Postoperative regional nodal irradiation (RNI) has been an issue of controversy over the last decades as the awareness of late side effects increased. While RNI was part of the treatment in almost all studies investigating postmastectomy radiotherapy (PMRT) [9], only a minority of patients received RNI after breast-conserving surgery (BCS) [11]. Moreover, the term RNI is not uniformly defined. Older studies generally included radiotherapy of the supraclavicular (SCN-RT), axillary (ALN-RT), and internal mammary nodes (IMN-RT). More recently, the use of IMN-RT has substantially declined as it was suspected to enhance cardiac toxicity and morbidity [12]. Likewise, dedicated ALN-RT of levels I–III has been considered obsolete for node-negative patients after axillary lymph node dissection (ALND) or sentinel node (SN) biopsy and even for selected SN-positive patients, for whom implications for radiotherapy will be discussed below.

Final assessment of survival in breast cancer requires a follow-up of 10–15 years, a time period that inevitably implies substantial changes in treatment planning and technique. Therefore, it is difficult to extrapolate results achieved in the past to those achievable in the present or future using modern techniques. This dilemma is exemplarily illustrated regarding literature about RNI: Before the advent of 3D planning as a routine tool and the use of individual contouring of the different lymphatic pathways, inadvertent inclusion of parts of the axillary and mammary nodes in the tangential fields was common but remained unrecognized as no dosimetric evaluation was performed [17, 18]. On the other hand, for intentionally irradiated lymph node areas, the target definition on the basis of anatomical landmarks alone (such as bony structures), may result in an underdosage, undetected for the same reasons [31]. Even in the recent randomized trials, 2D planning was mostly used and the dose to nodal areas accordingly prescribed. As a consequence, exact discrimination of the effect yielded by RNI in comparison to whole breast/chest wall irradiation (WBI/CWI) alone, may be difficult. These limitations also apply for the recent studies that provided new evidence for the benefit of RNI [22, 40, 58]. Two opposite speculations are conceivable: either the benefit of RNI might be more pronounced with high tech three-dimensional (3D) planning, as it ensures adequate dose coverage of the target; or the results may even be inferior, assuming that inadvertent irradiation of the lymph nodes may have contributed to tumor eradication with older techniques.

The aim of the present paper is to provide an overview on recent literature and to discuss the different aspects concerning indications for RNI in light of new data on the basis of a comprehensive literature search. Moreover, the problem of nodal coverage with different techniques will be illustrated by reproducing field arrangements in the different studies and superimposing them on CT slices with individually contoured node areas.

Finally, practical guidelines for targeting and technique of radiotherapy of the lymphatic pathways after surgery for breast cancer will be outlined. The present recommendations are complementing and critically reviewing the interdisciplinary S3 guidelines provided by the German Cancer Society (DKG) published in July 2012 [48] and updating the practical guidelines published in 2008 by the breast cancer expert panel of the German Society of Radiation Oncology (DEGRO) [49].

General indications for regional nodal irradiation

In case of 4 positive axillary nodes, the indication for RNI is undisputed. For patients with 1–3 positive nodes (pN1), data concerning the effectiveness of RNI are less unequivocal [9, 19, 56, 62]. This incertitude is reflected by the heterogeneity of recommendations in different international guidelines. Indirect evidence for the benefit of RNI for patients with 1–3 positive axillary nodes was derived from subgroup analyses of the Danish Breast Cancer Group study and the British Columbia trial. Both randomized studies yielded an improvement of survival if postmastectomy radiation (PMRT) including the complete lymphatic pathways was used. Comparison of outcome according to the number of metastatic nodes revealed that patients with 1–3 and those with ≥ 4 positive nodes had a similar absolute overall survival (OAS) advantage of roughly 10 % [38, 42]. As no comparison was made to chest wall irradiation alone, uncertainty remained how to quantify the contribution of RNI.

Patients allocated for primary systemic treatment (PST) should preferentially receive sentinel node biopsy (SNB) prior to treatment [28]. For patients with pathologically negative nodes after PST, who had initially presented with clinical lymphadenopathy, the decision for RNI should comply to the pretreatment stage, irrespective of the response to chemotherapy [1, 5, 35, 48]. This recommendation is not based on a high level of evidence as no randomized studies have been performed to assess the potential benefit of RNI in this situation. Data from retrospective studies are conflicting [13, 60]. Interestingly, in one series, the largest benefit of PMRT (including RNI) was observed for the subgroup of patients with a complete remission after PST (33 % vs.77 % p = 0.0016) [33]. Especially with regard to the emerging role of PST, a prospective controlled trial is highly warranted.

New evidence from randomized trials

Canadian trial (NCIC-CTG MA.20)

Beyond retrospective data, the first randomized study providing evidence for the benefit of RNI especially in patients with 1–3 lymph nodes (LN), was the NCIC-CTG MA.20 trial, presented at the ASCO 2011, not yet published as full paper [58]. The study comprised 1832 women with mostly 1–3 positive axillary nodes (85 %) and a minority of women (10 %) with negative nodes in the presence of high-risk factors (G3, lymphatic vessel invasion, T3 tumors, T2 tumors with < 10 axillary lymph nodes removed and HR-negative). Patients were randomized after breast-conserving surgery and ALND to either whole breast irradiation (WBI) or WBI and additional RNI. The target volume in the RNI group included levels I–III of the axillary nodes, supraclavicular and internal mammary nodes (Fig. 1 b). The 5-year locoregional recurrence-free survival was 96.8 % with and 95.5 % without RNI (p = 0.02). The 5-year disease-free survival (DFS) was significantly improved in the RNI group: 89.7 % vs. only 84 % patients with WBI alone (p = 0.003). This difference is twice as high as the absolute benefit in terms of local control, and therefore hypothetically attributable to the significant positive impact on distant metastases-free survival (DMFS) with an absolute 5.4 % reduction at 5 years in the RNI arm (p = 0.002). There was a trend towards improved OAS (92.3 % vs. 90.7 %), however, just below statistical significance (p = 0.07). The rate of lymph edema (any grade) was 4 % without vs. 7 % with RNI (p = 0.004) which is in accordance to recent literature [51]. The rate of pneumonitis was slightly increased after RNI (1.3 %vs.0.2 %) but altogether low. Data concerning cardiovascular toxicity were not yet provided. The authors concluded that RNI reduces the risk of locoregional and distant recurrence and improves DFS with a trend in improved OAS [58].

Fig. 1
figure 1

Regional lymph node and whole breast irradiation: Differences in portal designs in the EORTC and MA-20 study and in corresponding dose distributions, respectively. Frontal view of field projections with respect to 3D-contoured lymphatic regions according to respective protocol guidelines: a EORTC trial, b MA-20 trial. Dashed yellow lines indicate cross sections for demonstration of dose distributions (c, d, e, f, g, h). Cranial cross sectional views through supraclavicular fields and corresponding dose distributions for the c EORTC trial and d MA-20 trial. Upper planes through tangential fields for EORTC alignements including a separate portal for e IMLN and f MA-20 beam arrangements (“wide tangents”). Midplane distributions for tangential beams for g EORTC, and h MA-20 alignments. ALN I axillary lymph nodes level I, ALN II axillary lymph nodes level II, ALN III axillary lymph nodes level III, IMLN internal mammary lymph nodes, SCLN supraclavicular lymph nodes, PTV planning target volume

Full size image

European study EORTC 22922-10925

The European study EORTC 22922-10925 [40] (not yet published as full paper either) included 4004 women stage I–III with mostly pT1–2 tumors (95 %) and either involved axillary LN (55.6 %) and/or a medially located primary tumor (44.5 %). Patients were randomized after BCS (76.1 %) or mastectomy (23 %) to receive WBI/PMRT either with or without inclusion of the IMN and medial SCN (MSCN) (Fig. 1 a) with 50 Gy in 25 fractions. After mastectomy, chest wall irradiation was applied to 73.2 % of patients in both arms. Dose specifications for WBI/CWI were not required, presumably, relevant parts of the axilla were included in these fields. Nearly all LN-positive (99.0 %) and 66.3 % of the LN-negative patients received adjuvant systemic treatment. After adjustment for stratification factors, IM-MS RT significantly improved outcome at 10 years: OAS: 82.3 vs. 80.7 %, p = 0.049; DFS: 72.1 vs. 69.1 %, p = 0.044; metastases-free survival (MFS) 78.0 vs. 75.0 %, p = 0.020. The treatment effect on OAS was similar for pN1 vs pN2 patients but interestingly, most pronounced for node-negative patients (HR 0.79, 95 % CI 0.61–1.02). In case of > 10 positive nodes, no advantage was observed; however, the number of patients (about 2 %) was small. The highest benefit was observed in patients receiving chemo- as well as endocrine therapy (HR 0.72, 95 %CI 0.55–0.94). While the local recurrence rate was similar (5.3 vs. 5.6 %), regional lymph node recurrence was 2.7 % with RNI vs. 4.2 % without. In contrast to the findings in the Canadian study, the rates of any-grade lymph edema at 3 years were identical in both groups [32]. No increase in lethal complications was observed so far. Therefore, the authors conclude that radiotherapy of the IMN and medial SCN should be recommended for patients with involved axillary LN and/or medially located primary tumor.

French study

In the French trial [22], 1334 patients with mostly T1–2 tumors (85 %) and either positive axillary nodes (75 %) or central/medial tumors irrespective of nodal status (25 %) were included. All patients were treated with PMRT to the chest wall, including SCN (plus axillary apex, in node-positive cases) and were randomized to receive additional IMN-RT or not. IMN included the first 5 intercostal spaces. Roughly 60 % received chemotherapy, about 50 % endocrine systemic treatment. Overall, 10-year OAS was 62.57 % with IMN-RT and 59.3 % (p = 0.8) without. Node-negative patients (25 %) showed a trend towards a worse outcome with IMN-RT (n.s.), whereas node-positive patients seemed to profit from IMN-RT (n.s.). In patients with lateral tumors, who were treated with chemotherapy, the 10-year OAS was 67 % vs. 64 % (n.s.) in favor of IMN-RT, without chemotherapy 55 % vs. 50.5 % (ns) respectively. Node-positive patients with medial/central tumors had a worse outcome compared to lateral ones, for whom 10-year OAS after chemotherapy was 57 % vs. 49.8 % and 54 % vs. 47.5 % without chemotherapy both in favor of IMN-RT, but also without statistical significance. The authors concede that the study may have been underpowered to prove a significant survival benefit for IMN-RT. No increase in cardiac toxicity was observed in the IMN-RT group but may have been incompletely reported; the authors admit that their data do not permit a definite conclusion regarding toxicity and conclude that they cannot reliably recommend for or against IMN-RT [22].

Meta-analysis of these trials

A meta-analysis of these data [6] revealed a more distinct benefit of RNI on OAS with a hazard ratio of 0.82 (p = 0.011) (Tab. 1). The largest gain was observed for DMFS, possibly supporting the hypothesis of Hellman [21] that radiotherapy is “stopping metastases at their source”.

Tab. 1 Meta-analysis of survival data from two randomized trials
Full size table

Comments and conclusions of the DEGRO panel

  • Data from the MA-20 and EORTC studies suggest that all node-positive patients profit from comprehensive RNI including SCN and IMN.

  • The respective contribution of RNI by site (SCN vs. IMN) on improved outcome cannot be distinguished, axillary nodes of level I and II were partly included in the control arms using WBI/CWI.

  • Full publication of the MA-20 and the EORTC studies is pending; several important details or subgroup analyses are not yet available.

  • No increase in cardiovascular toxicity by RNI was reported. However, follow-up of the MA- 20 study is too short to assess long-term toxicity; in the French study, exclusion of toxicity was not possible according to the authors. In the EORTC study, no increase of lethal complications was observed, however, without providing details how cardiac toxicity was assessed. Moreover, 10 years of follow-up may still be too short for final conclusions.

  • Two of the studies permit comparison between the effect on locoregional and distant recurrence. In both trials the impact on MFS is greater than the difference in locoregional control, strongly suggesting that RNI prevents distant spread.

Is RNI still obsolete for pN0 patients?

Statement of the German S3 Guidelines 2012 [48]

Regional nodal irradiation by site

As separate irradiation of the different nodal areas has hardly ever been investigated in a randomized setting, it is impossible to quantify the contribution to improved tumor control achieved by RT of the distinct sites. In most studies, RNI comprised either all or at least two sites.

Radiotherapy of the internal mammary nodes

Risk factors for metastatic spread to the internal mammary nodes (IMN) have been elaborated in a large series of 1679 Chinese women who underwent extended mastectomy including dissection of the mammaria interna nodes. Patients with following conditions had high risk of IMNS metastasis: 4 or more positive ALNs, medial tumor and positive ALNs, T3 tumors, and age less than 35 years. The incidence of IMN metastases for those patients was more than 20 %. Positive IMN in accordance to the number of positive ALN with respect to index tumor location: 1–3 LN medial: 23 % and central 32 %; for those with 4–6 LN: medial 47 %, central 22 %. This is in contrast to the small number of patients diagnosed with clinically manifest IMN recurrence which is assumed to be only about 1 % [17]. A possible explanation for this discrepancy is the lack of routine imaging of this area as part of the follow-up program. Moreover, parts of the IMN may be inadvertently included in tangential field arrangements (Fig. 1 a, b). Another hypothesis is that micrometastasis in the IM-LN may represent a source for metastatic spread without growing to clinically detectable size before distant metastases have been diagnosed [6].

Statement of the German S3 Guidelines 2012 [48]

Targeting, technique, and dose for RNI

Three-dimensional treatment planning is mandatory; several anatomically-based instruction guidelines have been published to define individual contouring of the different lymph node regions [14, 25, 26, 45]. Substantial variations may be caused by patient positioning; especially the abduction of the arm plays an important role [26].

Internal mammary nodes

The target volume should be restricted to the ipsilateral side and not exceed below intercostal space 3–4 ([14], Fig. 1 a, b). Contouring of the heart is mandatory, detailed instructions are provided by the atlas of Feng et al. [22]. Delineation of the whole heart should start directly inferior to the left pulmonary artery. The left anterior descending artery (LAD) is one of the clinically most relevant structures for late toxicity, as it is the major coronary vessel in the closest vicinity to left-sided IMN. The LAD originates from the left coronary artery and runs in the interventricular groove between the right and left ventricle. Contouring and dose documentation in a dose–volume histogram may be helpful for quality control.

Supra/infraclavicular nodes

The term SCN is synonymous with inferior deep cervical LN. The medial part corresponds to the level IV neck nodes, while the lateral part corresponds to level Vb according to the AJCC classification of head and neck area [2].

The medial SC LNs consist of the inferior jugular nodal chain and the medial part of the transverse cervical nodal chain. When the target volume is restricted to this area (as in the EORTC 22922-10925 trial), the lateral field border extends to the mid clavicle. The lateral SC LNs consist of the lateral part of the transverse cervical nodal chain, which is the inferior part of the posterior triangle of the neck; in this case, the field extends to the coracoid process [14]. The term infraclavicular nodes is synonymous with apical or level III of the axillary nodes and are mostly included when SCN-RT is described (Fig. 1 c, d).

Axillary nodes

The ALN are divided into level I (low axilla) which extends laterally from the pectoralis minor muscle and level II (mid axilla) between the medial and lateral borders of the pectoralis minor muscle and the interpectoral LN (Rotter). Level III (apical axilla) corresponds to infraclavicular LN (Fig. 1 a, b). When ALN -RT is intended, the SCN are included and the lateral field border is extended according to the individual anatomy.

Dose and fractionation

Statement of the German S3 Guidelines 2012 [48]