer. Compared with younger patients, older patients are more likely to have higher comorbidity, worse tolerability for treatment, and shorter life expectancy, which may have a negative influence on survival benefits. Indeed, clinical treatments for older patients with early breast cancer were often less aggressive, and epidemiological studies also reported that there was significant reduction in use of radiotherapy with increasing age for patients with early breast cancer [9,10]. Moreover, the EBCTCG trial showed that the risk of local relapse fell with increasing age in both radiotherapy and no radiotherapy group for early breast cancer after BCS, as did the ARD [27,28]. The NSABP B-21 trial also showed that the risk of IBTR was reduced with advancing age in early breast cancer [24]. Liljegren et al. showed that risk for local relapse decreased by 3% (95% CI, 1% to 6%) per year of increasing age [29]. The results of the current study indicated that the 5-year and 10-year ARD for local relapse was 4.0% and 5.2%, respectively, and the 5-year and 10-year ARD for IBTR was 4.4% and 4.4%, respectively. Although our results indicated that radiotherapy after BCS could slightly reduce the risk of local relapse and IBRT, the absolute benefits from radiotherapy may be less because ARD was low, resulting in a high NNT for preventing one local relapse. Besides, our results showed that there were no differences for the survival. Similarly, the PRIME, PRIME II, and CALGB 9343 trials reported that radiotherapy could only improve local control but not survival [14,18,19].
As mentioned, adding radiotherapy to BCS did not translate into significant survival benefits in older patients with early breast cancer. It is probably that a survival benefit was difficult to be detected in older patients with early breast cancer treated with radiotherapy because they had good survival even without radiotherapy. Moreover, several studies have reported that breast cancer tend to behave more indolently in older patients, and the postmenopausal status was related to favorable survival [7,8]. And tamoxifen could also contribute to a substantial reduction in the risk of tumor relapse and thus weaken the benefits from radiotherapy [27,30]. Indeed, Fyles et al. included postmenopausal patients with tamoxifen administration and reported that the rate of 5-year OS was 93.2% and only 10 deaths were due to breast cancer in non-radiotherapy group [23]. CALGB 9343 trial showed that the OS, in non-radiotherapy group, was 86% and only 3 deaths were due to breast cancer after 5-year follow up, www.impactjournals.com/oncotarget and 10-year OS was 66% and 10-year breast-CSS was 98% after 12.6-year follow up [18,31]. Furthermore, radiotherapy may increase mortality from heart disease and lung cancer, resulting in counteracting the slight benefit for breast cancer [32,33]. Thus, individual and tailored strategies on indication of radiotherapy were important and should be precisely explored.
Clinically, it was important to identify subpopulation who could safely and effectively omit the radiotherapy after BCS. The main concern was the definition of older age for patients with early breast cancer. However, there was no consistent definition of older age and the definition of older age varied from 50 to 70 in included studies. In our study, the "older" not only focused on the objective age but also on the menopausal status because postmenopausal status maybe reflect physiological status for "older" and was a favorable survival factor. Our study included at least 88.8% patients were aged ≥ 60. Furthermore, sensitivity analysis based on age obtained similar results. Unfortunately, our study could not determine the optimal definition of older age due to a lack of data. Future studies are required to establish an Oncotarget 28222 www.impactjournals.com/oncotarget optimal cut-off value for older age, considering the fact that a higher cut-off value may miss a substantial number of patients in clinical practice. In addition, the tumor and patient characteristics were also important factors for the selection of specific subpopulation. The majority of patients in our study had relatively favorable tumor and patient characteristics: T1, negative lymph node, positive hormone receptor, postmenopausal status, and tamoxifen treatment. Several studies also showed that grade-3 tumor and lymphovascular invasion may be an important factor for the use of radiotherapy [13,27,34]. Whether the definition of favorable tumor characteristics should be based on additional biomarker (i.e., Ki-67, cyclooxygenase-2, and E-cadherin) in early breast cancer was unclear.
In clinical practice, it was important to judge whether the benefit of radiotherapy outweighed its adverse effects. Older patient frequently had worse performance status and higher comorbidity, and radiotherapy-related toxicities and complications may frequently occur [32,33], and the improvement of local control by radiotherapy could not translate into survival. Moreover, radiotherapy could result in inconvenience, worse cosmetic results and adverse effects, such as breast pain, breast edema, and skin-color changes, reducing quality of life [31]. Recently, several studies have showed that accelerated partial breast radiotherapy or intraoperative radiotherapy may become safe, effective and alternative strategies with acceptable toxicities and short time durations [35,36].
The main strength of the current meta-analysis study was that this meta-analysis systematically evaluated the efficacy of radiotherapy in early breast cancer and DCIS through assessing all the relevant end points and included all eligible studies to increase statistical power though increasing the number of patients. However, there were several limitations in the current study. First, this study was based on published data and we could not obtain detailed individual data. Future high-quality, well-designed multicenter studies are required. Second, although the two studies on DCIS showed that radiotherapy could not improve survival for DCIS patients in overall group, irrespective of age, we could not specifically evaluate the survival benefit of radiotherapy in older patients due to lack of data and study number. Future studies are needed to explore the impact of radiotherapy on survival in older patients with DCIS after BCS. Third, our study did not provide a conclusive result regarding the optimal definition of older age. Fourth, although our results indicated that radiotherapy after BCS could slightly reduce the risk of local relapse, there were no studies performing cost-analysis to evaluate whether radiotherapy was a costeffective treatment for local relapse. Thus, further studies are needed to evaluate its cost-effectiveness.
In conclusion, although our study indicates that radiotherapy after BCS could slightly reduce the risk of local relapse in older patients with favorable early breast cancer, radiotherapy can only gain low absolute benefits and high NNT for local relapse. Moreover, radiotherapy cannot translate into significant survival benefits. Thus, our study proposes that radiotherapy options should not only focus on the relative risk of local relapse but also on the absolute benefit, NNT and survival benefits.

Eligibility criteria
Studies were included if they met all of the following criteria: (1) assessment of the efficacy of postoperative radiotherapy after BCS in older patients with early breast cancer or DCIS; (2) the definition of older patients was postmenopausal patient aged ≥ 50 year or patient aged ≥ 55 year regardless of postmenopausal status; (3) the study design was prospective study with at least a subgroup analysis containing older patients; (4) the outcomes of interest were tumor relapse (local relapse, distant relapse, ipsilateral breast tumor relapse [IBTR], and contralateral breast cancer) or/and survival (overall survival [OS], cancer-specific survival [CSS], breast cancer-specific survival [breast-CSS], non-breast cancer-specific survival [non-breast-CSS], non-cancerspecific survival [non-CSS], and other cancer-specific survival [other cancer-CSS]); and (5) the effect estimates and corresponding 95% confidence intervals (CIs) could be extracted directly or calculated from published data indirectly. To retain maximum information, the most recent study was included and the data that only reported in excluded duplicated studies were extracted and added it into the included duplicated study if there were several studies based on the same population or data set.

Data extraction
Studies were reviewed and data were extracted by two reviewers, independently. For each study, the following information was extracted: first author,