Breast Radiotherapy (RT) Using Tangential Fields (TgF): A Prospective Evaluation of the Dose Distribution in the Sentinel Lymph Node (SLN) Area as Determined Intraoperatively by Clip Placement

Background Randomized trials have established that patients with limited involvement of sentinel lymph node (SLN) do not require axillary lymph node dissection (ALND). The similar outcome in patients with ≤2 positive SLN with or without additional ALND is attributed, in part, to tangential fields (TgF) RT. We evaluated the dose distribution in the SLN biopsy area (SLNBa) as determined intraoperatively by clips placement for radiotherapy (RT) optimization. Methods This prospective study included 25 patients who had breast conservation. Titanium clips were used intraoperatively to mark the SLNBa. All patients had 3D-conformal RT using standard (STgF) or high tangential fields (HTgF). Axillary levels, SLNBa, and organs at risk were contoured on a CT scan. Dose distribution and overlap between TgF and target volumes were analyzed. Results The average doses delivered to axilla levels I-III and SLNBa were 25, 5, 2, and 33 Gy, respectively. The average dose delivered to SLNBa was higher using HTgF with better coverage of the axilla. Only 12 of 25 patients (48 %) had their SLNBa completely covered by the TgF. There was no impact of TgF size on ipsilateral lung dose. The mean heart dose delivered using STgF was lower than HTgF. Conclusions In the era of SLNB, axilla and SNLBa RT technique has to be standardized to deliver adequate dose. We recommend the use of HTgF or direct axillary RT techniques (such as in AMAROS trial) in patients with metastases in SLN without ALND completion, when only TgF are expected to cure potential residual disease in the axilla.

scan. Dose distribution and overlap between TgF and target volumes were analyzed. Results. The average doses delivered to axilla levels I-III and SLNBa were 25, 5, 2, and 33 Gy, respectively. The average dose delivered to SLNBa was higher using HTgF with better coverage of the axilla. Only 12 of 25 patients (48 %) had their SLNBa completely covered by the TgF. There was no impact of TgF size on ipsilateral lung dose. The mean heart dose delivered using STgF was lower than HTgF. Conclusions. In the era of SLNB, axilla and SNLBa RT technique has to be standardized to deliver adequate dose. We recommend the use of HTgF or direct axillary RT techniques (such as in AMAROS trial) in patients with metastases in SLN without ALND completion, when only TgF are expected to cure potential residual disease in the axilla.
An extensive literature, including seven randomized trials and B32 trial update at 10 years, has established that axillary lymph node dissection (ALND) is not required in patients with negative sentinel lymph node biopsy (SLNB) in which axillary recurrence is rare. 1,2 SNLB represents the standard procedure for patients with early breast cancer (BC) and clinically node-negative (cN0). Thus, Saint-Gallen guidelines state that ALND should not be completed in cN0 patients with one to two macrometastatic (MAC) in the SLNs after breastconserving surgery (BCS) and tangential field (TgF) Yazid Belkacemi, Elie Calitchi, Romain Bosc belong to AROME Association of Radiotherapy and Oncology of the Mediterranean Area (www.aromecancer.org). Francilian Breast Intergroup (FBI) Paris France.
Ó The Author(s) 2014. This article is published with open access at Springerlink.com radiotherapy (RT). 3 The American Society of Clinical Oncology (ASCO) updated guidelines concluded recently that women with one to two metastatic SLNs planning to undergo BCS with whole breast radiation therapy (WBRT) should not undergo ALND. 4 Moreover, in the ACOSOG Z0011 trial, 6-year outcome after BCS plus WBRT was equivalent in SLNB and ALND patients with B2 positive SLNs. This equivalence was attributed to the potential cure of axillary residual disease with systemic therapy and TgF RT. 5 While radiation parameters and dose distribution in the axilla were not reported in the initial publication, Jagsi et al. tried recently to detail radiation treatments from the 605 available RT report forms. 5 No clear conclusions could be drawn from the analyses on whether additional regional nodal RT was necessary or beneficial for these patients. 6 The utility of TgF RT has been established as the standard of care. The issue that remains outstanding relates to the benefit to include the lymphatics.
Our study was undertaken to determine the dose distribution in the sentinel lymph node biopsy area (SLNBa) marked intraoperatively by clips. This could be helpful for RT optimization when only TgF are used for WBRT in patients with SLN involvement without ALND completion.

METHODS
This prospective study included 25 patients who have undergone BCS in a single institution between April 2012 and March 2013. The Henri Mondor Breast Center Multidisciplinary Committee has approved the protocol. Patients' characteristics are presented in Table 1.

Surgery Procedure
All patients underwent BCS and SLNB procedure for invasive BC (tumor size\3 cm, cN0). SLNB mapping was performed using technetium-99 m-labelled human albumin colloid particles. Blue dye was not systematically administered. SLNs were identified with a gamma detecting probe and/or blue dyed. After SLNs removal two titanium clips were placed to mark the location. No ALND was performed in this study.

Radiation Therapy Technique
All patients had 3D-conformal RT. Two radiation oncologists contoured axilla nodal volumes, SLNBa, and organs at risk using the RTOG contouring atlas. 7 The WBRT technique and indications followed the French guidelines described elsewhere. 8 Height of TgF was defined individually to target the breast volume. This study evaluated SLNBa coverage by TgF as determined intraoperatively by clips placement. For height TgF analyses, STgF was defined with the superior border set at 2 cm below the humeral head, whereas HTgF consisted of a superior border placed at the inferior edge of the humeral head.
The SLNBa was defined as a clinical target volume (CTV SLNB ) with 5 mm in diameter surrounding the clips. To account for position uncertainties, we defined SLNBa planning target volume (PTV SLNB ) with a 10 mm extension around the CTV SLNB (Fig. 1). Only two patients had seroma in SLNBa with a maximum diameter of 24 and 10 mm. In both cases, clips were not displaced by the seroma cavity. The latter was included in the PTV SLNB (Fig. 1). . a Example of clips topography in the sentinel lymph node biopsy area without associated seroma and tangential fields including totally the PTV SLNB . CTV SLNB clinical target volume of the sentinel lymph node biopsy area (in red); PTV SLNB planning target volume of the sentinel lymph node area (in light blue). b Example of clips topography in the sentinel lymph node biopsy area with associated seroma and tangential fields including partially the PTV SLNB . TgF were not adjusted to include totally the SLNBa in patients with negative SLN status. CTV SLNB clinical target volume of the sentinel lymph node biopsy area (in red); PTV SLNB planning target volume of the sentinel lymph node area (in blue). c Example of PTV SLNB and axilla levels CTV coverage by 95 and 50 % isodoses using standard tangential fields Dose-volume-histograms were analyzed according to axilla volumes receiving 95 % (V95) or 50 % (V50) of the prescribed dose. All values were compared according to the use of STgF (n = 20) or HTgF (n = 5). Overlaps between the TgF and the PTV SLNB were analyzed in three groups of TgF-PTV SLNB overlap percentages: 100 % overlap (''suitable group''; PTV SLNB completely in the TgF), C50 % overlap (''partially suitable group''; PTV SLNB partially in the TgF), and 0-49 % overlap (''unsuitable group''; B49 % of the PTV SLNB or completely outside the TgF; Fig. 2). Dose distribution was calculated in each of the three groups.

Statistical Analyses
All comparisons and correlations were performed using t tests using SPSS software. Multiples comparisons were analyzed using an ANOVA post hoc Bonferroni. The level of significance was stated at p \ 0.05.

RESULTS
The median number of harvested SLNs was 1 (average 1.6; range 1-6). One patient had micrometastasis (MIC; [0.2-2 mm) and one had 1 MAC out of 4 SLNs, respectively. Adjuvant systemic therapy and WBRT parameters are presented in Table 1.

Dose and Volume Coverage
The average doses delivered to axillary levels I, II, III, and SLNBa were 25, 5, 2, and 33 Gy, respectively.   Table 3.

SLNB Area Coverage
In the STgF group (n = 20), the coverage of SLNBa by the TgF was ''suitable'' in eight cases (40 %), ''partially suitable'' in six cases (30 %), and ''unsuitable'' in six cases (30 %). In the HTgF group, four and one patients were considered ''suitable'' or ''partially suitable,'' respectively. Finally, the SLNBa was completely covered by the TgF in 12 of 25 patients (48 %), independent of the TgF size. In the two patients with involved SLNs, STgF were modified as HTgF to include totally the SLNBa.

DISCUSSION
Our group showed recently that STgF planed for breast RT does not allow adequate coverage of the axilla. 9 These findings are important to consider in the context of the international guidelines on ALND avoidance when systemic therapy and RT are expected to cure potential residual disease in the axilla. [1][2][3][4][5] However, in the past decade, several studies have shown that STgF fails to adequately treat levels I-II. 1,[9][10][11][12][13][14][15] In this context, two major points on optimal regional RT technique to cover the axilla correctly should be considered. First, recent data from randomized trials and a meta-analysis have shown that nodal RT increases overall survival and particularly distant metastases free-survival. [16][17][18] Second, RT to the axilla has been shown as safe and equivalent to ALND in patients with MIC in SLNB. 19 These results highlight the importance of redefining adjuvant nodal RT in the SLNB era.
There is a direct relationship between prognosis and the number of involved LNs. In patients with positive SLN, the percentage of the SLN occupied by tumors and the number of SLNs removed are independently predictive of non-SLN involvement. In addition, the non-SLN involvement negatively influences survival. 20 Reed et al. reported that none of 13 patients with ITCs who underwent an ALND had additional positive nodes compared with 27 % of patients with MIC. At 5 years, distant recurrence rates in SLNnegative, isolated tumor cells, MIC, and MAC groups were 6, 8, 14, and 21 %, respectively. The presence of MIC in the SLN was associated with a significantly shorter diseasefree interval than was SLN negativity (p \ 0.02). 21 The RT objectives in case of axilla residual disease are to reduce the risk of locoregional recurrence and to prevent distant metastases from axillary sanctuary as hypothesized by Hellman: ''RT is stopping metastases at their source.'' 22 These two objectives are highly linked together and to the axilla contents dose. Francissen et al. reviewed 16 studies describing patients with MAC disease in the SLN without ALND completion. 23 After 43 months of follow-up, they observed only 24 axillary recurrences out of 3,268 (0.7 %) among whom three received RT. 23 The axillary recurrence rate is even lower in the database study by Yi et al. with only 0.1 % among the 1, 473 patients with MAC in the SLN. 24 Other smaller cohorts studies showed rates between 0 and 7.1 %. 1 However, there is a lack of RT technique data to conclude on the relationship between axilla underdosage and local recurrence. 2,5 In our study, STgF coverage of SNLBa was complete (''suitable'' group) in only eight cases (40 %), whereas HTgF covered the SLNBa in four of five (80 %) patients. Indeed, the significant variations of the anatomical location of the SLNs do not allow coverage of the SLNBa with STgF, which does not include the LNs at highest risk of containing tumor. Therefore, some authors suggested removing the superiorposterior corner multileaf collimators of the TgF to cover axilla levels. 13,25,26 In our study, STgF have been modified as HTgF to include the whole SLNBa in both patients with metastases in the SLN.
Tumoricidal radiation dose also should be questioned. The evaluation of the delivered doses to axilla contents in the Z0011 study remains uncertain. 5 The recent report from Jagsi et al. based on a centralized review of 228 (out of 605) patients provided only partial results on RT technique. Among the 185 (out of 228) patients with TgF-only treatment there was sufficient data to evaluate TgF height in 142 (76.8 %) patients. Because RT parameters and nodal volumes details were lacking, they could not evaluate the dose distribution in the axilla. However, they showed that direct nodal irradiation technique was mainly used in case of multiple nodal involvements for better coverage of the axilla compared with the HTgF technique. 6 Several reports have highlighted that axillary nodal coverage depends on the upper TgF border. Studies using STgF showed that only approximately 50 % of level I and 20-30 % of level II nodes might receive 95 % of the prescribed dose. 10,[26][27][28] However, several of these early studies used conventional simulation with surgical clips as anatomic landmarks to evaluate the dose distribution. For example, Reed et al. showed that STgF fail to treat the axillary level I-II anatomic volume adequately, with approximately 50 % receiving a therapeutic dose. They concluded that surgical clips from ALND grossly underestimate the level I-II axilla nodal volume and should not be relied on for therapy planning. 29 The use of HTgF in patients with MIC in the SLN is an important issue to consider when no ALND is indicated. In the Z0011 trial, HTgF were used in 50 and 52.6 % of patients randomized to the ALND and SLNB arms, respectively. Of note, only 43 (19 %) patients received direct regional RT using C3 fields. In this small group receiving a third field, there was a trend suggesting that treatment with posterior axillary boost field was more common in patients who had SLNB alone (12/21 vs. 6/22; p = 0.0066) and those receiving nodal RT had greater number of LNs involved (p \ 0.001). 6 Axilla coverage may be paramount to locoregional control and to decreasing the risk of metastatic dissemination from residual uncontrolled disease in the axilla. This is particularly important to consider regarding the recent data on overall and metastatic free-survival benefits from nodal RT in large clinical trials. [16][17][18] Our study was undertaken to address the question of dose distribution and SLNBa coverage according to TgF size. We showed that the SLNBa was completely covered by the TgF independently from its size in only 48 % of the patients. The average dose in the SLNBa was 33 Gy.
However, there was significantly higher delivered dose using HTgF. The average dose is considered as nontumoricidal in at least the partially suitable (34 Gy) and unsuitable (8 Gy) groups (Table 4). To overcome the underdosage of axilla, some authors have suggested tailoring TgF to targets. Kiel et al. recommended that the cranial field edge should be 1.2 cm below the humeral head and that 2.5 cm of the lung be included in the breast TgF to adequately cover the axilla. 27 Schlembach et al. demonstrated that the LNs at greatest risk are 2 cm below the humeral head in 95 % of cases. 26 Reznik et al. reported an increase of 20-30 % of the average dose delivered to axilla levels when using HTgF. 10 Considering the same borders, these values were lower in our earlier study in which STgF was used in the majority of patients. 9 Many authors have attempted to define the anatomical borders by surgically marked axilla volumes and evaluated dose distribution at each level. Krasin et al. showed that only 1 of 25 patients received 50 Gy in the Level I of the axilla, and no patient had an adequate coverage of the Level II-III. 11 Reed et al. showed a significant volume difference between the anatomical and the surgically marked axillary volumes in 18 of 50 patients undergoing ALND with more adequate coverage of the axilla in the latter. 29 Another way to define marked axilla is to use a sentinel clip at the caudal border in the anatomically defined axilla. Using this procedure, Orecchia et al. showed that only 1 of 15 patients received 40 Gy in the axilla in a context of significant volume reduction. 15 In a context of the ALND avoidance in selected patients, the debate on the axilla underdosage by TgF has to be addressed as the risk of non-SLNs involvement may depend on the anatomic location of SLN and its degree of involvement. 30 The latter as determined by conventional histology has been described as a predictive factor for additional axillary involvement. 21 Thus, as studies that have been undertaken to quantify intraoperatively the total tumor load in the positive SLNs showed that it is possible to predict additional non-SLN metastasis in the axilla with a high specificity, the authors suggested that this could be used to guide decisions for ALND completion. 30 From the radiation oncology view, for an adequate coverage of the axilla, the use of direct fields could be considered rather than TgF in the patients with SLN involvement without further ALND. 19

CONCLUSIONS
In patients undergoing BCS followed by WBRT, STgF provide a limited coverage of the axilla contents and deliver a nontumoricidal dose to potential axillary residual disease. The RT technique to deliver adequate dose to axilla and SLNBa has to be standardized with the use of HTgF or direct fields as described in AMAROS trial. 19 This is true, insofar as we have: only limited follow-up in the SLNB trials without RT technique details, no clear tumoricidal dose level for residual disease, and uncertainties on the disease in the remained axilla non-SLN.