Our results indicate that sandwich RT contributed to improvement in 8-year DFS including locoregional recurrence, but not in distant metastasis, in node-positive breast cancer patients. The impact on LRR depends on the start of RT rather than of CT. A randomized control trial on sequencing of CT and RT after BCS for patients with four or more positive nodes found that there was a higher local recurrence rate in the CT-first group and a higher distant recurrence rate in the RT-first group [11]. This study suggests that a high distant recurrence rate in the RT-first group may be caused by the lower drug doses in the CT-first group. The result of effect on LRR is consistent with our study. Also, in our study that analyzed patients who received the same doses of CT, there was no significant difference between the CT followed by RT group and sandwich RT group in distant metastasis or OS. On the other hand, a research suggested that a treatment schedule with CT first cannot compromise local control despite of the delay in RT [12]. A retrospective study showed that delay in starting RT for patients who underwent BCS and received chemotherapy before RT did not compromise 10-year outcomes. However, in this study, patients who received chemotherapy followed by RT and subsequent chemotherapy were not included. All of patients were treated with RT after completion of chemotherapy although time interval between surgery and RT was shorter than 112 days. Patients with N0 disease also were included and axillary lymph node dissection was not performed for some patients with node-positive disease [13] .
RT has impacts on LRR and OS in breast cancer. Several studies suggested that postoperative RT after systemic therapy could improve OS in node-positive breast cancer patients [9, 14–16]. Locoregional RT within six months of beginning chemotherapy improved both disease recurrence and mortality in a meta-analysis that reviewed 18 randomized controlled trials with 6,367 participants [3]. In contrast, several randomized clinical trials demonstrated that postoperative adjuvant RT reduced locoregional recurrence but did not affect OS [2, 17]. After 10 years of follow-up in MA.20 [18] and in EORTC 22922 [19], regional nodal irradiation was found associated with improvement in locoregional and disease-free survival and lower breast cancer mortality, but not overall survival. RT sequence did not affect 8-year OS in our study. The reason why RT sequence did not correlate with OS, even though high-risk patients were analyzed in this study, may be that all of the patients received RT and there was a relatively small difference of interval from surgery to RT between the two groups. The eight years of follow-up and the small number of patients studied may also be the cause.
The timing of effective RT varies. The interval from surgery to the start of RT is determined according to the CT and RT sequence. In our study, RT was started 5.5 months after surgery in the CT followed by RT group and there was an impact on LRR despite of the difference of 2.5 months between the two (assessed) groups. Hickey et al. reported that starting RT within 8 months after surgery did not reduce survival [16]. Huang et al. demonstrated that patients who received RT for more than 8 weeks after breast cancer surgery were 2.28 times more likely to develop locoregional recurrences [20]. A systemic review suggested that the delay in RT caused an increase in local recurrence but it was not related to distant metastasis or OS regardless of whether patients received CT or not [21]. These results are consistent with our study. The timing of RT can affect not only the local recurrence, but also DFS and OS. Delayed RT, 6 months after surgery, increased local recurrence rate and led to significantly poorer OS and DFS [22].
It is important to note whether chemotherapeutic drugs are administered while RT is delayed. As CT has a detrimental effect on survival as a systemic therapy and RT has an established role in the prevention of local recurrence, which treatment begins first can influence disease outcome in patients who have risk factors for recurrence. The results from randomized trials of CT followed by RT versus RT followed by CT showed that there is no significant difference between RT sequences in mortality and local or distant recurrence [16]. A recent meta-analysis showed that survival and recurrence in patients who received sandwich RT was not significantly different from that of patients who received concurrent RT. In contrast, the OS of patients who underwent concurrent chemoradiation was superior to that of sequential RT. It was found that recurrence was higher when chemotherapy was started before radiotherapy than when radiotherapy was initiated first. Avoiding delay in RT or no maintaining no gap between chemotherapy and RT might lead to their additive interaction and tumor response [23]. Our study found that sandwich RT decreased LRR and improved DFS despite all of the cases having negative margins and node-positive disease. These results may support the theory that sandwich method with less delay in starting RT can improve survival and reduce recurrence.
The evidence for effectiveness of sandwich RT is limited [11, 22, 24, 25]. Most previous studies on sandwich RT have included patients undergoing chemotherapy with CMF or anthracycline regimen. RT was delivered between CT cycles in these studies [7, 8]. These may support that a sandwich RT can be applied with the CMF regimen but RT insertion between cycles of chemotherapy using anthracycline is not recommended [26]. We showed that sandwich RT between AC and taxane can provide DFS benefit regardless of BCS or mastectomy. Taxanes added sequentially with AC chemotherapy improved DFS despite the delay in starting RT. Henderson et al. suggested that taxanes were the main factors that led to the improvement in survival compared in the only AC chemotherapy followed by RT group and the AC plus taxanes followed by RT group [27]. In our study, we showed the effect of only RT sequence, excluding addition of taxanes on DFS.
We found that sandwich RT improved LRR and DFS, especially in luminal A subtype in the subgroup stratified by molecular subtypes. This result is consistent with the data of Wang et al., which suggested that adjuvant radiotherapy reduced the risk of relapse in luminal A breast cancers [28]. The benefits from chemotherapy were generally smaller in luminal A breast cancers [29]. As the potential effect of CT is relatively small in luminal A subtype, it may not be able to dilute the impact of RT sequence on treatment outcome although while RT is delivered is delivered could allow the proliferation of micr-metastatic disease [20]. The effect of RT on OS in our study was different from the study of Mao et al. who analyzed patients treated with or without RT and showed a significant survival benefit after radiotherapy in younger patients (age at diagnosis <60 years) with luminal A subtype [30]. Some patients with little benefit from CT might be included in this study because multigenomic assays used to identify patients at increased risk for distant recurrence could not be performed during this study period. However, clinically high-risk tumors with node-positive disease were more frequently high-risk by multigenomic assays [31].
In the multivariable analysis of our study, RT sequence was a stronger factor for LRR rather than the extent of breast resection and the status of the resection margins which are related to the tumor burden after surgery. The RT sequence of whether sandwich RT or CT followed by RT can affect the length of time required for the remaining tumor cells to proliferate before radiation therapy. These findings can support that delaying RT, while chemotherapy is treated first, could increase local recurrence rates and delaying of systemic chemotherapy. As RT is more effective leading to less residual tumor burden, the residual tumor regrowth after surgery can increase because of a long interval between surgery and RT and lead to poorer outcomes [20].
This study support that RT can be given between anthracyclines and taxanes in node-positive luminal A subtype breast cancer patients and it can improve locoregional recurrence free survival. Delay of starting endocrine therapy did not affect overall survival when CT and RT were completed. Although literature on effect of RT sequence on toxicity or quality of life is very limited, theoretically, optimal timing of CT and RT can alleviate toxicity. As subsequent CT may increase hematological toxicities of initial chemotherapy, sandwich RT may make these toxicities to be more manageable. It may lead to improve quality of life when subsequent CT is administered and to increase capacity to maintain full-dose CT on the planned schedule. Since more effective local control can be beneficial only in patients at risk for local recurrence among luminal A subtype patients, there is a need to identify more robust predictors of local recurrence such as the expression profiling [32].
Our study has several limitations. Because of the small sample size, especially for the HER2 subtype, we could not analyze each four subtypes and compare luminal A subtype with non-luminal A subtype, including luminal B, HER2, and TNBC. Also, the low numbers of events occurred although patients of two groups had similar baseline characteristics and all of patietns were treated with anthracyclin and taxane. We need to follow disease outcomes in patients included in this study with longer follow-up periods to see if there is any change in the overall survival in longer term. Breast cancer is regarded as a systemic disease that spreads by local extension [33]. Whelan et al. suggested that RT may inhibit secondary systemic spread by reducing locoregional recurrence to improve survival when systemic CT is given [3].