De ﬁ ning the dose ‐ volume criteria for laryngeal sparing in locally advanced oropharyngeal cancer utilizing split ‐ ﬁ eld IMRT, whole ‐ ﬁ eld IMRT and VMAT

Purpose: To determine the optimal dose ‐ volume constraint for laryngeal sparing using three commonly employed intensity modulated radiation therapy (IMRT) dose < 33 Gy. There were no signi ﬁ cant differences in dose to target structures or non ‐ laryngeal organs at risk among techniques.

dose < 33 Gy. There were no significant differences in dose to target structures or non-laryngeal organs at risk among techniques.

K E Y W O R D S
dosimetry, IMRT, Larynx, oropharynx, split-field, VMAT

| INTRODUCTION
Head and neck cancer mortality in the United States has steadily fallen over the past several decades in part because of the increasing incidence of human-papilloma virus (HPV) associated oropharyngeal squamous carcinoma (OPSCC). As more patients are cured of their disease, the focus of curative radiotherapy has shifted towards minimizing treatment-related toxicities. [1][2][3] Dysphagia is one of the more commonly cited long-term side effects associated with head and neck radiotherapy and has been shown to significantly decrease quality of life (QOL) metrics following treatment. 4,5 It is of considerable interest, therefore, to minimize dose to the larynx and laryngeal substructures while maintaining optimized dose delivery to the tumor and at risk sites.
Historically, radiotherapy for OPSCC was delivered with a threefield approach, utilizing a parallel-opposed beam arrangement to treat the primary tumor and lymph nodes above the thyroid cartilage. These beams were matched to a low anterior neck (LAN) field to address the draining lymphatics of the mid-low neck and a laryngeal block was included to shield the larynx. 6,7 Although intensity modulated radiation therapy (IMRT) has since emerged as the preferred approach for the treatment of head and neck malignancies owing particularly to improved parotid sparing, 8,9 the effectiveness by which whole-field IMRT (WF-IMRT), which encompasses the entire target volume and low neck in a single plan, can protect the larynx compared to a larynx block LAN field remains controversial.
The split-field IMRT (SF-IMRT) technique was employed as an effort to minimize excess dose to the larynx by matching the IMRT field above the glottis to a LAN field with larynx block. [10][11][12] However, potential drawbacks with SF-IMRT are longer treatment times, dose uncertainties at the match line and the technical challenges of incorporating mid-neck and appositional electron boost fields in patients with gross disease at or below the match line.
In our previous study, we compared SF-IMRT to whole-field-IMRT (WF-IMRT) and found the former to provide significantly better laryngeal sparing (mean laryngeal dose: 18.7 vs 47 Gy). 13 There have been a number of notable technological advances since our initial study was performed. Volumetric modulated arc therapy (VMAT) represents a new IMRT technology that offers improved treatment efficiency and dose conformity, but it can be difficult to match to LAN fields. 14,15 Improved diagnostic imaging accuracy in CT-based target and organ at risk (OAR) delineation together with imageguided radiation therapy (IGRT) have additionally allowed smaller treatment margins to be used. Whether these improvements can improve larynx sparing compared to SF-IMRT in the treatment of OPSCC remains unclear.
Interpretation of the dose-volume relation for laryngeal dysfunction and identifying the optimal dose-volume relationship for laryngeal dysfunction is further complicated by differences in the 3D larynx volume definition among clinical studies and in clinical practice. These differences among definitions principally involve the extent of the inclusion of structures of the supraglottic larynx. Here, we revisit larynx sparing using modern IMRT techniques in a larger and more homogeneous cohort of only oropharyngeal cancers. In particular, because of the growing acceptance of VMAT leading to a greater use of whole field techniques, we wanted to reinvestigate if this technique was compromising larynx sparing as we hypothesized over a decade ago. To further facilitate standardization of the laryngeal dose, we evaluated two common definitions of laryngeal volume in patients with locally advanced OPSCC.

2.A | Patients
We evaluated patients with OPSCC treated with definitive radiotherapy using SF-IMRT at MD Anderson Cancer Center (MDACC) from 2008 to 2014. Those with the presence of bilateral or bulky nodal disease (≥6 cm), primary tumors extending to the larynx, or the absence of complete dosimetric planning data were excluded. In total, 70 patients were identified of which 30 patients were treated to the bilateral neck and included in this analysis. The study was approved by the MDACC institutional review board.

2.B | Delineation of laryngeal and OARs volumes
The larynx (MDACC larynx) was defined per our prior publication as bounded by the superior and inferior aspects of the thyroid cartilage. 13 We defined a second larynx volume based on the Radiation Therapy Oncology Group 1016 guidelines (RTOG larynx). This volume extended from the inferior aspect of the hyoid bone to the cricoid cartilage with inclusion of the infrahyoid but not suprahyoid epiglottis. We defined the supraglottic and subglottic larynx as the volumes of the RTOG larynx extending superiorly and inferiorly, respectively, beyond the borders of the MDACC larynx. A diagram depicting the superior and inferior borders of these structures is shown in Fig. 1

2.C | Radiotherapy planning procedures
Contours underwent institutional quality assurance peer-review by radiation oncologists specializing in the treatment of head and neck malignancies. 16 The radiation treatment plans were generated using

2.D | Statistical analysis
SF-IMRT, WF-IMRT, and VMAT plans were compared with regards to target coverage, OAR dose, total delivery time, monitor units required to deliver each plan, as well as the Paddick conformity index and the heterogeneity index (percentage of one standard deviation to the mean dose) of the primary targets. The Tukey-Kramer method was used to identify significant differences between treatment methods with P < 0.05 considered statistically significant. All analyses were performed using the JMP Pro 12 software package (SAS, Cary, NC).

3.B | Larynx and laryngeal substructure dose
Laryngeal/critical OAR doses and dosimetric indices are shown in

3.D | Treatment time and dosimetric indices
The VMAT plans had a significantly shorter treatment time

| DISCUSSION
Dysphagia is a significant treatment-associated toxicity following head and neck radiotherapy. Mean laryngeal doses exceeding 40 Gy are associated with risk of aspiration and prolonged feeding tube dependence. 17,18 Despite the need to minimize unnecessary larynx dose, published data evaluating the optimal method for laryngeal sparing in OPSCC patients receiving IMRT have yielded mixed results.
As early adopters of IMRT for head and neck cancer treatment, we previously demonstrated a considerably lower mean MDACC larynx dose with SF-IMRT (18.7 Gy), compared to WF-IMRT (47 Gy) in 13 patients with early stage OPSCC (T1-2, N0-1). 13 These results  Several factors may explain the lower larynx dose in this study compared to our prior observations. In addition to a larger sample size and improvements in CT-based contouring and IMRT treatment planning, this study used a non-uniform PTV for whole-field techniques such that PTV expansion into the larynx was restricted. In  were not used in treatment of patients and thus the final plans were "accepted" for this study when the planning objectives were achieved without further optimization beyond this.
In conclusion, modern IMRT techniques including SF-IMRT, WF-IMRT, and VMAT can provide excellent laryngeal sparing for patients with primary oropharynx cancer not involving the larynx who receive bilateral neck radiotherapy. There were no major clinical dosimetric differences with the exception of supraglottic larynx mean doses among IMRT techniques. VMAT was the most efficient treatment method requiring significantly fewer monitor units and considerably shorter treatment duration and thus lower risk of intrafraction movement. Achieving a larynx mean dose <20 Gy was feasible in a majority of cases and in nearly all cases, a mean larynx dose of <30 Gy was accomplished using all three techniques. VMAT should be con-