According to the global cancer statistics 2020, female breast cancer (BC) ranked the most commonly diagnosed cancer, accounting for 11.7% of total tumors with an estimated 2.3 million new cases (1). Radiation therapy (RT) played a key role in the management of BC after breast-conserving surgery or mastectomy, and regional nodal irradiation (RNI) was widely used for the treatment of node-positive BC due to its potential survival benefit (2). Since the publication of two large phase III randomized controlled trials MA-20 (3) and EORTC-22922/10925 (4), comprehensive RNI including supraclavicular lymph nodes (SVCs) and internal mammary lymph nodes (IMNs) had become the standardized adjuvant RT therapy for high-risk node-negative or involving one to three node-positive BC in international guidelines. However, despite an abundance of data supporting the benefit of comprehensive RNI for BC, uncertainty existed regarding the appropriate RNI volume for BC. Indeed, significant inter-physician variability existed in real-world practice especially in IMN irradiation (IMNI) because the inclusion of the IMN would lead to cardiopulmonary extra doses, which increase the risks for developing late adverse events such as cardiac events and pulmonary fibrosis. In a more recent randomized clinical trial (KROG 0806), the authors failed to demonstrate superiority of the IMNI group over those treated without IMNI for node-positive BC patients because of sample size calculation issue, poor protocol compliance, and RT quality assurance issue (5). Thus, an unanswered question about the optimal regional nodal irradiation volume for BC patients remains, and the optimal RNI strategy gained great interest for radiation oncologists with the aim of archiving maximal clinical benefit of RNI, with the minimal possible radiotherapy toxicity to maintain the quality of life. As a result, we performed the present comprehensive meta-analysis to investigate the efficacy of radiotherapy with IMNI vs. RNI without IMNI in BC patients. In addition, we also compared the efficacy difference of RNI regimens head-to-head via network meta-analysis in terms of combined clinically meaningful overall survival (OS), disease-free survival (DFS), and distant-metastasis-free survival (DMFS) benefits.

Over the past decades, primary breast tumor resection followed by adjuvant radiation to the lymphatic drainage for both early-stage and locally advanced BC had become the standard treatment strategy. However, the treatment volume of RNI significantly varied among those published studies. Thus, there remained a matter of debate on which regional nodal volume should be irradiated for BC patients in the era of effective systematic therapies (35). Prior to the present study, Budach et al. performed two meta-analyses to define the optimal RNI volume based on four prospective randomized trials, and the pooled results indicated a statistically significant improvement in overall survival for CW/WB+ comprehensive RNI when compared to CW/WB alone in stage I–III breast cancer (36, 37). In 2019, the Early Breast Cancer Trialists’ Collaborative Group conducted an updated meta-analysis to investigate the efficacy of regional nodal irradiation in early-stage BC patients according to treatment periods and found that RT to regional lymph nodes in older (1961–1978) studies increased the overall risk of death, while nodal RT in more recent (1989–2003) studies reduced breast cancer recurrence, breast cancer mortality, and overall mortality (38). Since the publication of these meta-analyses, more clinical trials had been performed in recent years, but the optimal treatment RNI strategy remained undetermined. We therefore performed this meta-analysis to assess the benefit of different RNI treatments in BC patients and compare the efficacy of different RNI regimens via network meta-analysis.

A total of 29,640 BC patients from 20 studies were included for analysis. In consistent with previous result (39), our pooled HR demonstrated that radiotherapy with IMNI in BC patients significantly improved OS, DFS, and DMFS when compared to RNI without IMNI (p<0.05). We then performed a sub-group analysis according to RNI volume and found that CW/WB+RNI with IMNI significantly improved OS (HR 0.85, p<0.01), DFS (HR 0.65, p<0.01), and DMFS (HR 0.90, p=0.02) when compared to CW/WB+RNI without IMNI. In contrast to the report of Budach et al. (36), no significant difference could be found between CW/WB+RNI with IMNI vs. CW/WB alone in terms of OS (p=0.19) and DFS (p=0.48), while a significantly decreased risk of developing distant metastasis could be observed in CW/WB+RNI with IMNI group when compared to CW/WB alone (p=0.04). Additionally, CW/WB+ mixed RNI did not improve OS, DFS, and DMFS when compared to CW/WB alone (p>0.05). We further explored the optimal RNI treatment strategy for BC patients by using network meta-analysis. Based on those published studies, CW/WB+RNI with IMNI remained a better treatment option for preventing BC progression and death, and CW/WB combined with comprehensive RNI could be recommended for BC patients after curative surgery.

IMNI would inevitably increase RT dose to the heart and lungs regardless of using modern radiation technology, which might be associated with an increased risk of developing RT-related toxicities (40). But in a large cohort study based on SEER database, Giordano et al. (41). found that the risk of death from ischemic heart disease associated with radiation for breast cancer had been substantially decreased over time. In a more recent meta-analysis of three randomized trials, Budach et al. found that IMNI was not associated with an excess of cardiac death or cardiac toxicity rate (36). Although there was a lack of formal guidelines on whether or not to treat IMN for BC patients, radiation oncologists in our institute had achieved a consensus that CW/WB+RNI with IMNI could be recommended for BC patients due to the low incidence of cardiac toxicities. Between June 2017 and August 2019, we performed a prospective randomized controlled trial to investigate the early cardiac event in pre-specified dose constraints for the heart vs. choice of radiation oncologists for breast cancer treated with postoperative intensity-modulated radiotherapy, which had been registered in ClinicalTrials.gov (NCT02942615). By 31 December 2019, 143 patients completed 1-year follow-up (77 in study group and 66 in control group). The Dmean of the heart was 374.9 ± 205.3cGy and 376.7 ± 204.7cGy in the study and control group, respectively (p=0.96). No clinical cardiac toxicity was observed. Subclinical cardiac events occurred in 29 patients in the study group and 29 in the control group (p=0.45) (42). Therefore, mean heart dose in BC treated with IMNI was low and did not increase the risk of developing early-stage cardiac toxicities. As for late cardiac toxicities, EORTC Trial 22922/10925 reported 15-year side effects after IMNI and showed that IMNI did not significantly increase the risk of developing cardiac fibrosis when compared to without IMNI group (1.1% vs. 1.9%, p=0.07) but not for any cardiac disease (9.4% vs. 11.1%, p=0.04). RT technique improvement would also impact the cardiac toxicity of RT. The Danish Breast Cancer Group performed a population-based study and found a higher risk of cardiac toxicity in left- vs. right-sided patients irradiated during the non-CT-based period, while no increased risk of coronary artery disease in left-sided versus right-sided patients was observed in the CT-based period (43). As a result, with the improvement of RT techniques, IMNI could be safely applied for BC patients, even in the left-side BC.

Another major concern was that BC was a heterogeneous disease, and the risk of developing recurrence significantly varied. For BC patients with multiple risk factors, including young age, large tumor size, lymph vascular invasion, medial/central tumor location, or high nuclear grade (44), the incidence of local regional recurrence could increase to 20%. Therefore, one size did not fit all, and not all BCs could benefit from comprehensive RNI. Based on our established clinical risk model for N1 breast cancer (45, 46), our group initiated a prospective randomized clinical trial to investigate whether a part of pN1 breast cancer patients could be safely omitted from IMNI by using a clinical-genomic model, and the trial was under recruiting(NCT04517266). The preliminary results of enrolled 75 patients showed that among clinically high-risk pN1 BC, which is defined as having at least two of the five clinical risk factors (age≤40, three positive LN(lymph node), T2 stage, grade 3, and Ki-67 index≥14%), 70% of them present with genomic high risk, and 30% present with genomic low risk, and whether those clinical high-risk but genomic low-risk BC patients could be omitted from IMNI still needs long-term follow-up of our research (47). Another potential method was performing lymphoscintigraphy by injecting radiotracer for axillary sentinel node biopsy and peritumor to identify high risk for IMN metastases. It had been reported that lymphoscintigraphy not only identifies axillary sentinel node biopsy but also depicts drainage to the internal mammary (IM) basin, present in approximately 20% of patients (48, 49). Internal mammary sentinel lymph node biopsy (IM-SLNB) was another way to clearly determine the status of IMN metastases, although routine performance of IM-SLNB in clinical axillary lymph node (ALN)-negative patients remained debated. Qiu et al. (50) conducted a prospective cohort study and found that IM-SLNB visualization rate was 71.9% among BC patients who received initial surgery and 33.1% among BC patients treated with neoadjuvant systemic therapy.

In the real-world practice, there is a severe heterogeneity issue of RNI administration. In the ALTTO trial, only a third of the pN1 patients (36.8%) received RNI, while 82.2% of patients with four or more positive lymph nodes were treated with RNI. Among patients treated with RNI, 60.9% of the patients RNI targeting only one regional node area, while only 3.9% received RNI targeting the three regional nodal areas (24). One possible explanation for this finding was that the ALTTO trial was conducted between 2007 and 2011, before the first publication of the MA.20 and EORTC 22922/10925 trials. The quality of the regional nodal plan was another issue that should caution radiation oncologists. Ling et al. performed a network to determine the compliance with regional nodal coverage, contouring quality, target coverage, and organ-at-risk dosimetric parameters and found that 18% of plans presented with unacceptable nodal contour quality and 15% of them had inadequate coverage (51). In the KROG-0806 trial, the individual case review demonstrated that overall protocol compliance, including IMNI, significantly varied, and only 59.0% of the prescribed dose was delivered to the IMNI group (52, 53).

Nevertheless, there were some limitations that needed to be concerned of. First, both prospective and retrospective studies were included in this meta-analysis. Although the quality of included studies was high, selection bias between groups could not be avoided despite the fact that we performed subgroup analysis according to RNI volumes that successfully reduced the heterogeneity to a low grade. Second, this study was conducted at the base trial level but not at the individual level. Therefore, we could not perform pooled analysis according to patient characteristics, such as nodal stage or tumor location, although it had been established that medial-located tumor is a well-known predictor for IMNI benefit. In addition, both early-stage and locally advanced BC were included in the present meta-analysis, which might be another source of heterogeneity. Finally, the present study included both old or modern systematic therapy, and advances in systematic treatment in the modern era might affect the benefit of RNI, which might be another source of heterogeneity.

In conclusion, the present study demonstrated that RNI with IMNI yielded a significant survival advantage for BC patients. Subgroup analysis according to RNI volumes showed that CW/WB+RNI+IMNI significantly improved DFS and DMFS when compared to CW/WB+RNI without IMNI, but not for OS. NMA found that CW/WB+RNI+IMNI was the optimum RNI treatment strategy for BC patients that reduced mortality and disease progression. Additionally, further studies evaluating the impact of RNI volume on late cardiac and lung toxicities are strongly needed.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.

Conceptualization: JC and W-XQ. Project administration: CX and LC. Methodology: W-XQ, GC, and JC. Data curation: JC and CX. Formal analysis: W-XQ. Manuscript preparation: W-XQ, CX, GC, JC, and LC. Final approval of manuscript: all authors.