NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1 Radiation Therapy for Oropharyngeal Squamous Cell Carcinoma: An ASTRO Evidence

2 Based Clinical Practice Guideline

3 4 David J. Sher, MD, MPH1, David J. Adelstein MD, FACP2, Gopal Bajaj, MD3, David M. Brizel,

5 MD4 , Ezra E. Cohen, MD5, Aditya Halthore, MD6, Louis B. Harrison, MD7, Charles Lu, MD8,

6 Benjamin J. Moeller, MD9, Harry Quon, MD, PhD10, James W. Rocco, MD, PhD11, Erich M.

7 Sturgis, MD, MPH12, Roy B. Tishler, MD, PhD13, Andy Trotti, MD14, John Waldron, MD15,

8 Avraham Eisbruch, MD16

9 10 1. Department of Radiation Oncology, UT Southwestern, Dallas , Texas 11 2. Department of Hematology and Medical Oncology, Cleveland Clinic Taussig Cancer 12 Institute, Cleveland, Ohio 13 3. Department of Radiation Oncology, Inova Dwight and Martha Schar Cancer Institute, 14 Falls Church, VA 15 4. Department of Radiation Oncology, Duke University Medical Center, Durham, North 16 Carolina 17 5. Department of Radiation Oncology, University of California, San Diego, California 18 6. Department of Radiation Medicine, Northwell Health, Lake Success, New York. 19 7. Department of Radiation Oncology, Moffitt Cancer Center, Tampa, Florida 20 8. Department of Radiation Oncology, UT MD Anderson Cancer Center, Houston, 21 Texas 22 9. Department of Radiation Oncology, North Shore-LIJ Health System, Bellerose, New 23 York 24 10. Department of Radiation Oncology, John Hopkins, Baltimore, Maryland 25 11. Department of Otolaryngology-Head and Neck Surgery, The Ohio State University 26 Wexner Medical Center, Columbus, Ohio 27 12. Department of Head and Neck Surgery and Department of Epidemiology, The 28 University of Texas MD Anderson Cancer Center, Houston, Texas 29 13. Department of Radiation Oncology, Brigham and Women's Physician Organization, 30 Boston, Massachusetts 31 14. Department of Radiation Oncology, Moffitt Cancer Center, Tampa, Florida 32 15. Department of Radiation Oncology, Princess Margaret Cancer Center, Ontario, 33 Canada 34 16. Department of Radiation Oncology, University of Michigan, Ann Arbor, Michigan 35 36 37 38 39 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

40 Conflict of Interest Disclosure Statement

41 The guideline panelists were required to complete disclosure statements before initiating

42 work on this project. These statements are maintained at the American Society for Radiation

43 Oncology (ASTRO) Headquarters in Fairfax, VA, and pertinent disclosures are published within

44 this report. The ASTRO Conflict of Interest Disclosure Statement seeks to provide a broad

45 disclosure of outside interests. The guideline panel chairs (AE and DS), in concert with the

46 ASTRO COI review committee, ASTRO legal counsel, and ASTRO guidelines subcommittee

47 chair and vice-chair, reviewed these disclosures and approved the participation of all task force

48 members for all key questions.

49

50 51 Acknowledgements

52 The authors thank the following individuals who served as expert reviewers of the

53 manuscript: Steven Frank, MD, Nancy Lee, MD, and Cristina Rodriguez, MD, none of whom

54 have any relevant conflicts of interest. Gratitude is extended to Warren Caldwell, for serving as a

55 patient advocate for this guideline. The authors also thank ASTRO staff members Margaret

56 Amankwa-Sakyi, Sokny Lim and Caroline Patton for the systematic literature review assistance

57 and administrative support.

58 The Oropharyngeal Squamous Cell Carcinoma Guideline panel, created by the guidelines

59 subcommittee of the Clinical Affairs and Quality Committee of ASTRO, prepared this

60 document. ASTRO guidelines present scientific, health, and safety information and may to some

61 extent reflect scientific or medical opinion. They are made available to ASTRO members and to

62 the public for educational and informational purposes only. Commercial use of any content in

63 this guideline without the prior written consent of ASTRO is strictly prohibited. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

64 Adherence to this guideline will not ensure successful treatment in every situation.

65 Furthermore, this guideline should not be deemed inclusive of all proper methods of care or

66 exclusive of other methods of care reasonably directed to obtaining the same results. The

67 physician must make the ultimate judgment regarding the propriety of any specific therapy in

68 light of all the circumstances presented by the individual patient. ASTRO assumes no liability

69 for the information, conclusions, and findings contained in its guidelines. In addition, this

70 guideline cannot be assumed to apply to the use of these interventions performed in the context

71 of clinical trials, given that clinical studies are designed to evaluate or validate innovative

72 approaches in a disease for which improved treatments are needed or are being explored.

73 The guideline panel recommends that providers discuss with patients’ shortly after

74 diagnosis what to expect regarding symptoms, treatment-related toxicities, outcomes including

75 risk of recurrence, and effects of the disease and the treatments on quality of life.

76 This guideline was prepared on the basis of information available at the time the panel

77 was conducting its research and discussions on this topic. There may be new developments that

78 are not reflected in this guideline and that may, over time, be a basis for ASTRO to consider

79 revisiting and updating the guideline.

80 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

81 Abstract

82 Purpose: To present evidence-based guidelines for the treatment of oropharyngeal squamous

83 cell carcinoma (OPSCC) with definitive or adjuvant radiotherapy (RT).

84 Methods and Materials: The American Society for Radiation Oncology (ASTRO) convened

85 the OPSCC Guideline Panel to perform a systematic literature review investigating the following

86 key questions: (1) When is it appropriate to add systemic therapy to definitive RT in the

87 treatment of OPSCC? (2) When is it appropriate to deliver post-operative RT with and without

88 systemic therapy following primary surgery of OPSCC? (3) When is it appropriate to use

89 induction chemotherapy in the treatment of OPSCC? (4) What are the appropriate dose,

90 fractionation, and volume regimens with and without systemic therapy in the treatment of

91 OPSCC? The GRADE methodology was followed in the construction and assessment of the

92 recommendation statements, and a predefined consensus-building process was implemented to

93 arrive at the final statements.

94 Results: Patients with stage IV and stage T3 N0-1 OPSCC treated with definitive radiotherapy

95 should receive concurrent high-dose intermittent cisplatin (HDIC). Patients receiving adjuvant

96 radiotherapy following surgical resection for positive surgical margins or extracapsular extension

97 should be treated with concurrent HDIC, and individuals with these risk factors who are

98 intolerant of cisplatin should not routinely receive adjuvant concurrent systemic therapy.

99 Induction chemotherapy (IC) should not be routinely delivered to patients with OPSCC. For

100 patients with stage IV and stage T3 N0-1 OPSCC ineligible for concurrent chemoradiotherapy,

101 altered fractionation RT should be used. Patients with tonsillar fossa-only, T1-2 N0-1 OPSCC

102 should receive ipsilateral radiotherapy. Recommendation statements do not vary by human

103 papillomavirus (HPV) status. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

104

105 Conclusion: The successful management of OPSCC requires the coordination and collaboration

106 of radiation, medical and surgical oncologists. When high-level data are absent for clinical

107 decision-making, treatment recommendations should incorporate patient values and preferences

108 to arrive at the optimal therapeutic approach.

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

127 Introduction

128 The oropharynx is an anatomic site within the head and neck that includes the tonsils,

129 base of tongue, soft palate, and the upper lateral and posterior pharyngeal walls. It is lined by

130 squamous epithelium, and thus squamous cell carcinoma composes the vast majority of

131 malignancies that arise from the oropharynx. The epidemiology and prognosis of oropharyngeal

132 squamous cell carcinoma (OPSCC) has changed dramatically over the past 30 years, such that its

133 treatment and expected outcomes are nearly unrecognizable from a generation ago.

134 Indeed, human papillomavirus (HPV) associated cancers of the oropharynx are now

135 widely recognized as a distinct clinical disease with regard to risk factor profiles, clinical and

136 molecular characteristics, treatment response and prognosis.1 Human papillomavirus HPV-

137 associated cancers arise from the lingual and palatine tonsils within the oropharynx, are poorly

138 differentiated and present with early T stage and advanced N stage. When compared to patients

139 with HPV-negative disease, patients with HPV-positive tumors are more frequently male, young,

140 and possess a favorable performance status. 3 The incidence of HPV-positive OPSCC has

141 increased by over 200% between 1988 and 2004, whereas the incidence of HPV-negative

142 OPSCC declined by half over that same era.2 Individuals with HPV-driven oropharynx cancers

143 have an estimated 50% reduction in risk of death when compared to patients with HPV-negative

144 tumors.3 The recognition of this powerful prognostic and predictive influence of HPV has

145 generated a wide array of treatment paradigms for OPSCC, which further highlights the

146 importance of evidence-based medicine to guide clinical practice.

147 The purpose of this clinical practice guideline is to systematically review the evidence for

148 effective treatment of OPSCC with definitive or adjuvant radiotherapy. Recommendation

149 statements are independent of HPV status, given the current absence of high-level, high-quality NOT TO BE COPIED, DISSEMINATED OR REFERENCED

150 data confirming treatment decisions should differ by cancer etiology. The panel used a

151 formalized literature review process, complemented by expert opinion where appropriate, to

152 make recommendations on the use of concurrent systemic therapy in the primary and post-

153 operative settings, adjuvant radiotherapy following curative surgery, neoadjuvant chemotherapy

154 prior to radiotherapy, and dose, fractionation and neck volume decisions faced in the

155 management of this disease.

156

157 Methods and Materials

158 Process

159 The Guidelines Subcommittee of the Clinical Affairs and Quality Council identified

160 OPSCC as a disease that provides many challenges to the radiation oncologist, as well as to the

161 medical and surgical oncology multidisciplinary team. Given its rapidly increasing prevalence

162 and the various treatment options with differing short- and long-term toxicity profiles, OPSCC

163 was considered a high-priority topic in need of an evidence-based practice guideline. In

164 accordance with established ASTRO policy, the Guidelines Subcommittee recruited a guideline

165 panel of recognized experts in oropharyngeal cancer, including radiation oncologists, medical

166 oncologists, otolaryngologists, and a patient advocate. Panel members were drawn from

167 academic settings, private practice, and residency. Four specific key questions (KQs) were

168 proposed, which addressed: (KQ1) the addition of concurrent systemic therapy to radiotherapy,

169 (KQ2) the delivery of post-operative radiotherapy with and without systemic therapy following

170 primary surgery, (KQ3) the use of induction chemotherapy, and (KQ4) the optimal dose-

171 fractionation regimens with and without systemic therapy, as well as nodal volumes for primary NOT TO BE COPIED, DISSEMINATED OR REFERENCED

172 tonsillar cancer. In September 2014, the ASTRO Board of Directors approved the proposal and

173 panel membership.

174 Through a series of communications by conference calls and emails between December

175 2014 and February 2016, the guideline panel, with ASTRO staff support, completed the

176 systematic review, created literature tables, and formulated the recommendation statements and

177 narratives for the guideline. The members of the task force were divided into four writing groups

178 by KQs, according to their areas of expertise. The initial draft of the manuscript was reviewed by

179 three expert reviewers (see Acknowledgements) and ASTRO legal counsel. A revised draft was

180 placed on the ASTRO Web site in MONTH 2016 for a four/six-week period of public comment.

181 Following integration of the feedback, the document was submitted for approval to the ASTRO

182 Board of Directors in MONTH 2016. The ASTRO Guidelines Subcommittee intends to monitor

183 this guideline and initiate an update when appropriate, according to existing ASTRO policies.

184

185 Literature Review 186 187 A systematic review of the literature was performed in early 2015 to form the basis of the

188 guideline. An analytic framework incorporating the population, interventions, comparators, and

189 outcomes (PICO) was first used to develop and refine search strategies for each KQ. The

190 searches were conducted in MEDLINE PubMed and designed to identify studies published in

191 English between January 1990 and December 2014 that evaluated adults with OPSCC who were

192 treated with primary radiotherapy, adjuvant radiotherapy, or radiotherapy with concurrent

193 systemic therapy. Both MeSH terms and text words were utilized and terms common to all

194 searches included: oropharyngeal neoplasms, carcinoma squamous cell, and radiotherapy.

195 Additional terms specific to each KQ were also incorporated. The outcomes of interest were NOT TO BE COPIED, DISSEMINATED OR REFERENCED

196 overall and progression-free survival, recurrence rates, toxicity, and quality of life. The initial

197 literature review was conducted in December 2014. The electronic searches were supplemented

198 by hand searches.

199 A total of 2615 abstracts were retrieved. The articles were then reviewed by ASTRO

200 staff, the co-chairs of the guideline, and the writing groups for each KQ. During the first round of

201 screening, 2452 articles were eliminated based on the inclusion and exclusion criteria. The

202 inclusion criteria were: patients ≥18 years or older, all stages of oropharyngeal cancer, and

203 publication date 1990 to 2014. Included treatments were: primary radiotherapy, primary

204 chemoradiation, primary surgery with adjuvant radiotherapy with or without systemic therapy, or

205 IC followed by radiation therapy or chemoradiotherapy. The exclusion criteria were: pre-clinical

206 or non-human studies, case reports/series, non-English language, available in abstract only,

207 pediatric patients, distant metastasis, non-squamous cell carcinoma, and otherwise not clinically

208 relevant to the key clinical questions. The panelists on KQ 1, 3 and 4 generally only considered

209 articles in which the percentage of OPSCC patients was greater than 50%; the panelists on KQ 2

210 did not have an absolute threshold because OPSCC patients typically comprise a minority of

211 individuals in post-operative studies. Ultimately, 119 full-text articles were chosen for inclusion

212 and abstracted into detailed literature tables to provide supporting evidence for the clinical

213 guideline recommendations.

214 Because expression of the p16 protein is typically used as a surrogate for HPV infection,

215 4 the two words – p16-positive and HPV-positive – are sometimes used interchangeably (and

216 incorrectly). However, throughout this guideline, the mention of p16 will be reserved for clinical

217 studies in which its status was assessed and reported.

218 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

219 Grading of Evidence, Recommendations, and Consensus Methodology

220 Guideline recommendation statements were generated from this literature review, with

221 high-quality evidence forming the basis of the statements whenever possible, in accordance with

222 Institute of Medicine (IOM) standards; 5 expert opinion supplemented the evidence base if high-

223 level data were not available. The GRADE (Grading of Recommendations, Assessment,

224 Development and Evaluations) methodology was followed in the construction and assessment of

225 the recommendation statements. GRADE is an explicit, systematic approach to defining the

226 recommendation strength and quality of evidence.6,7

227 Recommendation strengths were dichotomized as “strong” or “conditional.”

228 Strong recommendations were made when the panel was very confident that the benefits of the

229 intervention clearly outweighed the harms; conditional recommendations were made when the

230 ratio between risks and benefits was more balanced. Per the GRADE formalism, a “strong”

231 recommendation implies that the panelists believe “all or almost all informed people would make

232 the recommended choice for or against an intervention.” 7 A “conditional” recommendation

233 implies that “most informed people would choose the recommended course of action, but a

234 substantial number would not. When a recommendation is weak, clinicians and other health care

235 providers need to devote more time to the process of shared decision making by which they

236 ensure that the informed choice reflects individual values and preferences.”5 The importance of

237 patient preferences and shared decision-making was highlighted in each conditional

238 recommendation statement.

239 The GRADE methodology defines “quality of evidence” as a rating that indicates “the

240 extent of our confidence that the estimates of an effect are adequate to support a particular

241 decision or recommendation.”6 The quality of evidence underlying each recommendation NOT TO BE COPIED, DISSEMINATED OR REFERENCED

242 statement was categorized as either high, moderate, low. 6 Although GRADE does provide for a

243 “very low” quality, the panel did not consider this rating for consistency with prior ASTRO

244 guidelines. Descriptions of each quality level are below:

245  High: We are very confident that the true effect lies close to that of the estimate of

246 the effect.

247  Moderate: We are moderately confident in the effect estimate: The true effect is

248 likely to be close to the estimate of the effect, but there is a possibility that it is

249 substantially different

250  Low: Our confidence in the effect estimate is limited: The true effect may be

251 substantially different from the estimate of the effect

252 The degree of consensus among the panelists on each recommendation statement was

253 evaluated through a modified Delphi approach. A survey was sent by ASTRO staff to the panel

254 members, who rated their agreement with each recommendation on a five-point Likert scale,

255 ranging from strongly disagree to strongly agree (higher score corresponds with stronger

256 agreement). A pre-specified threshold of ≥ 75% of raters was determined to indicate when

257 consensus was achieved.8 If a recommendation statement did not meet this threshold, it was

258 modified and re-surveyed, or excluded from the guideline. The final set of recommendation

259 statements were achieved after 2 surveys. The guideline statements, along with the

260 corresponding ratings of recommendation strength, quality of evidence, and level of consensus,

261 are listed in Table 1.

262

263

264 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

265 Results 266 267 Key Question 1. When is it appropriate to add systemic therapy to definitive radiotherapy 268 in the treatment of oropharyngeal squamous cell carcinoma (OPSCC)? 269 270 1. In the scenario of stage IVA-B disease? 271 272 273  Statement 1A. Concurrent high-dose intermittent cisplatin should be delivered to patients with 274 stage IVA-B oropharyngeal squamous cell carcinoma receiving definitive radiotherapy. 275 o Recommendation strength: Strong 276 o Quality of evidence: High 277 278  Statement 1B. Concurrent cetuximab or carboplatin-fluorouracil should be delivered to patients 279 with stage IVA-B oropharyngeal squamous cell carcinoma receiving definitive radiotherapy who 280 are not medically fit for high-dose cisplatin. 281 o Recommendation strength: Strong 282 o Quality of evidence: High 283 284  Statement 1C. Concurrent weekly cisplatin may be delivered to patients with stage IVA-B 285 oropharyngeal squamous cell carcinoma receiving definitive radiotherapy who are not medically 286 fit for high-dose cisplatin, after a careful discussion of patient preferences and the limited 287 prospective data supporting this regimen. 288 o Recommendation strength: Conditional 289 o Quality of evidence: Low 290 291  Statement 1D. Concurrent cetuximab should not be delivered in combination with chemotherapy 292 to patients with stage IVA-B oropharyngeal squamous cell carcinoma receiving definitive 293 radiotherapy. 294 o Recommendation strength: Strong 295 o Quality of evidence: High 296 297  Statement 1E. Intra-arterial chemotherapy should not be delivered to patients with stage IVA-B 298 oropharyngeal squamous cell carcinoma receiving definitive radiotherapy. 299 o Recommendation strength: Strong 300 o Quality of evidence: High

301 302 303 Narrative 304 305 The role of concurrent cytotoxic chemotherapy combined with radiotherapy

306 Several phase III trials have randomized patients with locally advanced, non-metastatic

307 head and neck squamous cell carcinoma to treatment with conventionally- or altered-fractionated NOT TO BE COPIED, DISSEMINATED OR REFERENCED

308 radiotherapy alone versus treatment with concurrent chemoradiotherapy (Table 2). All of the

309 studies in this systematic review were predominantly composed of patients with oropharynx

310 cancer, and the clear majority of patients presented with stage IV disease. In the six studies that

311 directly compared standard fractionation radiotherapy to combined concurrent

312 chemoradiotherapy, dose and fractionation schemes were relatively uniform, with total delivered

313 doses approaching 66-70.2 Gy in 1.8-2 Gy fractions.9-15 The chemotherapy regimens were

314 variable, including bolus cisplatin9, weekly cisplatin, 25 carboplatin-fluorouracil, 15,16 daily

315 carboplatin,11 and bleomycin-mitomycin. 14

316 The five trials comparing altered fractionation radiotherapy (3 accelerated, 2

317 hyperfractionated) with and without concurrent chemotherapy also employed a variety of

318 systemic therapy regimens, including carboplatin-fluorouracil,17 daily cisplatin,18 cisplatin-

319 fluorouracil,19 high-intensity cisplatin-fluorouracil (bolus cisplatin every 2 weeks),20 and

320 fluorouracil-mitomycin.21,22 However, as opposed to the conventionally fractionated trials, in

321 which the radiotherapy was standardized between the two arms, two of these studies20,21 used

322 different radiotherapy schemes for the radiotherapy-alone arm. The ARO 95-06 trial12 treated

323 patients to a total dose of 77.6 Gy without chemotherapy and 70.6 Gy with chemotherapy, and

324 the GORTEC trial23 delivered 64 Gy in 5 weeks with chemotherapy and the same dose in 3

325 weeks without chemotherapy.

326

327 Impact on overall survival

328 In all but two of the 6 studies of conventionally fractionated radiation therapy,

329 chemoradiotherapy significantly improved overall survival.10,11,14,24,25 Most of these studies

330 reported the statistically significant improvement in OS at 3 years, with the absolute benefit of NOT TO BE COPIED, DISSEMINATED OR REFERENCED

331 chemoradiotherapy ranging between 14% and 24%.10,11,24 The few studies with longer follow-up

332 showed a persistent but smaller survival advantage to concurrent chemotherapy. For example,

333 the 5-year OS difference in GORTEC 94-0126 was 6% (22% vs. 16%, p=0.05), whereas in the

334 AIRO study11 using daily carboplatin, the 5-year survival difference was less than 3%. In

335 contrast, neither the Mumbai trial25 (weekly cisplatin) nor the ORO 93-0127 (carboplatin-

336 fluorouracil) showed an OS advantage with concurrent chemotherapy, although disease-free

337 survival was significantly improved with chemotherapy in both studies; it is notable that both of

338 these trials included a third arm of altered fractionation alone, and so the total numbers of

339 patients in each arm were less than 70, likely underpowering the studies for an OS endpoint.

340 Among the 5 studies of chemoradiotherapy with altered fractionation, three of them

341 showed an absolute OS advantage with the addition of concurrent chemotherapy, ranging from

342 5%20 (5-year ARO 95-06, fluorouracil and mitomycin) to 9.8%28 (5-year, Semrau et al,

343 carboplatin and fluorouracil) to 24%19 (3-year, Wendt et al, cisplatin and fluorouracil). Among

344 the other two trials, concurrent daily cisplatin improved cancer-specific survival in SAKK

345 10/9418 by an absolute 12% at 10 years, and concurrent carboplatin-fluorouracil with very

346 accelerated radiotherapy in Bourhis et al.23 reduced the risk of any cancer progression by nearly

347 25%, but toxic deaths precluded a survival benefit.

348 The data from the Meta-Analysis of Chemotherapy in Head and Neck Cancer (MACH-

349 NC Meta-analysis) 29 corroborated the conclusions drawn from these individual phase III

350 randomized trials. In the most recent update from this group, Blanchard et al.29 analyzed over

351 5800 patients with oropharynx cancer and showed that the concomitant addition of

352 chemotherapy to radiotherapy resulted in an 8.1% absolute benefit in OS at 5 years30, translating

353 into a relative 22% reduction in overall mortality. By contrast, adjuvant or neoadjuvant NOT TO BE COPIED, DISSEMINATED OR REFERENCED

354 chemotherapy administration was not associated with a survival benefit. There was no significant

355 heterogeneity in chemotherapy treatment effect by fractionation regimen, although only the use

356 of platinum-fluorouracil (HR 0.83, 95% CI 0.75-0.91) or platinum monotherapy (HR 0.70, 95%

357 CI 0.59-0.84) was significantly associated with survival. Additional univariable analyses

358 suggested that the survival benefit with concurrent chemotherapy was largely restricted to

359 younger patients (HR for 61 years or older, 0.97, 95% CI 0.87-1.08) and patients with stage IV

360 disease (HR for stage III disease, 1.01, 95% CI 0.88-1.14). However, multivariable analysis

361 showed that only favorable performance status and type of chemotherapy (i.e. platinum-based

362 chemotherapy with or without 5-FU) were significantly associated with overall survival.

363

364 Impact on locoregional control

365 The three trials of conventionally fractionated radiotherapy that reported locoregional

366 control (LRC) outcomes showed a statistically significant improvement in patients receiving

367 concurrent chemotherapy. Denis et al. and Ruo Redda et al.11,16 demonstrated an absolute 3-year

368 LRC benefit with concurrent chemotherapy of 24% and 8.7%, respectively 11,16, although

369 notably, the LRC difference in Ruo Redda et al.11 was only statistically significant among stage

370 IV patients. Ghosh-Laskar25 found that weekly cisplatin improved LRC by an absolute 17% at 5

371 years. The four studies of concurrent chemotherapy with altered fractionation that presented

372 distinct LRC probabilities consistently found that concomitant treatment significantly reduced

373 locoregional failure, with the absolute benefits of 8% (10-year, SAKK 10/94), 18,28 12% (5-year,

374 Semrau et al), 28 19% (3-year, Wendt et al), 19and 13%22 (3-year, 28 Budach et al).

375 In the recent MACH-NC meta-analysis 31 that did not distinguish by tumor site, Pignon

376 and colleagues showed that concurrent chemoradiotherapy significantly improved LRC, with an NOT TO BE COPIED, DISSEMINATED OR REFERENCED

377 absolute improvement of 9.3% and 13.5% among all studies and those with platinum-

378 fluorouracil, respectively. These absolute improvements translated into hazard reductions of 0.74

379 and 0.66 for all studies and those with platinum-fluorouracil, respectively (p<0.0001).

380

381 Impact on distant metastases

382 One of the challenges in interpreting the impact of chemotherapy on distant metastasis is

383 the inability of trials to distinguish between metastasis arising from a first locoregional failure

384 versus metastasis as the first site of failure. That said, none of the randomized studies of

385 concurrent chemotherapy in conventionally fractionated radiotherapy found an improvement in

386 distant control with systemic therapy, and of the five trials of altered fractionation, only the

387 Swiss Group for Clinical Cancer Research (SAKK) trial18 of daily cisplatin showed an

388 improvement in distant metastasis-free survival. In this study, concurrent chemotherapy

389 improved 10-year distant metastasis free survival by 15% at 10 years, but only 9 total patients in

390 the entire study developed an isolated metastasis without prior locoregional failure; thus, one

391 may assume that this presumptive improvement in distant control was a function of reduced

392 locoregional recurrence.

393 In contrast, the MACH-NC meta-analysis31, which benefitted from the increased power

394 of thousands of patients, showed a small but statistically significant improvement in distant

395 control with concurrent treatment; the absolute improvement was 2.5% for all patients, with a

396 relative improvement of 0.88 (95% CI 0.77-1.0, p=0.04). As stated above, this reduction in

397 distant metastasis with concurrent systemic treatment may reflect less propagation from fewer

398 locoregional recurrences.

399 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

400 Complications

401 While most of these phase III randomized studies show that adding chemotherapy to

402 radiotherapy improves survival, the combination clearly increased the risk of severe, acute

403 toxicities. For example, Adelstein et al.24 reported that patients receiving standard fractionation

404 RT with or without concurrent cisplatin had an overall grade 3 to 5 incidence of 85% and 51%,

405 respectively, of which leukopenia, anemia, and renal toxicities were the most significant. In this

406 Intergroup trial, 9 approximately 15% of patients treated with bolus cisplatin did not complete

407 treatment, and delivering three cycles of bolus cisplatin has been consistently challenging. For

408 example, in RTOG 0129, 32 which compared accelerated versus conventional radiotherapy with

409 concurrent bolus cisplatin in both arms, only 69% of patients in the conventional fractionation

410 arm received all 3 cycles, and approximately 24% of patients received only two courses of bolus

411 cisplatin.

412 In a review of the original GORTEC studies, Bourhis et al. showed that 71% of patients

413 receiving radiation with carboplatin and 5-FU developed acute grade 3 or 4 mucositis, compared

414 to only 39% of those receiving RT alone.33 Even weekly cisplatin in the Mumbai trial increased

415 the risk of grade 3 or 4 mucositis (35% vs. 21%, p value not reported), and daily carboplatin also

416 increased the risk of grade 2 toxicities (type not recorded). The combination of mitomycin-C and

417 bleomycin increased the risk of grade 3 or 4 mucositis by almost one-third.

418 The studies of concurrent chemotherapy with altered fractionation also found increases in

419 acute toxicity. The Starr et al. study showed statistically significant worse acute toxicities in the

420 arm receiving carboplatin and 5-FU—grade 3 and 4 mucositis was 68% versus 52% and grade 3

421 and 4 vomiting was 8.2% versus 1.6%.17 Similarly, the use of concurrent cisplatin-fluorouracil

422 with split-course hyperfractionation more than doubled the risk of acute grade 3-4 mucositis NOT TO BE COPIED, DISSEMINATED OR REFERENCED

423 (38% vs. 16%). In the GORTEC accelerated radiotherapy trial, 20 the very accelerated RT-alone

424 arm led to more acute mucosal toxicity, but a high early death rate from non-cancer cause

425 (almost 20% at 1 year) revealed the potentially life-threatening complications of intensive

426 chemotherapy with extreme acceleration. On the other hand, the higher radiation dose in ARO

427 95-06 21 actually showed more mucosal and skin toxicity in the radiotherapy alone cohort, and

428 the SAKK study18 using daily cisplatin found fewer than 10% of chemotherapy patients

429 developed a grade 3-4 hematologic toxicity, with no difference in acute radiotherapy effects.

430 Despite these acute toxicities, these trials consistently showed no significant increase in

431 high-grade late morbidity with concurrent therapy. A careful toxicity analysis of patients treated

432 on GORTEC 94-0126 showed that only grade 3-4 dental toxicity was significantly worse with

433 chemotherapy (44% vs. 12%). The ORO 93-0127 study showed numerically more tissue, skin and

434 mucosal side effects with chemotherapy, but these were not statistically tested and reported to be

435 typically transient. Neither daily carboplatin, weekly cisplatin nor mitomycin-bleomycin

436 increased the risk of late toxicity. Moreover, none of the five trials involving altered fractionation

437 with chemotherapy showed a significant increase in serious late toxicities.

438

439 Comparisons between chemotherapy regimens

440 While the studies mentioned above consistently showed LRC and typically OS benefits

441 from concurrent chemotherapy, it is important to note that the choice of systemic agents varied

442 considerably among each trial—bolus cisplatin, carboplatin/5-fluorouracil, cisplatin/5-

443 fluorouracil, low-dose carboplatin alone, and bleomycin/mitomycin C, were each selected and

444 administered at varying schedules and frequencies. Unfortunately, there are far fewer studies

445 comparing different chemotherapy regimens and schedules to each other. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

446 RTOG 970334 randomized over 241 patients to conventionally fractionated radiotherapy

447 plus either concurrent cisplatin-fluorouracil in the last 2 weeks of radiation therapy, daily

448 hydroxyurea-fluorouracil, or weekly cisplatin-paclitaxel; patients receiving daily chemotherapy

449 were treated every other week. Interestingly, all of the arms compared favorably to historical

450 controls, although the arms were not directly compared to each other. However, none of these

451 alternative regimens have been prospectively tested against more standard agents such as bolus

452 cisplatin.

453 The RTOG did not publish an additional randomized trial on competing chemotherapy

454 agents in head and neck cancer until RTOG 0522, which compared bolus cisplatin with bolus

455 cisplatin plus cetuximab, concurrent with accelerated radiotherapy.35 Nearly 900 patients were

456 enrolled in this trial, which confirmed that the addition of cetuximab did not improve outcomes.

457 Locoregional control, disease-free survival, and OS were not significantly different between the

458 two arms, and acute grade 3 and 4 toxicities were uniformly higher in the cetuximab arm, with a

459 low but statistically significant increase in toxic death. Thus, cetuximab should not be combined

460 with cytotoxic chemotherapy unless on a clinical trial.

461 Erlotinib is an oral biologic agent that targets EGFR by tyrosine kinase inhibition. It has

462 been used successfully in treating certain types of non-small cell lung cancer and, like

463 cetuximab, has demonstrated synergism when combined with chemotherapy or radiation in the

464 pre-clinical setting. Martins et al.36 performed a phase II randomized trial of cisplatin and

465 radiotherapy with or without concurrent erlotinib for stage III and IV locally advanced head and

466 neck cancers, most of which were oropharynx cancer. In results that echoed RTOG 0522,35 no

467 improvement was seen in complete response rates, LRC, progression-free survival, or OS at a NOT TO BE COPIED, DISSEMINATED OR REFERENCED

468 median follow up of 26 months, showing that an alternative method of EGFR inhibition is also

469 ineffective in combination with concurrent cisplatin.

470 The RTOG also obliquely asked a chemotherapy question in RTOG 01294,37, whose

471 primary hypothesis involved accelerated fractionation. This study randomized 743 patients to

472 either 3 cycles of bolus cisplatin with conventionally fractionated radiation therapy (RT) or 2

473 cycles of bolus cisplatin with accelerated RT.4,37 The trial showed no difference in any endpoint

474 between the arms, including distant metastasis, again confirming that the primary role of

475 chemotherapy is radiosensitization. In a subset analysis, the authors showed that patients

476 receiving only 1 cycle of cisplatin experienced markedly inferior survival. However, among

477 patients treated with conventionally fractionated RT, there was no statistically significant

478 survival decrement among the 24% of individuals who only received 2 cycles (HR 1.15, 95% CI

479 0.81-1.62). Although this finding casts some doubt on the benefit of the third cycle in the context

480 of standard fractionation, it is notable that (a) the upper end of the confidence interval was 1.62,

481 (b) the survival curve for patients receiving 2 cycles appears to separate from the curve for the

482 cohort receiving 3 cycles, and (c) this analysis was only based on 86 patients.

483 Finally, there is only one phase III trial comparing different modes of cisplatin

484 administration. Historically, there has been interest in delivering cisplatin intra-arterially, to

485 maximize the cisplatin dose to the tumor while infusing sodium thiosulphate systemically to

486 minimize cisplatin toxicity.38 The RTOG performed a multi-institutional phase II trial39 based on

487 this approach that produced results sufficiently promising for further study, and in fact, a

488 randomized controlled trial of this technique has been published. In this Dutch study by Rasch et

489 al.,40 239 patients with stage IV head and neck SCC received once-daily radiotherapy to 70 Gy

490 and were randomized to intravenous (three cycles of bolus infusion) versus intra-arterial (4 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

491 courses) cisplatin. 40 At 3 years, there was no difference in OS (51% vs. 47%), distant metastasis-

492 free survival (69% vs. 66%), and LRC (65% vs 63%). High-grade (i.e. > grade 2) renal toxicity

493 was more common in patients receiving intravenous cisplatin (9% vs. 1%), although only one

494 patient experienced permanent nephrotoxicity, and this patient was not dialysis-dependent. On

495 the other hand, neurological toxicity (> grade 2) was significantly more common in the intra-

496 arterial cohort (8 patients vs. 1) Given the technical and logistical challenges of administration

497 and the absence of any oncologic benefit but an increase in neurological toxicity, intra-arterial

498 chemotherapy should not be administered unless on a clinical protocol.

499

500 Summary of cytotoxic chemotherapy

501 The data are consistent that concurrent chemotherapy in combination with radiation

502 therapy improves LRC for patients with locally advanced oropharyngeal squamous cell

503 carcinoma, treated with either conventional or altered fractionation. In most of these studies, the

504 reduction in locoregional progression translated into a significant and meaningful OS benefit.

505 Although chemotherapy significantly increased acute toxicities, late effects were comparable to

506 treatment with radiotherapy alone, such that the survival benefits of concurrent therapy clearly

507 outweigh the non-trivial but short-term risks. Since the vast majority of the patients in these trials

508 presented with stage IV disease, concurrent systemic therapy should be delivered in this

509 population of patients.

510

511 The role of cetuximab combined with radiotherapy

512 Cetuximab is an IgG1 monoclonal antibody against epidermal growth factor receptor

513 (EGFR), which is a growth factor receptor overexpressed in head and neck cancer and associated NOT TO BE COPIED, DISSEMINATED OR REFERENCED

514 with poor prognosis. Pre-clinical data suggested synergy between cetuximab and radiotherapy,

515 and a preliminary phase I trial produced encouraging tolerability and response rates. These

516 promising outcomes generated a phase III randomized trial comparing radiotherapy alone (using

517 standard or altered fractionation) with radiotherapy plus cetuximab. The results of this trial were

518 first published in 2006 with a median follow-up of 54 months, revealing a clinically and

519 statistically significant improvement in OS with the addition of the antibody (hazard ratio 0.74,

520 95% CI 0.57-0.97), with an absolute 3 year survival difference of 10% (p=0.05). Similar to

521 cytotoxic chemotherapy, the benefit from combined modality therapy was in LRC, with a

522 relative 32% reduction in locoregional failure and absolute difference of 13% at 3 years

523 (p=0.01). There was no difference in distant metastases. There was no significant difference in

524 any acute adverse event except acneiform rash and infusion-related events, which were

525 substantially more common with cetuximab (17% vs. 1% for grade 3 or greater rash, p<0.001;

526 3% vs 0% for grade 3 or greater infusion reaction, p=0.01). The risk of mucositis (any grade)

527 was the same between the arms.

528 The results were updated in 2010 with a median follow-up of 60 months and additional

529 survival data on 40% of the surviving patients. The OS advantage was maintained (HR 0.73,

530 95% CI 0.56-0.95) was an absolute 5-year survival difference of 9%. Patterns-of-failure

531 information was not reported. Patients who developed a prominent rash experienced markedly

532 superior survival in comparison to those who did not (HR 0.49, median survivals 68.8 months vs.

533 25.6 months, p=0.002). A series of subset analyses suggested that the following cohorts

534 experienced significant benefit from cetuximab: oropharynx site, American Joint Committee on

535 Cancer (AJCC) T1-3, USA site, altered fractionation treatment, node-positivity, superior

536 performance status, male gender, and younger age (< 65 years). Of the 110 patients 65 years or NOT TO BE COPIED, DISSEMINATED OR REFERENCED

537 older, there was a non-significant trend favoring radiation therapy alone. It is critical to note,

538 though, that even though patients were stratified by Karnofsky Performance Status (KPS) (90-

539 100 vs lower), nodal status (N0 versus N+), tumor stage (T1-3 vs. T4), and radiation

540 fractionation regimen, the numbers in the subgroups were quite small, and so these results should

541 be considered hypothesis-generating.

542 A phase II randomized prospective trial sought to improve upon the results of Bonner et

543 al. by comparing concurrent cetuximab-radiotherapy with the same regimen plus 12 weeks of

544 adjuvant cetuximab.41 The authors of this study hypothesized that continued epidermal growth

545 factor receptor (EGFR) blockade by maintenance cetuximab would serve to eradicate any

546 residual disease after concurrent treatment. Although there was a trend for improved response

547 rates at 12 weeks, with adjuvant cetuximab, there was no significant improvement in LRC or OS

548 between the two arms. Therefore cetuximab should be delivered per the Bonner protocol when it

549 is used in combination with radiotherapy.

550

551 Summary of systemic therapy for stage IVA-B disease

552 There are no randomized trials that have defined an optimal chemotherapy regimen, but

553 individual studies have most commonly incorporated cisplatin and/or fluorouracil. Although

554 high-dose intermittent (i.e. bolus) cisplatin can be challenging to deliver, it has the longest track

555 record in large multi-institutional trials in the United States, with a well-known and largely

556 predictable side effect profile. Carboplatin-fluorouracil has also been shown to improve OS in

557 comparison to radiotherapy alone and is thus a viable alternative to bolus cisplatin. However,

558 there is far less practitioner experience with delivering this regimen concurrently with head and

559 neck radiotherapy, which is significantly more resource-intensive than cisplatin monotherapy. In NOT TO BE COPIED, DISSEMINATED OR REFERENCED

560 addition, there are few reported toxicity data on carboplatin-fluorouracil in the intensity –

561 modulated radiation therapy (IMRT) era, and whether the low-dose bath seen with IMRT would

562 be particularly accentuated by a well-known mucositic agent (fluorouracil) is unknown. Thus, for

563 patients expected to tolerate high-dose intermittent cisplatin administration, the panel strongly

564 recommends its use for patients with stage IVA-B OPSCC.

565 A non-trivial percentage of patients with OPSCC may not tolerate bolus cisplatin

566 administration for a variety of reasons, including hearing loss or compromised renal function. In

567 this circumstance, the panel strongly recommends the use of concurrent carboplatin-fluorouracil

568 or cetuximab with definitive radiotherapy. As detailed above, the former regimen has a

569 consistent history of improving OS in combination with radiotherapy, and concurrent cetuximab

570 was shown in a phase III trial to improve OS in comparison to radiotherapy alone.35 There are

571 several retrospective studies comparing outcomes with cetuximab-radiotherapy with

572 chemoradiotherapy, but these publications are conflicting and too laden with confounding and

573 selection bias for consideration in this guideline. Therefore the panel did not prioritize one drug

574 over another for patients who do not receive high-dose cisplatin. The results from RTOG 1016

575 should provide important guidance in selecting the optimal concurrent systemic therapy for

576 patients with p16-positive oropharyngeal cancer.

577 Although weekly cisplatin may be an acceptable alternative to high-dose administration,

578 the evidence suggesting a survival benefit with its use is significantly weaker and based on

579 extrapolation rather than high-level evidence. So while the panel concluded that weekly cisplatin

580 may be given if the patient is not a candidate for bolus delivery, the lack of data guiding its use

581 must be made very clear to the patient, especially in the face of high-level data supporting OS

582 gains with competing regimens. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

583 The panel recognized that there is a population of patients with stage IVA-B (and stage

584 III) OPSCC who will not be candidates for concurrent systemic therapy. KQ4 discusses the

585 benefits and risks of altered fractionation radiotherapy to provide additional recommendations on

586 optimal treatment of this cohort.

587

588

589 2. In the scenario of stage III disease? 590 591  Statement 1F. Concurrent systemic therapy should be delivered to patients with T3 N0-1 592 oropharyngeal squamous cell carcinoma receiving definitive radiotherapy. 593 o Recommendation strength: Strong 594 o Quality of evidence: Moderate

595 596  Statement 1G. Concurrent systemic therapy may be delivered to patients with T1-T2 N1 597 oropharyngeal squamous cell carcinoma receiving definitive radiotherapy who are considered at 598 particularly significant risk for locoregional recurrence, after a careful discussion of patient 599 preferences and the limited evidence supporting its use. 600 o Recommendation strength: Conditional 601 o Quality of evidence: Low

602 603 3. In the scenario of stage I-II disease? 604 605  Statement 1H. Concurrent systemic therapy should not be delivered to patients with stage I-II 606 oropharyngeal squamous cell carcinoma receiving definitive radiotherapy. 607 o Recommendation strength: Strong 608 o Quality of evidence: High

609 610 Narrative 611 612 The addition of concurrent systemic therapy to primary radiotherapy for patients with

613 stage III oropharyngeal squamous cell carcinoma may be appropriate in certain patient

614 subgroups. Randomized controlled trials have generally included both stage III and IV patients,

615 with the stage III subgroup representing a minority (4 – 32%) of the total study population in the

616 studies reviewed. These trials are underpowered to address the magnitude of benefit of NOT TO BE COPIED, DISSEMINATED OR REFERENCED

617 concurrent systemic therapy in the stage III subgroup. In addition, stage III patients are

618 heterogeneous, and include those with T3 N0-1 and T1-2 N1 tumors.

619 Since concurrent systemic therapy improves outcomes primarily through

620 radiosensitization, the key issue is the expected LRC probability with radiotherapy alone in this

621 population. Retrospective studies have shown that local tumor control is strongly related to

622 tumor volume and tumor stage. For example, Lok et al.42 reported the effect of tumor volumes on

623 outcomes of 340 OPSCC patients managed with radiation and concurrent chemotherapy at

624 Memorial Sloan Kettering between 1998-2009. Volumes of the primary site gross tumor (GTVp)

625 were calculated from the original IMRT plans. When dichotomized at the median volume (32.79

626 cm3) GTVp was demonstrated to be an independent predictor of 2 year OS on multivariate

627 analysis (94.3% vs 82.7%, p=0.0003). Similarly, two year probabilities of local failure (LF) at

628 the primary site were 1.3 % vs 10.4% (HR=6.01, p=0.004) based on a GTVp cut-off of 32.79

629 cm. 3 The authors observed that GTVp was a more reliable predictor of OS and LF than T

630 category (dichotomized as T1-2 vs T3-4).

631 Garden et al. 43 have reported a retrospective analysis of 1046 OPSCC treated at MD

632 Anderson between 2000 and 2007. Locoregional control at 5 years for 640 patients presenting

633 with T1-2 disease was 91% for those given concurrent chemotherapy and 95% for those treated

634 with radiation alone. In contrast, patients with T3-4 disease treated with concurrent

635 chemoradiotherapy achieved a LRC probability of only 77%, which dropped to 63% with

636 radiotherapy alone. A comparable retrospective analysis from the University of Florida described

637 the LRC outcomes in 130 OPSCC patients treated with definitive radiotherapy, 89% and 61% of

638 whom received AccF and/or concurrent chemotherapy, respectively. When evaluating their local

639 control outcomes, the authors found that T1 and T2 lesions experienced 5-year control NOT TO BE COPIED, DISSEMINATED OR REFERENCED

640 probabilities of 93% and 91%, respectively, whereas T3 and T4 tumors recurred locally 82% and

641 67% of the time, respectively. These data are persuasive that patients with larger volume disease

642 need additional local therapy beyond conventional radiotherapy alone. In fact, the MACH-NC

643 meta-analysis31 showed a significant OS improvement with chemotherapy for patients with stage

644 III head and neck cancer, although the results were not broken down by T stage. Since

645 concurrent chemotherapy is consistently associated with LRC and OS benefits in the

646 aforementioned randomized trials, the panel strongly recommends that patients with T3 N0-1

647 OPSCC receive concurrent systemic therapy with primary radiotherapy.

648 The degree to which low-bulk disease, namely non-T3 stage III (T1-T2, N1) OPSCC,

649 benefit from treatment intensification with concurrent systemic therapy is uncertain, and it

650 remains controversial whether the added toxicity of concurrent systemic therapy is warranted in

651 this population. Consider that the volume of a T2 tumor may range from 8 cm3 to 64 cm3, and

652 thus the baseline risk of local recurrence may vary significantly in this tumor category. Other

653 tumor features may also impact the perceived locoregional failure risk for a patient with T1-2 N1

654 disease, such as its human papillomavirus (HPV) status and the patient’s smoking history. For

655 most patients with the T1-2 N1 OPSCC, the panel believes that radiotherapy alone should be

656 sufficient to obtain LRC. However, certain patients with T1-2 N1 OPSCC who are considered at

657 particularly significant risk for locoregional recurrence may receive concurrent systemic therapy

658 with primary radiotherapy, since the absolute benefit of combined modality therapy may warrant

659 its toxicities in this population. The potential benefit of concurrent systemic therapy is even less

660 compelling in smaller tumors, unless there are other features suggesting radioresistance. Patient

661 preferences must be engaged on the toxicity of concurrent therapy versus modestly improved

662 tumor control, and the limited data guiding this recommendation should be made explicit. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

663 The use of concurrent systemic therapy has not been rigorously examined in patients with

664 stage I and II OPSCC, a population with favorable LRC outcomes with radiotherapy alone. For

665 example, in the MACH-NC meta-analysis, 31 just over 500 patients with stage I-II disease were

666 included in the study (versus over 6,000 for stage IV cancer), and the hazard ratio for

667 chemotherapy was approximately one. Therefore these patients should not be treated with

668 concurrent systemic therapy.

669

670 Key Question 2: When is it appropriate to deliver post-operative radiotherapy with and 671 without systemic therapy following primary surgery of oropharyngeal squamous cell 672 carcinoma (OPSCC)?

673 1. In the scenario of positive margins and/or extracapsular nodal extension (ECE)?

674 675  Statement 2A. Concurrent high-dose intermittent cisplatin should be delivered with post- 676 operative radiotherapy to patients with positive surgical margins and/or extracapsular nodal 677 extension; this high-risk population includes patients independent of HPV status or the extent of 678 extranodal tumor.

679 o Recommendation strength: Strong 680 o Quality of evidence: Moderate 681 682  Statement 2B. Concurrent weekly cisplatin may be delivered with post-operative radiotherapy to 683 patients who are considered inappropriate for standard high-dose intermittent cisplatin after a 684 careful discussion of patient preferences and the limited evidence supporting this treatment 685 schedule.

686 o Recommendation strength: Conditional 687 o Quality of evidence: Low

688 689  Statement 2C. For the high-risk post-operative patient unable to receive cisplatin-based 690 concurrent chemoradiotherapy, radiotherapy alone should be routinely delivered without 691 concurrent systemic therapy; given the limited evidence supporting alternative regimens, 692 treatment with non-cisplatin systemic therapy should be accompanied by a careful discussion of 693 the risks and unknown benefits of the combination.

694 o Recommendation strength: Strong 695 o Quality of evidence: Moderate NOT TO BE COPIED, DISSEMINATED OR REFERENCED

696 697  Statement 2D. Patients treated with post-operative radiotherapy should not receive concurrent 698 weekly carboplatin.

699 o Recommendation strength: Strong 700 o Quality of evidence: Moderate 701 702  Statement 2E. Patients treated with post-operative radiotherapy should not receive cetuximab, 703 either alone or in combination with chemotherapy, although such regimens are currently under 704 investigation.

705 o Recommendation strength: Strong 706 o Quality of evidence: Low 707 708  Statement 2F. Patients treated with post-operative radiotherapy should not routinely receive 709 concurrent weekly docetaxel given the limited evidence supporting its use, although such 710 regimens are currently under investigation.

711 o Recommendation strength: Strong 712 o Quality of evidence: Low

713 714  Statement 2G. Patients treated with post-operative radiotherapy should not receive concurrent 715 mitomycin-C, alone or with bleomycin, given the limited evidence and experience supporting 716 its use.

717 o Recommendation strength: Strong 718 o Quality of evidence: Moderate

719 720  Statement 2H. Post-operative chemotherapy should not be delivered alone or sequentially with 721 post-operative radiotherapy.

722 o Recommendation strength: Strong 723 o Quality of evidence: High 724

725 Narrative 726 727 Is there a patient population that will benefit from the addition of chemotherapy to post-

728 operative radiotherapy? What drug and treatment schedule should be used?

729 The results of two landmark trials of adjuvant chemoradiotherapy, EORTC 2293144 and

730 RTOG 950145, were published simultaneously in 2004 (Table 3). Both studies compared post- NOT TO BE COPIED, DISSEMINATED OR REFERENCED

731 operative radiotherapy (PORT) with post-operative chemoradiotherapy, using concurrent high-

732 dose intermittent cisplatin (100 mg/m2 every three weeks for three doses) in patients deemed at

733 high risk for disease recurrence. High-risk disease was defined as positive surgical margins, or

734 extracapsular nodal involvement in both trials, but the RTOG study45 also included involvement

735 of two or more regional nodes, and the EORTC study44 included perineural disease, vascular

736 embolism, or level 4 or 5 nodal involvement in oral cavity or oropharynx primary cancers. It

737 should be noted that in the RTOG trial, a positive surgical margin was defined as microscopic

738 tumor at the tumor specimen edge. In the EORTC trial44 however, tumor within 5 mm of the

739 specimen edge was considered a positive margin. Initial results from both studies demonstrated

740 an improvement in LRC and disease/progression-free survival with chemoradiotherapy. Overall

741 survival was statistically better in the EORTC study44 but only trended towards improvement in

742 the first RTOG report. Long-term follow-up from the RTOG trial has been reported46, however,

743 and it no longer demonstrated any statistical benefit from the chemotherapy in the primary

744 comparisons.

745 It is of note that distant metastases were not reduced by the chemotherapy in either trial,

746 and all three cisplatin doses could be given to only 49% and 61% of the EORTC44 and RTOG45

747 patients respectively. Acute toxicity, specifically mucositis, myelosuppression, nausea and

748 vomiting, was increased in the chemoradiotherapy patients. However, no statistically significant

749 increase in total late grade 3-5 toxicity could be identified after chemoradiotherapy in either

750 study. This acute toxicity burden must temper any enthusiasm to use concurrent chemotherapy

751 in an unselected population of patients receiving PORT.

752 The difference in survival outcomes, and the long-term results from the RTOG study

753 merit further discussion. Bernier and colleagues47 performed a pooled analysis of the 750 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

754 patients from both trials, and confirmed an improvement in all endpoints, including OS, in the

755 patients treated with chemotherapy. They suggested, however, that not all of the “high-risk”

756 eligibility criteria conferred sufficient risk to merit the addition of concurrent chemotherapy. In

757 their retrospective, unplanned subgroup analysis from these two trials, the addition of concurrent

758 cisplatin proved beneficial only in those high-risk patients with either positive surgical margins

759 or extracapsular nodal involvement. While benefit in patients with the other risk features

760 remained a possibility, it could not be statistically demonstrated. Similar conclusions were drawn

761 from the long-term RTOG 9501 follow-up report.46 Again using retrospective, unplanned

762 subgroup analysis, an improved loco-regional control and disease-free survival (with a strong

763 trend towards an improved OS) was identified in the chemoradiotherapy-treated patients with

764 positive surgical margins or extracapsular nodal extension. These observations have led to the

765 current strong recommendations for the addition of concurrent high-dose intermittent cisplatin to

766 adjuvant radiotherapy in these two high-risk post-operative populations. Patients with other risk

767 features, whose surgical procedure and/or pathologic findings imply a particularly significant

768 risk of locoregional recurrence may still benefit from concurrent cisplatin, although the evidence

769 is limited and inconclusive.

770 The improved prognosis and increasing incidence of HPV-positive oropharynx cancer,

771 however, could not be assessed in these trials and may have impacted the observed results.48 In

772 the EORTC study 30% of the patients had an oropharynx cancer compared to 42% in the RTOG

773 report (48% on the chemoradiotherapy arm and 37% of the radiotherapy alone arm). It is

774 unknown if these good prognosis HPV-positive patients require or benefit from more intensive

775 adjuvant therapy, even in the presence of conventional “high-risk” features. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

776 The implications of extracapsular nodal extension in oropharynx cancer, and in

777 particular, the HPV-positive patient have also been questioned. Retrospective reports from

778 several groups have suggested that extracapsular nodal disease has limited prognostic importance

779 in these patients, especially when it is only microscopically present in an HPV-positive patient.49-

780 52 Prospective studies that replicate these provocative retrospective results are needed before

781 PORT alone can be recommended for HPV-positive head and neck cancer with ECE. In addition,

782 to enhance its prognostic power, ECE reporting has recently been refined with the introduction

783 of a five point scale (0-4), and further multi-institutional study is required for this modification to

784 be used in clinical practice.53

785 Reasonable concern clearly exists about the applicability of RTOG 950145 and EORTC44

786 22931 to the growing population of patients with HPV-positive oropharynx cancer48 and the

787 prognostic implications of early extracapsular nodal involvement. To date, however, these two

788 trials provide the best available evidence and continue to drive the current adjuvant

789 recommendations. Well-designed HPV-specific trials are currently underway to better address

790 these many unresolved questions. Similar concerns have also led to greater sophistication and

791 nuance in the studies designed for the definitive management of oropharynx cancer. Such

792 concerns do not render current treatment standards invalid; they only lend urgency to further

793 investigation.

794

795 Can radiotherapy and a weekly concurrent cisplatin administration schedule be used instead of

796 giving 100 mg/m2 given every three weeks? NOT TO BE COPIED, DISSEMINATED OR REFERENCED

797 The high-dose, every three week cisplatin administration schedule (100 mg/m2 q3weeks

798 for 3 doses) has proven difficult to administer. It is associated with gastrointestinal, renal,

799 mucosal and neurologic toxicity, and most large multi-institutional studies (including RTOG

800 9501 and EORTC 22931) have reported that only 50-70% of patients can receive all three drug

801 doses. In recent years there has been considerable interest in the lower dose weekly drug

802 administration schedules (30-40 mg/m2/week) based on the unproven assumption being made

803 that such regimens will produce less toxicity, and, as such will allow equivalent or greater drug

804 administration with equal treatment efficacy. No prospective, randomized comparisons have yet

805 been completed comparing these two drug treatment schedules, however, and their equivalence

806 is unknown.

807 In the post-operative setting Bachaud et al. 54,55 reported the results of a very small phase

808 III trial comparing PORT alone with radiotherapy and weekly cisplatin; 50 mg/week

809 (approximately 25-30 mg/m2) for 7-9 doses in 83 patients (14% oropharynx cancer). Both

810 disease-free and OS, as well as OS without locoregional treatment failure, were significantly

811 improved in the chemotherapy treated patients. Distant metastases were not impacted by the

812 chemotherapy, although toxicity was significantly worse.

813 Geiger et al.56 conducted a retrospective analysis of 104 post-operative patients treated

814 with radiotherapy and either concurrent high-dose intermittent (100 mg/m2 q3weeks for 3 doses),

815 or low dose weekly (25-30 mg/m2/week for 6 doses) cisplatin. Sixty-one percent of the patients

816 had an oropharynx cancer, and 86% of these oropharynx cancers were HPV/p16 positive. Total

817 drug administration was significantly greater in the high-dose intermittent group, but no

818 difference was observed in either relapse-free or OS between patients treated with the two NOT TO BE COPIED, DISSEMINATED OR REFERENCED

819 different cisplatin schedules. There was also no outcome difference identified when the analysis

820 was limited to the HPV-positive oropharynx cancer patients.

821 Because of the increasing community acceptance of weekly cisplatin regimens despite

822 this lack of demonstrated equivalence, the NRG adopted weekly dosing as the control arm in

823 RTOG 1216, its current phase II/III trial exploring the substitution of docetaxel and cetuximab

824 for cisplatin in the postoperative adjuvant treatment of patients with high risk disease. Despite

825 the inclusion of weekly cisplatin as the control arm in this study, the panel does not feel the

826 evidence supporting its use is currently sufficient to strongly recommend its delivery in lieu of

827 high-dose intermittent cisplatin when the latter would be tolerated, as the high-dose regimen was

828 evaluated in two large prospective phase III randomized trials with positive results.

829

830 Are there alternative systemic treatment regimens that can be used concurrently with

831 radiotherapy, particularly in the medically compromised patient?

832 Cisplatin remains the only systemic agent with level 1 evidence supporting its use

833 concurrently with radiotherapy in high-risk post-operative patients. Several other systemic agents

834 have been tested in small phase II and phase III trials with limited success (Table 4). All studies

835 report greater toxicity when chemotherapy is added to radiotherapy without a proven survival

836 benefit.

837 Radiotherapy and concurrent single agent mitomycin C (15 mg/m2 for 1-2 doses) has

838 been compared to a radiation therapy control arm in three separate randomized trials at Yale

839 University.57-61 Pooled results from the 205 post-operatively treated patients (23% oropharynx

840 cancer) have been reported. Conventional high-risk features were not required for entry on these NOT TO BE COPIED, DISSEMINATED OR REFERENCED

841 trials although 27% had positive surgical margins and 34% had more than one positive node.

842 Locoregional control proved statistically superior in those patients treated with mitomycin C;

843 however, no difference was observed in the rate of distant metastases or in overall survival.

844 Post-operative radiotherapy with mitomycin C (15 mg/m2) was also studied by Smid,

845 Zakotnik et al.,62,63 who combined it with bleomycin (5 mg twice weekly) and compared it to

846 radiotherapy alone in 114 patients (30% oropharynx cancer); 59% of whom had either positive

847 surgical margins or extracapsular nodal extension. Both LRC and disease-free survival were

848 significantly improved in the chemotherapy treated patients, although OS only trended better

849 (55% vs. 37%, p=0.09). Distant metastases were not improved by the use of chemotherapy.

850 This work, from both groups, provided modestly encouraging evidence of a potential role

851 for mitomycin C in this setting, although patient sample sizes were small. Mitomycin C has not

852 been further pursued in part due to the stronger data that emerged about concurrent cisplatin, and

853 in part due to general concerns about the toxicity of this agent. Because of both the lack of an OS

854 benefit and absence of modern experience of combining this agent with radiotherapy, the panel

855 strongly recommends not combining mitomycin C with post-operative radiotherapy.

856 Concurrent carboplatin has also been studied. Racadot et al.64 report the results of a

857 randomized comparison between PORT alone, and PORT with concurrent carboplatin (50 mg/m2

858 twice weekly) in 144 patients (49% oropharynx cancer) with node positive disease. No outcome

859 differences were observed. Similarly, Argiris et al.65 studied a post-operative patient population

860 (32% oropharynx cancer) defined by positive margins, multiple node positivity, angiolymphatic

861 or perineural invasion (PNI), or extracapsular nodal involvement, randomizing between

862 radiotherapy and radiotherapy with concomitant carboplatin 100 mg/m2 weekly. Only 72 patients NOT TO BE COPIED, DISSEMINATED OR REFERENCED

863 were evaluated on the trial and no statistically meaningful outcome difference was observed,

864 perhaps reflecting the insufficient sample size. Further studies of post-operative carboplatin and

865 radiotherapy have not been reported, and since both randomized trials were negative, the panel

866 strongly recommends against using concurrent carboplatin in the post-operative setting.

867 A large multi-institutional phase III study from the United Kingdom randomly compared

868 PORT with radiotherapy and chemotherapy (either methotrexate alone, or methotrexate,

869 vincristine, bleomycin and fluorouracil) in 253 post-operative patients (23% oropharynx cancer).

870 Patients were described as “generally at high risk due to margin status or advanced stage of

871 disease.” No outcome benefit was observed in the chemotherapy treated patients using this non-

872 platinum containing regimen.66

873 Cetuximab and the other EGFR inhibitors have been considered promising agents in the

874 management of head and neck squamous cell cancer. When combined with radiotherapy in the

875 definitive treatment of locoregionally advanced disease, a survival benefit (but no impact on

876 distant metastases) has been demonstrated in comparison to radiation therapy alone. This result

877 has suggested the possibility that radiotherapy and cetuximab may also be a valuable post-

878 operative adjuvant regimen to explore. The NRG is currently comparing this combination to

879 radiation therapy alone in “intermediate risk” patients after surgical resection in RTOG 0920.67

880 However, since there are no prospective or even retrospective studies suggesting efficacy in this

881 scenario, the panel strongly recommends against using cetuximab with PORT if not on clinical

882 trial.

883 The taxanes (paclitaxel and docetaxel) have been investigated by the RTOG in a series of

884 post-operative trials.68 RTOG 0024 explored immediate post-operative paclitaxel followed by NOT TO BE COPIED, DISSEMINATED OR REFERENCED

885 concurrent chemoradiotherapy with weekly paclitaxel and cisplatin in 70 high-risk post-operative

886 patients (34% oropharynx cancer) in a single arm phase II trial.68 Toxicity appeared acceptable

887 and results comparable to historical controls using high-dose cisplatin. This study was followed

888 by RTOG 0234, a phase II randomized trial comparing post-operative cetuximab and

889 radiotherapy with either concurrent weekly cisplatin or weekly docetaxel, in 203 high risk

890 patients (36% oropharynx cancer).69 The radiotherapy, cetuximab and docetaxel arm appeared

891 more promising and this combination is now being formally tested in a phase II/III trial70 in

892 comparison to radiotherapy and cisplatin. Pending the results of this study, though, the use of

893 docetaxel with PORT – a combination never prospectively evaluated – should not be routinely

894 implemented with post-operative radiotherapy.

895 It should again be noted that the benefit from cisplatin reflects an improvement in loco-

896 regional control, not a reduction in distant metastases. There is no phase III evidence

897 demonstrating an improvement in distant metastatic recurrence from any systemic agent in this

898 setting. In patients with a contraindication to cisplatin, radiation therapy alone provides a LRC

899 benefit, and is the recommended treatment choice, even when the risk of disease recurrence is

900 high.

901

902 Is there a role for post-operative adjuvant chemotherapy alone, or for the sequential

903 administration or chemotherapy and radiotherapy, rather than concurrent treatment?

904 The MACH-NC Meta-analysis 31,71 reported on 12 trials of more than 2500 patients,

905 comparing adjuvant chemotherapy after locoregional treatment to locoregional treatment alone. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

906 No impact could be identified on either disease recurrence or death from the use of the adjuvant

907 chemotherapy. Adjuvant chemotherapy is not considered a standard approach in this disease.

908 Laramore et al.72 reported the results of Intergroup study 0034, a phase III randomized

909 trial conducted in 448 patients (25% oropharynx cancer) comparing PORT alone, with sequential

910 chemotherapy (fluorouracil and cisplatin for 3 cycles) followed by radiotherapy. No difference in

911 LRC, disease-free survival or OS could be identified between the two treatment arms, although a

912 statistically significant increase in distant metastases was seen in the patients not given

913 chemotherapy. Sequential treatment with post-operative chemotherapy and radiotherapy is not

914 recommended in this disease.

915

916 2. In the scenario of intermediate-risk pathologic factors such as lymphovascular invasion 917 (LVI), perineural invasion (PNI), T3-4 disease, or positive lymph nodes?

918 919  Statement 2I. Patients with intermediate-risk factors should not routinely receive concurrent 920 systemic therapy with post-operative radiotherapy.

921 o Recommendation strength: Strong 922 o Quality of evidence: Moderate 923 924  Statement 2J. Patients with intermediate-risk factors whose surgical procedure and/or 925 pathologic findings imply a particularly significant risk of locoregional recurrence may receive 926 concurrent cisplatin-based chemotherapy after a careful discussion of patient preferences and the 927 limited evidence supporting its use in this scenario; alternative systemic treatment regimens 928 should only be used in the context of a clinical trial.

929 o Recommendation strength: Conditional 930 o Quality of evidence: Low 931 932  Statement 2K. Post-operative radiotherapy should be delivered to patients with pathologic T3 or 933 T4 disease.

934 o Recommendation strength: Strong 935 o Quality of evidence: Low NOT TO BE COPIED, DISSEMINATED OR REFERENCED

936 937  Statement 2L. Post-operative radiotherapy should be delivered to patients with pathologic N2 or 938 N3 disease.

939 o Recommendation strength: Strong 940 o Quality of evidence: Low 941 942  Statement 2M. Post-operative radiotherapy may be delivered to patients with pathologic N1 943 disease after a careful discussion of patient preferences and the limited evidence of outcomes 944 following surgery alone in this scenario.

945 o Recommendation strength: Conditional 946 o Quality of evidence: Low 947 948  Statement 2N. Post-operative radiotherapy may be delivered to patients with LVI and/or PNI as 949 the only risk factor(s) after a careful discussion of patient preferences and the limited evidence of 950 outcomes following surgery alone in this scenario.

951 o Recommendation strength: Conditional 952 o Quality of evidence: Low

953 954 Narrative 955 956 The risk of locoregional recurrence following definitive surgical resection and neck

957 dissection is influenced by several factors. The clinical and pathologic features most clearly

958 associated with subsequent locoregional progression are positive margins and extracapsular

959 nodal extension, the implications of which were discussed above. On the other hand, due to the

960 paucity of prospective randomized trials of adjuvant radiotherapy alone versus observation

961 following surgery, the vast majority of data implicating other risk factors with locoregional

962 recurrence are retrospective. Thus, there is almost no high-level evidence to guide adjuvant

963 treatment recommendations in the negative margin, non-ECE setting, despite several other risk

964 factors serving as eligibility criteria for active prospective randomized trials evaluating the role

965 of adjuvant radiation therapy with or without chemotherapy. Nevertheless, this population

966 comprises a significant percentage of patients for whom PORT must be considered, and so the NOT TO BE COPIED, DISSEMINATED OR REFERENCED

967 panel used higher-quality retrospective data and expert opinion to derive treatment

968 recommendations (Table 5).

969 Pathologic nodal stage is a commonly used pathologic risk factor in the adjuvant

970 decision-making process. Resection of bulkier nodes may leave behind occult microscopic

971 disease, and with more pathologically involved nodes, there is theoretically a higher likelihood

972 of additional microscopic nodal deposits in dissected or undissected tissue. Indeed, the evidence

973 supports a significant risk of regional progression in patients with N2 or N3 disease. In one of the

974 few prospective trials of adjuvant radiotherapy dose escalation, Peters et al. showed that among

975 all patients, those with more than 1 positive node (i.e. N2b) experienced a higher risk of

976 locoregional recurrence (30% vs 16%, p=0.08), despite radiotherapy delivery to all patients.

977 Langendijk et al.73 performed a large retrospective analysis to define subgroups of patients with

978 locally advanced head and neck cancer (20% oropharyngeal cancer) at higher risk for

979 locoregional recurrence following surgery and post-operative radiotherapy. On multivariable

980 analysis, patients with pN2b-N2c and pN3 disease had significantly higher risks of locoregional

981 recurrence in comparison to pN0-N2a status (RR 1.7 and 3.5, respectively), and patients with an

982 N3 neck were placed in the “very high risk” category (5 year LRC of 58%) regardless of ECE or

983 the primary tumor characteristics. Similarly, Ambrosch et al.74 reported a large series of 503

984 patients treated with neck dissection with or without adjuvant radiotherapy. The 48 patients with

985 N2 disease treated without adjuvant radiotherapy experienced a 3-year neck recurrence risk of

986 24%, in comparison to 7% of the 188 N2-positive treated with post-operative radiotherapy.

987 Notably, among all patients with regional recurrence, successful salvage was feasible in only

988 24% of patients, and 67% of patients died specifically from the neck progression. In a smaller

989 surgical series from Mayo Clinic, 2 out of 11 (17%) pathologic N2b oropharynx cancer patients NOT TO BE COPIED, DISSEMINATED OR REFERENCED

990 observed after neck dissection developed a regional recurrence, as did 1 of 2 pathologic N3

991 patients.

992 The risk of regional recurrence in patients with a pathologically N1 neck is more

993 controversial. In a paper restricted to individuals with pN1 disease without extracapsular

994 extension (29% oropharynx), Jackel et al.75 reported the risks of locoregional recurrence with

995 and without adjuvant radiotherapy. Patients observed after primary surgery (n=72) experienced a

996 crude risk of any regional recurrence of 21%, with 10% of observed patients developing an

997 isolated nodal failure; the estimated 3-year risk of isolated neck recurrence in this population was

998 11.2% (vs. 2.9% in patients who were irradiated, p=0.09); three of the 7 patients with isolated

999 regional recurrence died from cancer. Somewhat conflicting data on the N1 neck were presented

1000 in the Ambrosch study,74 in which the 3-year neck recurrence risk for the 48 pN1 patients treated

1001 with surgery alone was 6.3%. This number was only slightly higher than the neck recurrence risk

1002 in pN0 patients observed after surgery (5.5% out of 213 patients). However, only 28% of these

1003 patients had oropharyngeal cancer, and the majority was either oral cavity or larynx, which have

1004 different patterns of nodal spread. In particular, oropharyngeal cancer is significantly more likely

1005 to metastasize to the retropharyngeal nodes, which are treated with radiotherapy but not accessed

1006 with conventional neck dissection; the absolute risk of retropharyngeal lymph node involvement

1007 in OPSCC has been reported between 20-30%.76,77 Smaller, modern series in oropharynx cancer

1008 also suggest high control rates following neck dissection alone for N1 disease. None of the 10

1009 patients observed with pN1 disease in the Mayo clinic series developed a regional recurrence,

1010 and none of the 10 patients with pN1 disease in the Penn prospective study of transoral robotic

1011 surgery (TORS) 78 alone developed a neck progression. One significant consideration in

1012 evaluating all of these data is the reality that the quality of the neck dissections was not formally NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1013 standardized, both in terms of surgical technique (e.g. levels dissected and number of nodes

1014 harvested) and pathologic evaluation processes. Thus one must recognize this limitation in

1015 applying these results to a given neck dissection.

1016 Two related surveillance, Epidemiology, and End Results Program (SEER) studies

1017 investigated the survival benefit of adjuvant RT following surgery for non-nasopharynx head and

1018 neck carcinoma with positive lymph nodes, and both showed a significant OS advantage for all

1019 nodal stages. For instance, Kao et al.79 showed that the absolute survival gains for patients with

1020 N1, N2a, N2b, and N2c-3 were 8.2%, 16.6%, 23.9%, and 12.2%, respectively (all p<0.0001).

1021 Patients with oropharyngeal cancer experienced an absolute survival improvement of 8.2% with

1022 adjuvant radiotherapy (p=0.0003). Although these survival improvements are impressive, any

1023 SEER study is limited by the nature of its retrospective, population-based data collection. For

1024 example, performance status, chemotherapy administration, ECE status, and pre-treatment

1025 imaging were all unknown. Nevertheless, these data do support a benefit of adjuvant

1026 radiotherapy for all node-positive disease.

1027 Taken together, this evidence is persuasive that individuals with pathologic N2 and N3

1028 disease experience an unacceptably high regional recurrence risk without radiotherapy, and

1029 therefore the panel strongly recommends that these patients receive adjuvant RT. Patients with

1030 pathologic N1 oropharynx cancer are clearly at lower risk for nodal progression following neck

1031 dissection, but there is substantial uncertainty on the true probability of regional recurrence in

1032 this population. Moreover, while it is intuitive and accepted that patients not receiving

1033 radiotherapy will have improved quality-of-life in comparison to those who do, the poor salvage

1034 ability (and potential mortality) of a neck recurrence must be considered in this decision. Since

1035 the panel favors a potential progression and/or survival benefit over modest, albeit meaningful, NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1036 quality-of-life improvements, it conditionally recommends adjuvant radiotherapy in this

1037 population. That said, it is critical to discuss the uncertainty in the recurrence risk with the

1038 patient, and the physician should engage patient preferences and values on the balance between

1039 the risks of radiotherapy versus the potential for unsalvageable regional progression.

1040 Pathologic T stage has also been frequently investigated in retrospective studies as a

1041 predictor of locoregional recurrence. Leemans et al80 reported on 244 head and neck cancer

1042 patients (13% oropharynx) treated with primary surgery with or without adjuvant radiotherapy.

1043 Clinical T stage was the most significant predictor of local recurrence, with a crude recurrence

1044 risk of 5.3% for T1-2 versus 16.2% for T3-4 disease (p=0.015), and the actuarial absolute

1045 difference at 5 years was almost 20%. In a retrospective analysis of 420 oropharyngeal and

1046 hypopharyngeal patients from the Centre Henri Becquerel treated with surgery and PORT

1047 radiotherapy, advancing T stage was strongly associated locoregional relapse (relative risk 1.85,

1048 p<0.0001) on multivariable analysis, with T stage treated as a continuous variable; absolute risks

1049 were not stated. In the Langendijk73 analysis in which all patients received adjuvant

1050 radiotherapy, T stage did not correlate with LRC on univariable analysis. However, after

1051 identifying patients with positive or close margins (5 mm or less, which is typically seen in

1052 larger tumors of the oropharynx), patients with T3-4 disease experienced significantly worse

1053 LRC than those with T1-2 disease (absolute difference 10%). The majority of these failures were

1054 in the T3 population (5-year LRC of 51%).

1055 A retrospective analysis of p16-positive patients by Haughey et al.81 nicely showed a

1056 significant increase in the risk of any recurrence as a function of pathology T stage: 1.5%, 5.5%,

1057 11.5%, and 28.6% for T1, T2, T3 and T4 respectively. Indeed, on multivariable analysis of

1058 disease-free survival, both clinical T stage (T3-4 vs T1-2) and pathologic T stage (T3-4 vs T1-2) NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1059 increased the hazard by over 3-fold. However, there were only 2 local recurrences in the entire

1060 series (both in T3 patients, 1 of whom did not receive adjuvant radiotherapy), so T stage was not

1061 a predictor of local-only recurrence. In the Ambrosch et al.74 study that primarily focused on

1062 regional recurrences after neck dissection, pT3/pT4 disease was significantly associated with

1063 worsened OS on multivariable analysis (risk ratio 1.6, p=0.0009), but local failure patterns were

1064 not mentioned.

1065 Both of these latter studies highlight a major challenge in analyzing the literature to

1066 derive adjuvant treatment recommendations in patients with T3-4 disease: almost all of these

1067 individuals received post-operative radiotherapy. While the retrospective data are convincing

1068 that these patients have a worse overall prognosis, it is hard to calculate an absolute local

1069 recurrence risk with and without radiotherapy. Nevertheless, the panel concluded that the

1070 existing literature support an increased risk of local recurrence with pathologic T3 or T4 disease

1071 even with radiotherapy, and moreover, there are insufficient LRC data to support the oncologic

1072 safety of observation following resection of T3-4 oropharyngeal cancer. Therefore the panel does

1073 strongly recommend adjuvant RT in this population.

1074 It is particularly difficult to estimate the risk of locoregional recurrence in patients for

1075 whom PNI or LVI is the only adverse pathologic factor. These characteristics are often found in

1076 patients with other known risk factors for recurrence and this confounding can be difficult to

1077 resolve.81 Secondly, retrospective studies don’t always comment on these factors because they

1078 are not always properly recorded in the pathology report. Finally, historically radiation

1079 oncologists have always treated patients with LVI or PNI, and there is very little information on

1080 outcomes without adjuvant radiotherapy. For example, the large retrospective analysis from

1081 Langendijk et al.73 found that patients with PNI had a significantly higher risk of locoregional NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1082 recurrence (31% vs 22% at 5 years, p<0.026), as did patients with angioinvasive (33% vs 21%,

1083 p=0.011) or lymphangioinvasive growth (38% vs. 23%, p=0.046). However, none of these

1084 factors were significant on multivariable analysis. Haughey et al. found on univariable analysis

1085 that angioinvasion was borderline associated with disease-free survival (p=0.062) and powerfully

1086 associated with disease-specific survival (HR 13.16, p=0.002). These relationships became non-

1087 significant on multivariable analysis. Moreover, only 26 patients had this pathologic factor, and

1088 of these 26, five recurred but 4 were distant metastases. There was no association between PNI

1089 and outcome in this large analysis.

1090 There are data supporting these characteristics as independent risk factors for

1091 locoregional recurrence. In a retrospective analysis of 80 consecutive head and neck patients

1092 (28% oropharynx) treated with primary surgery at the University of Utrecht, LVI was

1093 significantly associated with local failure on both univariable analysis (50% failure with vs. 16%

1094 without, p=0.001) and multivariable analysis (p=0.005). Perineural invasion was qualitatively

1095 associated with local failure (27% with and 15% without), but it was non-significant on both

1096 univariable and multivariable analysis.82 In the retrospective study from the Centre Henri

1097 Becquerel,83 individuals with tumor emboli in the vessels or nerves experienced an

1098 approximately 10% increased risk of locoregional recurrence. However, in a multivariable

1099 analysis of local relapse incorporating T and N stage, margin status, age, tumor emboli was

1100 marginally non-significant (relative risk 1.48, p=0.061). In another bi-institutional retrospective

1101 analysis of oral cavity and oropharyngeal cancer (33% oropharynx), McMahon et al.84 showed

1102 that PNI was a significant predictor of local recurrence in both institutions, and vascular and LVI

1103 was a significant predictor of local recurrence in only one of them. However, on multivariable

1104 analysis incorporating all patients, only PNI (over all other variables) was significantly NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1105 associated with local recurrence. Neither PNI nor LVI were independently associated with

1106 disease-specific survival on multivariable analysis. Fagan et al.85 retrospectively evaluated the

1107 prognostic influence of PNI in 142 patients (66 with oral cavity or oropharynx) treated with

1108 primary surgery with or without adjuvant radiotherapy. Over half of the patients had PNI, and it

1109 was significantly associated with local recurrence (23% vs. 9%, p=0.02) and disease-specific

1110 mortality. Finally, Lanzer and colleagues analyzed locoregional and distant recurrence patterns

1111 in 291 head and neck cancer patients, 32% of whom had oropharynx cancer. Among all patients,

1112 the presence of PNI was associated with a higher risk of ipsilateral and contralateral lymph node

1113 recurrence, but there were no independent associations between PNI and disease-free or overall

1114 survival.86

1115 Two retrospective analyses restricted to oral cavity cancer highlight the uncertainty in the

1116 adjuvant management of this population. In a large series of 460 patients with node-negative oral

1117 cavity cancer87, Liao et al. showed that the presence of PNI was not associated with local control

1118 or OS and the addition of radiotherapy to patients with PNI did not improve any outcome; only

1119 44 patients with PNI were observed. On the other hand, Chinn et al.88 reported on a cohort of 88

1120 node-negative oral cavity patients from the University of Michigan, finding that PNI was

1121 associated with worse LRC (35% vs. 12%, p=0.037) and a shorter disease-free interval on

1122 multivariable analysis; in addition, the addition of radiotherapy significantly improved both

1123 endpoints.88 However, only six patients with PNI were observed, limiting the interpretation of

1124 this comparison.

1125 In summary, these entirely retrospective data are mixed on the relationship between

1126 locoregional failure and LVI and PNI, but there is certainly the suggestion that they reflect more

1127 aggressive locoregional disease. Given the morbidity and mortality risk of local recurrence, NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1128 adjuvant radiotherapy may be used for patients with either pathologic factor as the only adverse

1129 characteristic. Both the lack of high-quality evidence and the conflicting retrospective data

1130 guiding this recommendation need to be described to patients, and their preferences and values

1131 on toxicity versus the risk of recurrence need to be explored.

1132 3. In the scenario of no pathologic risk factors?

1133 1134  Statement 2O. Post-operative radiotherapy may be delivered to patients without conventional 1135 adverse pathologic risk factors only if the clinical and surgical findings imply a particularly 1136 significant risk of locoregional recurrence, after a careful discussion of patient preferences and 1137 the potential harms and benefits of radiotherapy.

1138 o Recommendation strength: Conditional 1139 o Quality of evidence: Low 1140 1141 Narrative 1142 1143 There are limited prospective data that report outcomes following primary surgery alone

1144 for oropharyngeal cancer. In a multi-institutional prospective trial evaluating a post-surgical

1145 adjuvant therapy paradigm, Ang et al.89 demonstrated that when the surgical bed was at a low

1146 risk for relapse, defined by the absence of any adverse pathologic features (i.e. T1-2, N0,

1147 negative margin, no PNI, non-oral cavity site), observation was associated with a local-regional

1148 relapse of approximately 10% at 5-years, despite significantly older staging and surgical

1149 techniques. Investigators from University of Pennsylvania characterized the oncologic results of

1150 30 patients treated with TORS alone on a prospective trial.78 Positive margins were defined as

1151 tumor within 2 mm of the margin, and the surgeons paid exquisite detail to the margin

1152 evaluation. All patients in this cohort had final negative margins, although one patient required

1153 an additional resection for carcinoma in situ at the inked margin. Three patients had ECE, and 15

1154 patients were node-positive (10 N1, 5 N2a or N2b). With a minimum follow-up of 18 months,

1155 one patient (3% of total) with initial T2N0 stage developed a local recurrence, successfully NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1156 salvaged with chemoradiotherapy. Three patients (10% of total) experienced a regional

1157 recurrence, two of whom had N2b nodal stage at original presentation and refused the

1158 recommendation to receive adjuvant radiotherapy. One patient with T2N0 cancer recurred in the

1159 ipsilateral high-neck. All patients were alive without evidence of disease at the time of the

1160 publication.

1161 Psychogios et al.90 published a retrospective series of 228 patients with T2 N0-3

1162 oropharyngeal cancer treated with primary surgery with or without adjuvant radiation treatment.

1163 Out of the 69 patients with negative margins and no pathologically involved lymph nodes, there

1164 was no difference in disease-specific survival between the 20 observed patients versus the 49

1165 treated with radiation therapy (81.5% vs. 80.0%, p=0.361).

1166 A joint publication from Mayo-Jacksonville and Mayo-Arizona retrospectively analyzed

1167 outcomes following transoral laser microsurgery alone in 69 patients, all of whom had negative

1168 margins. 91 Twenty-five of these patients had an indication for RT (i.e. ECE, N2 status, LVI) but

1169 declined, and 44 individuals did not have an indication for RT. After a mean follow-up of 44

1170 months, only 3 patients recurred at the primary site (two T1, one T2 cancer, for a 5-year local

1171 control estimate of 95%). Forty-four patients were recommended to undergo observation, and 4

1172 of these patients recurred in the neck, leading to 5-year LRC estimates of only 90%, 73%, and

1173 70% for stage I, II and III, respectively. However, two of those 4 patients did not have a neck

1174 dissection and were observed. The 5-year disease-specific and OS probability for stage I and II

1175 disease was 88% and 79%, respectively. Of the 25 patients with an indication for adjuvant

1176 radiotherapy, 4 developed a regional failure, with a 5-year estimate of LRC of 74%. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1177 One of the challenges in interpreting pathologic data following transoral excision of

1178 oropharyngeal cancer is the lack of meaningful data on the appropriate surgical margin, which is

1179 not standardized. The most comprehensive study of margin status in head and neck cancer was

1180 based on 827 oral cavity cancer patients (343 stage I-II) and suggested that local control rates

1181 were significantly improved with a margin width of 7 mm or more, although even wider margin

1182 distances also seemed to predict for local failure.87 Of course, none of these patients had

1183 oropharyngeal cancer, and it is unclear whether these results can be translated to a separate

1184 disease site with different anatomy and even biology. For example, Wong and colleagues found

1185 no difference in local recurrence between close (1-5 mm) and negative (> 5 mm) margins treated

1186 with surgery alone, although only 32 of 192 patients had oropharyngeal cancer.92 There is a

1187 general consensus among head and neck surgeons that invasive cancer within 5 mm or less is an

1188 adverse prognostic feature requiring adjuvant therapy, 92 but whether this rule of thumb holds

1189 with modern surgical techniques in the oropharynx is unknown. Thus, while patients with

1190 pathological stage I-II disease with wide margins, a pathologically negative neck and no other

1191 adverse pathologic factors can typically be observed, patients whose surgical procedure or

1192 margin width are more concerning for local recurrence may be considered for adjuvant therapy.

1193 Careful discussion and collaboration between the members of the multidisciplinary team is

1194 necessary to optimize locoregional therapy, and patient preferences on the relative risks and

1195 benefits of radiotherapy versus observation must be understood.

1196 The active phase II study Eastern Cooperative Oncology Group (ECOG) 331193 will

1197 significantly improve our ability to estimate LRC rates following surgery alone for patients with

1198 p16-positive oropharynx cancer. Patients with T1-2, N0-N1 (without extracapsular extension) NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1199 and clear (> 3 mm) margins, will be observed after transoral surgery, and the final results of this

1200 study will further refine these recommendations.

1201 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1202 Key Question 3: When is it appropriate to use induction chemotherapy in the treatment of 1203 oropharyngeal squamous cell carcinoma (OPSCC)? 1204 1205  Statement 3A. Induction chemotherapy should not be routinely delivered to patients with 1206 OPSCC. 1207 o Recommendation strength: Strong 1208 o Quality of evidence: High 1209

1210 Narrative 1211 1212 The curative treatment for the majority of patients with locally advanced OPSCC requires

1213 a multidisciplinary effort, often involving various combinations of radiotherapy, chemotherapy,

1214 and surgery. Despite decades of clinical research (Table 6), however, the optimal integration of

1215 chemotherapy with radiotherapy in this setting has been controversial, and study interpretation

1216 has been complicated by changing local therapy paradigms and epidemiologic shifts. It has been

1217 recognized since the mid-1980s that patients demonstrated high response rates to neoadjuvant or

1218 IC, and the clinical utility of IC was established in landmark trials of advanced larynx (VA

1219 Larynx) and pyriform sinus (EORTC) cancers, which revealed that radical surgery was avoidable

1220 without any decrement in survival. However, whether IC improves outcomes in OPSCC requires

1221 a different calculus, since OSrather than organ preservation is the key issue and outcome of

1222 concern.

1223 One of the earliest studies of IC using modern regimens was an Italian multi-center

1224 randomized trial,94,95 which compared 4 cycles of cisplatin-fluorouracil (PF) followed by

1225 locoregional therapy with locoregional therapy alone. Operable patients were treated with

1226 resection and adjuvant RT, and inoperable patients received definitive RT alone (65-70 Gy).

1227 Slightly over half of all patients on both arms of the study had OPSCC. When the operable and

1228 inoperable patients were pooled, there was essentially no impact of IC on OS, or local control

1229 (LC), but a significant improvement in distant relapse-free survival (DRFS) was seen. A subset NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1230 analysis of the inoperable group demonstrated significantly improved LC, DRFS and OS; in

1231 terms of toxicity, over a third of IC patients had at least one dose reduction, and grade 2 or

1232 greater toxicities were common, including hematologic, stomatitis, vomiting, renal and cardiac

1233 changes.

1234 The French cooperative group GETTEC96 performed the only available randomized

1235 study of IC strictly in patients with OPSCC, demonstrating a significant improvement in OS

1236 (median OS 5.1 vs. 3.3 years) with PF-based IC followed by local therapy (either surgery and

1237 PORT alone or definitive radiotherapy alone to 70 Gy) compared to an identical non-IC arm.

1238 These survival gains were seen despite the trial closing early due to concerns of the

1239 appropriateness of a chemotherapy-alone arm. The toxicity reported here focused exclusively on

1240 the induction component of the treatment and consisted of hematologic (26%), digestive (22%)

1241 and mucosal (13%) side effects.

1242 The updated meta-analysis of chemotherapy in head and neck cancer 31 provided

1243 additional information on the benefits of IC. This study found that in comparison to radiotherapy

1244 alone, concurrent chemoradiotherapy improved LRC (HR 0.74, p<0.0001), distant metastasis

1245 (HR 0.88, p=0.04), and overall survival (HR 0.81, 95% CI 0.78-0.86). In contrast, the analysis of

1246 31 trials that compared IC plus local therapy with local therapy alone found that IC did not

1247 improve overall survival(OS) or LRC, but it did significantly and meaningfully improve the risk

1248 of distant failure (HR 0.73, p=0.0001). However, IC did improve OS in the subset of trials using

1249 PF (HR 0.90, 95% CI 0.82-0.99), supporting the hypothesis that a reduction in distant metastases

1250 with modern chemotherapy regimens may translate into an OS benefit, despite the high

1251 competing risk of locoregional failure in this population. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1252 It is critical to note that radiotherapy alone – not concurrent chemoradiotherapy – was the

1253 standard non-surgical locoregional therapy in both of these randomized trials, as well as the non-

1254 surgical arms included in the meta-analysis. Since concurrent chemoradiotherapy is now

1255 understood to improve LRC and OS, the inadequate local therapy arm in these trials of IC may

1256 entirely explain the survival gain with neoadjuvant chemotherapy. This limitation in trial design

1257 was only recently rectified in published randomized trials, as will be described below.

1258 The insight that PF-based therapies were associated with a survival benefit produced

1259 substantial interest in improving the efficacy of IC, eventually leading to 3-drug regimens, of

1260 which TPF (taxane, cisplatin, and 5-FU) was the most frequently tested. This triplet was

1261 shown to be highly active in early trials and set the stage for two landmark randomized trials

1262 comparing TPF with PF. The TAX 324 study examined what has become the North American

1263 model for IC, termed “sequential therapy,” involving IC followed by concurrent

1264 chemoradiotherapy.97

1265 Patients were eligible if they had squamous cell carcinoma of the oropharynx, oral

1266 cavity, larynx or hypopharynx; slightly more than 50% of the patients in each arm had OPSCC.

1267 Patients received 3 cycles of IC followed by daily radiation therapy (2 Gy per fraction to 70-74

1268 Gy total) with weekly carboplatin. The addition of docetaxel significantly improved OS (3-year

1269 OS 62% vs. 48%), which was apparently mediated through an improvement in locoregional

1270 failure (38% vs. 30%), rather than distant metastasis. Hematologic toxicity was substantial, as

1271 grade 3-4 neutropenia (83% vs. 56%) and grade 3-4 thrombocytopenia (11% vs. 4%) were also

1272 each significantly increased in the docetaxel arm. Importantly, more than 20% of patients in both

1273 arms never completed chemoradiotherapy per protocol. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1274 The TAX 323 study98 had different entry criteria, only allowing unresectable patients,

1275 and a slightly altered treatment schema consisting of 4 cycles of IC, followed by radiation

1276 therapy alone (standard or accelerated fractionation). Oropharyngeal cancer represented

1277 slightly less than 50% of the total. Similar results were obtained as in TAX 324, 97,99 though

1278 the magnitude of clinical benefit was smaller (median OS 18.8 vs. 14.5 months).

1279 Approximately 10% of patients in the TPF arm did not receive radiotherapy, and an additional

1280 5% did not complete the full course. In a follow up quality of life (QoL) study, van Herpen et

1281 al.100 demonstrated trends for improved QoL with TPF both during treatment and through 6

1282 months of follow-up, although the magnitude of improvement was not deemed “clinically

1283 meaningful.”

1284 After these data established that TPF was superior to the previously accepted standard

1285 regimen, PF, the next logical question was whether IC adds substantially to concurrent

1286 chemoradiotherapy, which had been the established standard-of-care for locally advanced

1287 disease. Two small phase II studies first investigated this hypothesis. Pacagnella and colleagues94

1288 compared TPF IC followed by concurrent chemoradiotherapy (70 Gy with 2 cycles of concurrent

1289 PF) to concurrent chemoradiotherapy alone in 101 patients. Just over 50% of the patients in this

1290 study had OPSCC. The primary endpoint (rate of radiologic complete response at 6-8 weeks

1291 post-radiation) was significantly increased for the IC arm (21% vs. 50%). One limitation of this

1292 study was the non-standard concurrent regimen in this trial, consisting of two cycles of

1293 chemotherapy spaced 5 weeks apart, with conventionally fractionated RT. Another randomized

1294 phase II study, from India,101 assigned 105 eligible oropharyngeal cancer patients to either

1295 concurrent chemoradiotherapy (weekly cisplatin with 66-70 Gy) or 2-3 cycles of TPF IC NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1296 followed by the same chemoradiotherapy. There were no significant differences in clinical

1297 outcomes or toxicities between the two arms.

1298 Three recent phase III studies directly compared optimal induction and optimal

1299 concurrent treatment. The Dana-Farber Cancer Institute-based multi-institutional group

1300 (PARADIGM)102 looked at a comparison between TPF IC followed by chemoradiotherapy

1301 (weekly carboplatin or docetaxel with 70-72 Gy) versus a concurrent regimen consisting of

1302 concomitant boost radiation therapy and 2 cycles of concurrent bolus cisplatin. The University of

1303 Chicago-based DeCIDE study103 used a slightly different model: the chemoradiotherapy

1304 consisted of Taxotere, 5-FU and hydroxyurea with BID radiotherapy (1.5 Gy per fraction), every

1305 other week, versus two 21-day cycles of IC followed by the same CRT regimen. The primary

1306 end point for both trials was OS. Just over half of patients in these studies had OPSCC. Both

1307 studies closed prematurely without meeting their accrual targets, and both ultimately were

1308 underpowered, due in part to the underestimation of OS in the standard arm. Neither

1309 demonstrated significant improvements in clinical outcomes between the IC and traditional

1310 chemoradiotherapy arms, although patients treated with IC clearly experienced higher toxicity.

1311 In PARADIGM,102 more induction patients developed febrile neutropenia, and there were more

1312 serious adverse events (52 versus 22) with IC. Similarly, there were substantially more serious

1313 adverse events in the induction arm of DeCIDE 103 (47% vs. 28%), and 5 patients died due to

1314 treatment-related toxicity in the induction arm, versus none in the CRT cohort.

1315 The Spanish Head and Neck Cancer Cooperative Group104 performed a 3-armed phase 3

1316 study comparing the two induction combinations (TPF vs. PF) and concurrent

1317 chemoradiotherapy. Patients were treated with either 3 cycles of IC, followed by a concurrent

1318 regimen consisting of 70 Gy with 3 cycles of bolus cisplatin, or the same concurrent regimen NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1319 alone. Inclusion criteria specified unresectable, locally advanced head and neck cancer, and 55%

1320 of the patients had OPSCC. No significant improvement in any clinical outcome was found

1321 between any of the 3 arms of the study when analyzed on an intention-to-treat basis. However, a

1322 remarkable fraction of patients treated with IC – approximately 30% – never received

1323 radiotherapy, and an additional number could not complete the planned chemoradiotherapy.

1324 Moreover, of those induction patients who proceeded to CRT, there was significantly more grade

1325 3-4 stomatitis and dysphagia in patients previously treated with IC. An unplanned “per protocol”

1326 subset analysis was able to demonstrate improved progression-free survival (HR 0.7) for the

1327 induction arms versus the concurrent arm, without an associated improvement in OS, potentially

1328 reflecting the selection bias for induction patients tolerant of and responsive to chemotherapy.

1329 These three randomized trials found no progression-free or OS benefit with IC, yet all

1330 three studies confirmed higher rates of serious adverse events. Thus, induction chemotherapy

1331 should not be routinely implemented in patients with stage IV OPSCC.

1332

1333 Considerations in the discussion of induction chemotherapy

1334 The panel considered the potential indications for IC at length, and as expected, there was

1335 robust discussion. Several key limitations of the available studies on the question of IC for

1336 OPSCC were noted. First, the vast majority of the referenced studies enrolled locally advanced

1337 head and neck cancer patients with diverse primary subsites, including the oral cavity and

1338 pharyngolaryngeal axis. Oropharyngeal cancer representation on most of these studies ranged

1339 from only 40-60%. Clearly, prognosis varies widely across the diseases encountered in the non-

1340 OPSCC subsites, and the resulting mixed outcomes in the pooled analyses may obscure

1341 conclusions that could be drawn specifically from the data on OPSCC patients. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1342 Second, the epidemiology of OPSCC has shifted during the evolution of the IC literature,

1343 with the emergence of HPV-related disease, and none of the reported IC trials had the

1344 opportunity to factor these shifts into their statistical plans. Recent studies have shown better-

1345 than-anticipated disease control rates, likely reflecting the increased predominance of HPV-

1346 related OPSCC (representing over 80% of OPSCC patients on the DeCIDE study, for example).

1347 With such sizeable changes in expected-to-observed outcomes, the initial power calculations for

1348 both PARADIGM and DeCIDE 102,103 were incapable of accurately predicting the number of

1349 patients required to address the primary hypothesis, and the resulting statistical power suffered

1350 from the small number of events.

1351 Furthermore, there is significant uncertainty in the optimal patient population that would

1352 potentially benefit from IC. As seen in several individual trials and confirmed in the meta-

1353 analysis, IC reduces the risk of distant metastasis. Thus patients with more modest risks of

1354 distant disease are unlikely to see OS gains with systemically-active chemotherapy; the toxicity

1355 of IC and the risk of compromising successful local therapy will outweigh the improved

1356 treatment of micrometastatic disease. Only 10% and 23% of patients in PARADIGM102

1357 presented with N3 or T4 disease, respectively, and thus one could argue that the baseline risk of

1358 metastatic disease was too low to expect to see a survival advantage with IC, irrespective of the

1359 prevalence of HPV-positive disease.

1360 Subset analyses of DeCIDE103 provide some insight into the potential benefit of IC in

1361 these advanced cases. Among the 96 patients with N2c or N3 nodal disease, there was a non-

1362 significant trend (p=0.19) for improved survival with IC. Moreover, the cumulative incidence of

1363 distant metastasis in patients with LRC was significantly worse with CRT alone (absolute

1364 difference approximately 10%, p=0.043), supporting the hypothesis that improved systemic NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1365 control may translate to an OS benefit in patients at sufficiently high risk for distant metastasis,

1366 such as in those with advanced nodal presentations.

1367 Indeed, although historically the risk of locoregional failure has been substantially greater

1368 than that of metastasis, certain populations of patients do appear to be at particularly high risk for

1369 distant spread. For example, in a retrospective analysis of over 300 patients treated at the

1370 University of Chicago105 (slightly under half with OPSCC), patients with N2c or N3 disease

1371 experienced a greater than 2-fold risk of distant failure in comparison to lower volume nodal

1372 stage; nearly 40% of these patients treated with CRT-alone developed distant metastases. More

1373 recent data restricted to oropharyngeal cancer suggest similar patterns of failure. Investigators

1374 from Princess Margaret Cancer Centre performed a retrospective analysis of patients with

1375 oropharyngeal cancer treated with RT alone or CRT and described subgroups of both HPV-

1376 negative and HPV-positive disease that were at significantly higher risk for distant spread106.

1377 Individuals with HPV-negative, T3-4 or N3 cancer, experienced a distant metastasis risk of 28%,

1378 and those with HPV-positive, T4 or N3 cancer experienced a metastasis risk of 22%, which may

1379 be sufficiently high to see a survival gain with IC. On the other hand, the locoregional failure

1380 risks in the high-risk HPV-positive and negative cohorts were 18% and 38%, respectively, so

1381 that the competing risks of locoregional failure may mitigate any theoretical survival gain from

1382 micrometastasis eradication. Regardless of the therapeutic paradigm, though, it is important to

1383 recognize that there are sub-populations of even HPV-positive patients whose survival outcome

1384 need significant improvement. In fact, in a separate analysis, these investigators found that

1385 younger HPV-positive patients with T4 or N3 disease achieved a 5-year survival of only 57%

1386 with CRT, implying that substantial improvements in their treatment paradigm are critically

1387 needed.107 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1388 Unfortunately, it is unlikely that the comparative benefit of IC in a higher-risk cohort will

1389 ever be answered or even revisited through future clinical trials. Induction chemotherapy is

1390 currently incorporated into clinical trials in order to select responsive HPV-positive patients for

1391 subsequent de-escalation,108 which is a research question to be answered in the coming years.

1392 One may conclude from the MACH meta-analysis109 that IC using modern agents appears to

1393 modestly improve survival over RT alone or surgery through its effect on distant metastasis,

1394 which supplies the proof of principle that treating micrometastatic disease may improve oOS.

1395 Nevertheless, the high-level data do not support this conclusion when the standard comparator is

1396 chemoradiotherapy. Three phase III randomized trials comparing IC and chemoradiation therapy

1397 (CRT) with CRT alone were all clearly negative; IC was associated with more toxicity without a

1398 proven survival benefit. While each study has recognized limitations, in light of this high-level

1399 evidence, the panel chose not to specify a high-risk cohort of patients for whom IC may be

1400 considered. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1401 Key Question 4: What are the appropriate dose, fractionation, and volume regimens with 1402 and without systemic therapy in the treatment of oropharyngeal squamous cell carcinoma 1403 (OPSCC)? 1404 1405 In the scenario of definitive non-surgical therapy? 1406 1407  Statement 4A. A dose of 70 Gy over 7 weeks should be delivered to gross primary and nodal 1408 disease in patients with stage III-IV oropharyngeal squamous cell carcinoma selected to receive 1409 standard, once-daily definitive radiotherapy. 1410 o Recommendation strength: Strong 1411 o Quality of evidence: Moderate 1412 1413  Statement 4B. The biologically equivalent dose of approximately 50 Gy in 2 Gy fractions or 1414 slightly higher should be delivered electively to clinically- and radiographically-negative regions 1415 at-risk for microscopic spread of tumor. 1416 o Recommendation strength: Strong 1417 o Quality of evidence: Low 1418

1419  Statement 4C. Altered fractionation should be used in patients with stage IVA-B oropharyngeal 1420 squamous cell carcinoma treated with definitive radiotherapy who are not receiving concurrent 1421 systemic therapy. 1422 o Recommendation strength: Strong 1423 o Quality of evidence: High 1424

1425  Statement 4D. Either accelerated radiotherapy or hyperfractionated radiotherapy may be used in 1426 patients with oropharyngeal squamous cell carcinoma treated with altered fractionation definitive 1427 radiotherapy after a careful discussion of patient preferences and the limited evidence supporting 1428 one regimen over the other. 1429 o Recommendation strength: Conditional 1430 o Quality of evidence: High 1431

1432  Statement 4E. Either standard, once-daily radiotherapy or accelerated fractionation may be used 1433 when treating oropharyngeal squamous cell carcinoma with concurrent systemic therapy after a 1434 careful discussion of patient preferences and the risks and benefits of both approaches. 1435 o Recommendation strength: Conditional 1436 o Quality of evidence: High 1437

1438  Statement 4F. Altered fractionation should be used in patients with T3 N0-1 oropharyngeal 1439 squamous cell carcinoma treated with definitive radiotherapy who do not receive concurrent 1440 systemic therapy. 1441 o Recommendation strength: Strong NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1442 o Quality of evidence: Moderate 1443

1444  Statement 4G: Altered fractionation may be used in patients with T1-2 N1 or T2 N0 1445 oropharyngeal squamous cell carcinoma treated with definitive radiotherapy alone who are 1446 considered at particularly significant risk of locoregional recurrence, after a careful discussion of 1447 patient preferences and the limited evidence supporting its use in this scenario. 1448 o Recommendation strength: Conditional 1449 o Quality of evidence: Low 1450

1451 Narrative 1452 1453 A relatively large body of literature has reported outcomes for OPSCC patients in the

1454 setting of randomized clinical trials. Important caveats in interpreting this literature, much of

1455 which directly tested dose and fractionation, must include the recognition that most of these

1456 studies included patients with advanced squamous carcinomas arising in sites other than the

1457 oropharynx, including oral cavity, larynx and hypopharynx. Although these studies described

1458 the relative proportions of these primary sites, their conclusions were often based on a pooled

1459 analysis. Human papillomavirus-associated OPSCC has been increasing in prevalence and

1460 proven to have a much more favorable prognosis than non-HPV related OPSCC.2 The proportion

1461 of OPSCC patients with HPV-related cancers in most landmark studies of fractionation remains

1462 unknown and is likely to be significantly lower than in contemporary patient populations, raising

1463 the question of the generalizability of older studies to modern cohorts. The panel recognizes

1464 there is important ongoing work examining the different outcomes between and within HPV

1465 positive and negative OPSCC in order to define prognostic subgroups. 89,107,110 This research has

1466 led to the development and implementation of prospective clinical trials examining treatment de-

1467 intensification strategies for favorable prognostic subgroups.93 In the future, these studies may

1468 result in significant changes in the recommended radiotherapy doses for the management of

1469 favorable-risk disease. To date, the results of radiotherapy de-intensification approaches are not NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1470 yet available and there are no prospective randomized data suggesting that HPV-positive patients

1471 gain any less benefit from altered fractionation regimens. Thus any recommendation made for

1472 oropharyngeal cancer patients within these guidelines apply equally to HPV-positive and

1473 negative individuals.

1474 Primary radiation therapy has long been a standard-of-care for the management of

1475 OPSCC, and acceptable doses to gross disease and elective regions is the first consideration in

1476 the discussion of dose and fractionation regimens. There have been no randomized trials

1477 comparing total radiotherapy dose delivered once-daily for OPSCC, and a variety of dose-

1478 fractionation regimens have evolved over the years. In the United Kingdom, the treatment

1479 approach classically delivered 50-55 Gy in 15-20 fractions over 3-4 weeks, while in Europe and

1480 most of North America, more extended fractionation schemes were employed, delivering 66-72

1481 Gy in 32-40 fractions over 6.5 to 8 weeks.111 Though retrospective in nature, comparisons

1482 between these regimens suggested that shorter courses delivering less total dose with larger

1483 fraction sizes resulted in lower rates of control for advanced disease with increased late

1484 effects.111 This observation, combined with the enhanced toxicity in combining concurrent

1485 chemotherapy with fraction sizes greater than 2 Gy, resulted in the emergence of a total dose of

1486 68-70 Gy delivered in 33-35 fractions over 6.5 to 7 weeks as the most common once-daily

1487 fractionation regimen.

1488 The delivery of 70 Gy to gross disease has been selected as the standard treatment

1489 approach for many practice-defining randomized trials of both fractionation and concurrent

1490 chemotherapy, including RTOG 9003, 112 EORTC 22851,113 the US Intergroup study,9 and

1491 RTOG 0129.37 Thus, when choosing a conventionally fractionated, once-daily radiotherapy NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1492 regimen for patients with stage III-IV OPSCC, the panel considers 70 Gy over 7 weeks to be the

1493 reference standard.

1494

1495 Standard Prophylactic Dose to Elective Volumes at Risk of Microscopic Disease

1496 The radiotherapeutic management of all stages of OPSCC requires treatment to be

1497 delivered electively beyond areas of apparent gross disease. Elective volumes include margins

1498 around gross disease as well as nodal levels at risk of harboring microscopic nodal metastasis.

1499 The location and extent of these elective volumes will be dictated by specific characteristics of

1500 the primary tumor (size and anatomic extent) and grossly involved nodes. The anatomic

1501 delineation of node levels at risk should be consistent with published guidelines.114 The dose and

1502 volume of elective irradiation is often the driver of dose to critical organs-at-risk, including

1503 salivary glands and dysphagia-associated structures such as the larynx and pharyngeal

1504 constrictors, emphasizing the importance of considering the dose necessary to sterilize

1505 microscopic disease.

1506 Although the evidence is nearly entirely retrospective, prophylactic radiotherapy at 2 Gy

1507 per day to a total dose of 50 Gy has been widely regarded as the standard dose for the elective

1508 treatment of areas at risk for subclinical or microscopic involvement. This dose was first

1509 suggested by Fletcher based on observations of control rates in head and neck and breast cancer

1510 patients.115 Withers et al.111 expanded these observations to a wider range of patients, concluding

1511 a dose of 50 Gy in 2 Gy fractions was necessary to achieve a 90% reduction in the incidence of

1512 recurrence within the electively irradiated tissues. Although this dose has been utilized in many

1513 clinical trials, many do not specifically report rates of nodal failure, or distinguish between in-

1514 field versus elective failure. Nevertheless, some insights may be gleaned from these trials. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1515 For example, the DAHANCA 6/7 trials116 limited dose to a maximum 50 Gy in 2 Gy

1516 fractions to elective neck regions and observed isolated neck failure (elective and therapeutic

1517 regions) in only 69/1476 (4.7%) patients. Goshal et al.117 delivered 44-45 Gy to the elective neck

1518 regions and reported isolated regional failures in 37/285 (12.9%) patients. The ARTSCAN

1519 randomized trial7 comparing conventionally-fractionated radiotherapy with treatment over 4.5

1520 weeks118 treated the elective nodal regions to 46 Gy and reported isolated nodal failure in 34/733

1521 (4.6%) patients. These three randomized studies have cumulatively reported isolated nodal

1522 failures in 140/2494 (5.6%) of patients receiving the equivalent of 50 Gy in 2 Gy fractions or less

1523 to elective nodal regions. These recurrences would have included failures occurring in pre-

1524 existing nodes, and therefore the failure rates in the elective volumes could be assumed to be less

1525 than this.

1526 Retrospective analyses that have specifically identified electively irradiated neck failures

1527 include Lambrecht et al.119, observing only 2 recurrences out of 368 (<1%) in elective neck

1528 regions receiving 44-50 Gy. More recently, investigators from MD Anderson reported a risk of

1529 elective regional failure with primary control of only 4/776 (< 1%), and many of these controlled

1530 patients were treated with a low anterior neck field to 50 Gy.120 With the emergence of routine

1531 IMRT planning and simultaneous integrated boost/“dose-painting” approaches, radiotherapy

1532 plans are now often delivered in a single phase, rather than the conventional, serial cone-down

1533 approach using 2 Gy per fraction. Thus, a plan delivering 70 Gy in 35 fractions to sites of gross

1534 disease must simultaneously deliver 56 Gy in 35 fractions of 1.6 Gy to areas of subclinical or

1535 microscopic involvement to achieve the radiobiologic equivalent of 50 Gy in 25 fractions. Using

1536 this approach, Duprez et al.121 have a reported regional relapse risk in 3/286 (1%) patients within

1537 the elective neck volumes receiving 56 Gy. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1538 The panel therefore believes there is sufficient evidence to strongly recommend the

1539 ongoing use of a radiobiologic equivalent of 50 Gy delivered in 2 Gy fractions to electively treat

1540 areas at risk of subclinical or microscopic involvement. The high rates of neck control observed

1541 with this approach raise the possibility of elective dose reduction, which would be expected to

1542 reduce dose to the salivary tissue and pharyngeal constrictors. In fact, if one considers that

1543 concurrent chemotherapy increases the biologically effective dose of any given dose of

1544 irradiation,122 treatment with concurrent chemoradiotherapy may afford the opportunity for dose

1545 reduction. At this time, however, there is insufficient data to recommend elective doses reduction

1546 below the equivalent of 50 Gy in 2 Gy fractions, although this is an area of active research.123

1547

1548 Altered fractionation radiotherapy without concurrent chemotherapy

1549 Poor historical outcomes for patients with stage III-IV OPSCC treated with

1550 conventionally-fractionated radiotherapy alone10 led to efforts to improve oncologic results by

1551 intensifying treatment. This intensification utilized two distinct approaches. One approach was

1552 the addition of cytotoxic systemic therapy delivered concurrently with radiotherapy, and the

1553 other was to alter the radiotherapy fractionation schemes from standard, once-daily

1554 administration.

1555 The concept of altered fractionation (AF) refers to fractionation regimens that differ from

1556 standard, once-daily treatment. Altered fractionation-regimens are based on radiobiologic

1557 principles that suggest giving a higher total dose and/or giving it over a shorter total treatment

1558 time (to counter tumor repopulation) will result in improved tumor control. Similar principles

1559 suggest delivering this dose with reduced fraction sizes (< 2 Gy) should result in reduced late

1560 tissue effects permitting delivery of a higher total dose. Altered fractionation regimens may be NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1561 classified as either accelerated, delivering a similar or reduced total dose over a shorter total

1562 time, or hyperfractionated, delivering an increased total dose over a similar time with multiple

1563 small daily fractions. Many AF regimes represent a hybrid of these approaches.

1564 Altered fractionation regimens for the treatment of head and neck squamous carcinomas,

1565 including OPSCC, have been compared to standard radiotherapy in a number of randomized

1566 phase III clinical studies (Table 7). The majority of these studies have demonstrated that AF

1567 regimens improve LRC by 8-20% compared to once-daily radiotherapy alone, but the increased

1568 control in the primary site and neck generally did not translate into an OS advantage. In addition,

1569 this disease control improvement frequently came at the cost of increased but manageable acute

1570 toxicity, with a variable impact on the risk for late toxicities.

1571 Accelerated regimens with or without a total dose reduction have been the AF approach

1572 most often compared with standard radiotherapy. In the landmark RTOG 9003 trial124 reported

1573 results on 1073 patient randomized to one of three AF arms compared to standard radiotherapy

1574 (70 Gy in 2 Gy fractions over 7 weeks). Two of these arms represented accelerated treatment

1575 delivered over 6 weeks. One arm delivered twice-daily (BID) treatment throughout but had a

1576 two-week break, to achieve a total dose of 67.2 Gy, while the other arm employed a scheme

1577 without a break, delivering once daily 1.8 Gy for six weeks with a second 1.5 Gy fraction as a

1578 concomitant boost to the primary on the last 12 treatment days for a total dose of 72 Gy. The

1579 concomitant boost arm demonstrated an absolute improvement of 9.5% and 7.6% in two year

1580 LRC and disease free survival (DFS) (p=0.05 and 0.054 respectively) which came at a cost of an

1581 absolute increase of 24% in grade 3 or higher acute toxicity (multiple endpoints). The split

1582 course arm demonstrated no significant improvement in LRC or DFS and neither arm improved

1583 OS. The recent final analysis of RTOG 9003112 demonstrated that the initially improved LRC NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1584 with accelerated radiotherapy lost its statistical significance with longer follow-up, and overall

1585 survival (OS) was still unaffected by this regimen. Although the continuous acceleration arm did

1586 improve disease-free survival (HR 0.82, p=0.05), there were non-significant trends for worse late

1587 toxicity in this arm. Notably, patients who were disease-free at 1-year were significantly more

1588 likely to be feeding tube dependent with continuous accelerated radiotherapy (1-year probability

1589 13.4% vs. 4.6%, p=0.02).

1590 The DAHANCA group125 randomized 1485 patients with squamous cell carcinoma of the

1591 head and neck (larynx, pharynx and oral cavity) to receive standard radiotherapy (66-68 Gy in

1592 2Gy fractions) delivered 5 days per week versus the same dose delivered in an accelerated

1593 fashion with 6 fractions per week. They observed a statistically significant 10% benefit in 5 year

1594 LRC (70% vs 60%) and a 13% benefit in DFS (73% vs 66%) for the accelerated regimen. No

1595 benefit in OS was observed. Patients treated with the accelerated regimen experienced a 20%

1596 higher risk of acute confluent mucositis, and the duration of this acute mucositis was longer, but

1597 there was no significant difference in longer-term mucosal toxicities. A similar LRC benefit

1598 (absolute 12%) was observed with this regimen when the study was later replicated in 908

1599 patients with more advanced disease, 126 with a parallel increase in acute but not late mucosal

1600 toxicity. In an interesting subgroup analysis of the original DAHANCA study,127 the LRC

1601 benefit of radiotherapy acceleration was maintained in p16 positive (mostly oropharynx) patients

1602 and was perhaps even greater than that observed in p16 negative (mostly non-oropharynx)

1603 patients.127

1604 The EORTC 22851 trial113 randomized 500 patients to receive 72 Gy delivered in twice-

1605 daily fractionation over 5 weeks versus standard fractionation (70 Gy in 2 Gy fractions over 7

1606 weeks). Although this study demonstrated a 13% improvement in LRC at 5 years the actuarial NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1607 rates of grade 3-4 late toxicity in the accelerated arm was 37% at 3 years versus 15% for the

1608 control arm. This finding, coupled with increased rates of acute toxicity and the lack of an

1609 overall or cancer-specific survival benefit, led the authors to conclude that a less toxic scheme

1610 should be used and emphasizes that there is a limit to the degree of safe acceleration.

1611 It should be noted not all studies have demonstrated significant benefits to accelerated

1612 radiotherapy. The TROG group128 reported outcomes for 343 patients (67% OPSCC) randomized

1613 to accelerated radiotherapy (59.4 Gy in 1.8 Gy bid over 24 days) versus standard 70 Gy in 2 Gy

1614 fractions over 7 weeks). In this modest sized study they observed a nonsignificant trend towards

1615 improved disease free and disease specific survival at 5 years (41% and 46% vs 35% and 40%

1616 respectively, p=0.323 and 0.398) for the accelerated arm. The multicenter Swedish ARTSCAN

1617 study7 randomized 750 patients (48% OPSCC) to standard radiotherapy (68 Gy in 2 Gy fractions

1618 over 7 weeks) versus accelerated radiotherapy (1.1 Gy + 2 Gy/day over 4.5 weeks to 68 Gy). No

1619 significant difference was observed in OS or loco-regional control between the two arms.

1620 The use of pure hyperfractionation has been demonstrated to improve outcomes. The

1621 EORTC 22791 study 129 reported the results of 325 patients with T2-T3, N0-N1 OPSCC

1622 randomized to receive 70 Gy in 35 daily fractions over 7 weeks or hyperfractionated treatment

1623 delivering 1.15 Gy BID to a total dose of 80.5 Gy over 7 weeks. A significant improvement was

1624 seen in 5 year rates of local control at the primary site (59% vs 40% p=0.02) in the

1625 hyperfractionation arm, although improvement in OS was only borderline significant (p=0.08).

1626 There was a significant increase in acute mucositis with hyperfractionation (18% absolute

1627 increase in grade 3 mucositis), but there was no difference in any late effect between the two

1628 arms. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1629 One arm of RTOG 9003112 utilized a hyperfractionated approach, delivering 68 BID

1630 fractions of 1.2 Gy to a total dose of 81.6 Gy. The final analysis observed not only an improved

1631 and sustained rate of LRC (relative reduction in locoregional failure of 21%) but also an OS

1632 advantage when patients were censored at 5 years (HR 0.81, p=0.05). This benefit was lost with

1633 longer follow-up, which may relate to second primary cancers or death from other causes

1634 obscuring a disease control benefit from hyperfractionation. Treatment with hyperfractionation

1635 was associated with significantly more acute toxicity: 54% of patients in this cohort experienced

1636 at least one grade 3 toxicity, versus 35% of patients treated once-daily. However, there was no

1637 significant increase in late toxicities with hyperfractionation, although disease-free patients were

1638 more likely to be gastrostomy-dependent in comparison to disease-free individuals treated with

1639 standard fractionation (17.3% vs. 4.6%, p<0.01).

1640 These results were largely echoed by a large meta-analysis (MARCH) based on

1641 individual patient data, which examined outcomes for 6,515 patients enrolled in 15 randomized

1642 trials comparing standard to AF radiation therapy (accelerated or hyperfractionated) for

1643 advanced head and neck cancers.130,131 Oropharyngeal cancer comprised approximately 44% of

1644 all patients. This meta-analysis reported absolute reductions in the 5-year risk of locoregional

1645 recurrence of 9.4% and 7.3% in patients treated with hyperfractionation or accelerated

1646 fractionation (without dose-reduction) regimens, respectively. Most of this benefit was due to

1647 reduced recurrence at the primary site. The absolute probability of OS at 5 years was

1648 significantly improved by 3.4% with altered fractionation. There was no significant difference in

1649 this benefit for tumors arising in the oropharynx compared to hypopharynx larynx or oral cavity,

1650 but the survival advantage with AF was lost in patients older than 71 years. The survival benefit

1651 was significantly greater with hyperfractionation (8.2% vs. 2% at 5 years for patients receiving NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1652 hyperfractionation and accelerated radiotherapy without a dose-reduction, respectively).

1653 Although one may conclude that hyperfractionation may be the superior approach, the trials of

1654 acceleration and hyperfractionation evaluated fundamentally different patient populations;

1655 accelerated radiotherapy was more often employed in larynx cancer and patients with earlier-

1656 stage disease, such that improved LRC was less likely to translate into an OS benefit.

1657 Although individual trials were unable to confirm an OS advantage with altered

1658 fractionation, the majority of studies have demonstrated consistent and meaningful

1659 improvements in LRC and trends towards reductions in overall mortality. The influential

1660 MARCH meta-analysis provided further evidence for the superiority of AF in patients with

1661 locally advanced head and neck cancer. Given the adverse clinical consequences of locoregional

1662 failure and the potential for a survival gain with its use the panel strongly recommends AF

1663 radiation therapy should be used for patients with stage IVA-B OPSCC managed with primary

1664 radiation therapy without concurrent systemic therapy. In comparison to treatment with

1665 concurrent chemotherapy, in which late toxicities were comparable to patients treated with

1666 radiotherapy alone, there were somewhat consistent signals that AF modestly worsened late

1667 toxicities. One particularly important late toxicity is swallowing dysfunction, which in the worst

1668 case can lead to the long-term requirement for enteral feeding. However, modern radiotherapy

1669 techniques may minimize these risks, as a large multi-institution pooled analysis of 2,315

1670 OPSCC patients managed with IMRT concluded that there was no difference in one year rates of

1671 gastrostomy tube dependence between patients managed with standard versus accelerated

1672 radiotherapy.132 Nevertheless, in patients treated with altered fractionation, particular attention

1673 must be paid to the dysphagia organs-at-risk and early and persistent swallowing rehabilitation. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1674 It is unclear whether hyperfractionation or acceleration is a better treatment paradigm.

1675 The only study comparing the two concepts was RTOG 9003, 112 which slightly favored

1676 hyperfractionation, but this study was not designed to compare the experimental arms; the

1677 MARCH meta-analysis also suggested an improvement with hyperfractionation, but as detailed

1678 above, there are significant flaws with this indirect comparison. Nevertheless, hyperfractionation

1679 regimens delivered without concurrent chemotherapy are not in wide-spread use for the

1680 management of OPSCC. This pattern is in part due to the increased resources required to deliver

1681 these treatments, coupled with the emergence of concurrent chemotherapy approaches that

1682 mitigated the benefit of altered fractionation. Accelerated radiotherapy, particularly using the

1683 DAHANCA approach, 125 is significantly more convenient than BID treatments over 7 weeks,

1684 and provides a more practical platform for both clinical trials and routine practice. That said, at

1685 the present time, the panel concludes the available evidence supports the use of either form of

1686 AF for the management of stage IVA-B OPSCC in patients not receiving systemic therapy, and

1687 the relative strengths and weaknesses of the strategies should be carefully discussed with

1688 patients.

1689

1690 Altered fractionation radiotherapy for patients with stage III oropharyngeal cancer

1691 Deriving recommendations for patients with stage III OPSCC is more challenging; stage

1692 III disease may present with a wide range of primary tumor volumes (T1-T3, N1). The majority

1693 of patients in randomized studies of AF presented with stage IV disease, and thus the advantages

1694 of this approach are less clear with lower-bulk disease. Although most of the studies included

1695 patients with stage III disease, the proportion of patients with this stage was typically less than

1696 30%, and the outcomes were rarely stratified by T and N or AJCC stage. One exception is NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1697 EORTC 22791, 129 which randomized patients with T1-3 N0-1 oropharyngeal cancer to

1698 hyperfractionation or conventional daily radiotherapy. Although the study showed an overall

1699 benefit in LRC, this gain was strictly in patients with T3 tumors. Similarly, in EORTC 22851, 113

1700 which randomized stage III and IV patients (all disease sites) to accelerated or conventional

1701 radiotherapy, the LRC advantage with acceleration was only seen in N2-3 or T4 patients. Both

1702 studies emphasize the concept that relative benefit of AF is greatest with larger volume disease.

1703 As detailed in the narrative of KQ1, single institution retrospective reports have also

1704 described significant differences in local control as a function of T-stage or tumor volume. The

1705 preponderance of data strongly suggest that patients with larger volume disease need additional

1706 local therapy beyond conventional radiotherapy alone. Since concurrent systemic therapy is most

1707 consistently associated with superior LRC and OS, patients with T3 N0-1 disease who are

1708 candidates for concomitant systemic therapy should receive it, as detailed in KQ1. For

1709 individuals who would not tolerate such treatment, the panel strongly recommends AF alone.

1710 The data guiding radiotherapy fractionation recommendations for stage II and III (T1-2

1711 N1) are far less compelling, and there is significant volume heterogeneity even in this cohort. For

1712 individuals with T2 N0 and T1-2 N1 stage III disease considered at particularly significant risk

1713 for primary and/or nodal recurrence, AF may be employed, but the clinician must weigh the

1714 patient’s estimated risk of locoregional failure with conventional treatment against the

1715 recognized toxicities of altered fractionation. The relative lack of data guiding this

1716 recommendation should be carefully discussed with patients.

1717 The use of modestly accelerated radiotherapy alone was shown to result in excellent

1718 outcomes in RTOG 0022, 133 which treated 67 patients with T1-2 N0-1 OPSCC with 66 Gy in 30

1719 fractions. Most (75%) of these patients had T2 disease. There were 7 locoregional failures, all in NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1720 current or former smokers (7/36, accounting for crude 19% risk of failure in this cohort),

1721 although 2 of these recurrences were in patients with major dosimetric violations. Only six

1722 patients had new or continuing grade 3 or 4 toxicity after 15 months. In total, these prospective

1723 data suggest that this accelerated regimen is a reasonable approach for selected patients.

1724

1725 Altered fractionation radiotherapy with concurrent chemotherapy

1726 The therapeutic alternative to AF for locally advanced head and neck cancer is concurrent

1727 chemoradiotherapy, the benefits of which were discussed in a prior narrative (KQ1). A large

1728 meta-analysis has reported improved survival and LRC with the addition of concurrent

1729 chemotherapy to radiotherapy,31 a benefit maintained using either standard or altered

1730 fractionation. The emergence of data demonstrating improved outcomes with either AF approach

1731 or the addition of chemotherapy to standard, once-daily radiotherapy naturally led to studies

1732 examining the addition of chemotherapy to AF radiation therapy. Initially the addition of

1733 chemotherapy to AF radiotherapy was compared to AF radiotherapy alone. Several randomized

1734 studies have demonstrated improved LRC and survival with such a strategy.17,134-136 For

1735 example, Brizel et al.134 achieved this with similar levels of toxicity, but this required a reduced

1736 radiotherapy dose and treatment break. Bensadoun et al. maintained the same dose of

1737 radiotherapy in both arms (75.6-80.4 Gy) while delivering full dose cisplatin (100 mg/m2) and 5-

1738 fuorouracil infusion. Although the reported toxicity data was similar between the two arms they

1739 described a 13.6% risk of “early death” in the combined treatment arm, emphasizing the

1740 potential risk in this treatment strategy.

1741 The next logical question was whether accelerated chemoradiotherapy was superior to

1742 conventionally-fractionated chemoradiotherapy, a hypothesis that was examined in two large NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1743 phase III studies. RTOG 012932 randomized 743 patients to standard fractionation (70 Gy in 35

1744 fractions over 7 weeks) delivered with 3 cycles of concurrent cisplatin or accelerated

1745 fractionation (70 Gy in 35 fractions over 6 weeks) delivered with 2 cycles of concurrent

1746 cisplatin.37 The mature results of this study (median follow-up 7.9 years) indicated no difference

1747 in progression free survival or OS with similar acute and late toxicities. Groupe d'Oncologie

1748 Radiothérapie Tête et Cou (GORTEC) 99-0223 was a three arm study randomizing 840 patients

1749 to either standard chemoradiotherapy (70 Gy in 7 weeks plus 3 cycles carboplatin-fluorouracil)

1750 versus accelerated chemoradiotherapy (70 Gy in 6 weeks plus 2 cycles carboplatin-fluorouracil)

1751 versus very accelerated radiotherapy alone (64.8 Gy in 1.8 Gy bid over 3.5 weeks).23 No

1752 difference was observed between the chemoradiotherapy and accelerated chemoradiotherapy

1753 arms in 3 year rates of locoregional failure, distant metastasis, progression-free survival, OS or

1754 acute or late toxicity, with both chemotherapy arms superior to the very accelerated radiotherapy

1755 approach in all aspects.

1756 The addition of chemotherapy to aggressive AF approaches must be considered with

1757 caution given the potential for increased acute and late effects associated with this approach. For

1758 this reason, both RTOG 012932 and GORTEC 99-02137 delivered less chemotherapy in the AF

1759 arm than the standard fractionation arm. This strategy seems to have been successful in

1760 maintaining equivalent toxicities between the arms without compromising tumor control. In

1761 RTOG 0129, there were no statistically significant differences in acute or late toxicities, although

1762 there was a trend for increased long-term gastrostomy dependence in the accelerated arm (13.2%

1763 vs. 5.9% at 5 years, p=0.08). On the other hand, more patients received the correct numbers of

1764 cisplatin cycles in the accelerated arm (88% vs. 69%), owing to the challenge of managing the

1765 toxicities with bolus cisplatin. Similarly, in GORTEC 9902, 137 92% of patients in the NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1766 accelerated chemoradiotherapy arm received the correct number of cycles, in comparison to only

1767 71% of individuals in the conventional arm.

1768 Since two large phase III trials have shown no significant difference in oncologic

1769 outcomes or toxicities between conventional and accelerated fractionation (AccF) with three and

1770 two cycles of bolus cisplatin or carboplatin/fluorouracil, respectively, the panel concludes either

1771 fractionation approach may be considered for patients with stage IVA-B oropharyngeal cancer

1772 treated with these chemotherapy regimens. The relative risks and benefits of more chemotherapy

1773 versus more intensive radiation treatment should be carefully discussed with the patient, and the

1774 decision should be made on a patient-specific basis, in concert with their preferences and values.

1775

1776 In the scenario of adjuvant post-operative radiotherapy? 1777 1778  Statement 4H. Adjuvant post-operative radiotherapy should be delivered to regions of 1779 microscopically-positive primary site surgical margins and extracapsular nodal extension at 2 1780 Gy/fraction once daily to a total dose between 60 and 66 Gy. 1781 o Recommendation strength: Strong 1782 o Quality of evidence: Moderate 1783 1784  Statement 4I. Adjuvant post-operative radiotherapy delivered without concurrent systemic 1785 therapy should treat regions of microscopically-positive primary site surgical margins and 1786 extracapsular nodal extension at 2 Gy/fraction once daily to a total dose of 66 Gy, although there 1787 are limited data guiding this recommendation. 1788 o Recommendation strength: Conditional 1789 o Quality of evidence: Weak 1790 1791  Statement 4J. Adjuvant post-operative radiotherapy should be delivered to the tumor bed and 1792 involved, dissected lymph node regions at 2 Gy/fraction once daily to a total dose of 60 Gy in the 1793 absence of primary site positive margins and extracapsular nodal extension. 1794 o Recommendation strength: Strong 1795 o Quality of evidence: Moderate 1796 1797

1798 Narrative 1799 1800 A dearth of prospective randomized data exist addressing dose-fractionation issues in the

1801 post-operative, high-risk setting. Peters et al.138 randomized patients receiving PORT alone (18% NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1802 oropharynx) to one of four dose levels depending on the type and number of adverse pathologic

1803 risk factors. Patients deemed lower-risk were first randomized to 52.2-54 Gy versus 63 Gy, but

1804 an interim analysis showed this lower dose was inadequate, and the study was completed

1805 comparing 57.6 Gy in 32 fractions with 63 Gy in 35 fractions in this lower-risk population. The

1806 higher-risk patients were randomized to 63 Gy in 35 fractions versus 68.4 Gy in 38 fractions.

1807 Because the risk stratification was a function of both the type and number of adverse

1808 characteristics, patients with ECE and positive margins (less than 5 mm) were included in all

1809 dose levels. The authors found no difference in 2 year recurrence rates for all higher-risk

1810 patients. However, in a post hoc subset analysis, patients with ECE experienced superior 2-year

1811 LRC with 63 Gy and 68.4 Gy versus 57.6 Gy, but no difference between the higher dose levels

1812 (74% and 72% versus 52%, respectively, p=0.03). Consistent with this finding, in a retrospective

1813 analysis of 102 oral cavity/oropharynx patients treated with differing doses of PORT, Zelefsky et

1814 al.139 showed that non-oral tongue patients with positive or close margins receiving a minimum

1815 of 60 Gy had significantly improved LRC versus lower doses (92% vs. 44% at 7 years,

1816 p=0.0007).

1817 Ang et al.4 performed a randomized assessment of AccF vs conventional fractionation

1818 (CF) in high-risk post-operative patients (63 Gy/5 weeks vs 63 Gy/7weeks). There were trends

1819 towards improved LRC (p=0.11) and OS (p=0.08) with acceleration, the latter result partially

1820 owing to a 33 % incidence of distant failure. Accelerated fractionation increased grade 3 mucosal

1821 toxicity (62% vs 36%). Sanguineti et al.140 also compared AccF vs CF (64 Gy/5 weeks vs 60

1822 Gy/6 weeks). There were no differences with respect to local-regional control, distant failure, nor

1823 overall survival. Grade 3 mucosal toxicity was significantly increased, however (53% vs 27%).

1824 Suwinski et al.141 also compared post-operative AccF against CF. Approximately 40% of the NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1825 patients had oral cavity/oropharynx primaries, and AccF resulted in significantly better local-

1826 regional control (74% vs 53%) in this subset. There was no difference in OS with accelerated

1827 treatment. Again the cost was a significantly higher incidence of grade 3 mucosal toxicity with

1828 AF (60% vs 33%).

1829 Caution should be exercised in attempting to apply these data to the clinical setting. The

1830 vast majority of patients with high-risk pathologic features now receive concurrent

1831 chemoradiotherapy rather than RT alone. Additionally, a minority of the patients treated on the

1832 aforementioned trials had OPC primary tumors (17-40%). These trials enrolled most of their

1833 subjects during the 1980s and 1990s, and OPC itself has evolved since then from a smoking

1834 related disease to one that is predominantly caused by HPV infection. Most investigational

1835 efforts for the latter are focused on therapeutic de-escalation given its more favorable prognosis.

1836 Changes in RT delivery techniques must also be kept in mind when interpreting these

1837 data. Most of the patients were treated with Co-60 or 4 MV photons using 2-D parallel opposed

1838 fields prescribed at mid-plane. It is likely that the prescribed daily dose of 1.8 Gy was 10-15%

1839 higher (2.0-2.1 Gy) in the neck as a consequence of dose inhomogeneity. In this respect, past is

1840 prologue for patients treated with contemporary IMRT dose painting techniques that commonly

1841 deliver daily doses of 2.2 Gy to high risk regions. Lukens et al. have shown a significantly

1842 increased risk of soft tissue necrosis when daily fractions of this size are used for PORT in

1843 OPSCC.142

1844 In summary, the small volume of randomized data suggests improved LRC with

1845 treatment doses beyond 57.6 Gy for patients with positive margins and/or ECE. The three

1846 randomized trials of accelerated radiotherapy provided mixed results on LRC, but they were

1847 consistent in showing no OS benefit and markedly increased mucosal toxicity with intensified NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1848 treatment. Therefore the panel strongly recommends that patients receiving adjuvant PORT for

1849 positive margins and/or ECE are treated using daily fraction sizes of 2 Gy to a total dose between

1850 60 and 66 Gy. This dose (60 Gy) is comparable to the 63 Gy in 7 weeks from Peters et al. while

1851 reducing the total package time (i.e. time from surgery to the end of radiotherapy).

1852 There are insufficient data to properly answer whether dose-escalation beyond 60 Gy

1853 provides additional clinical benefit in patients with high-risk pathology. As detailed above, the

1854 Peters trial138 showed no further benefit to 68.4 Gy in this population. That said, the landmark

1855 RTOG 950145,46 and EORTC 2293144 trials used post-operative doses of 60-66 Gy (RTOG) and

1856 66 Gy (EORTC) to regions of highest risk. Without concurrent cisplatin, patients in the RTOG

1857 trial with positive margins and/or ECE had a 10-year locoregional recurrence risk of 33%, and

1858 individuals in the EORTC study treated with radiotherapy alone had a 5-year locoregional

1859 recurrence risk of 31%. Given these poor LRC outcomes with radiotherapy alone, the expert

1860 panel is more concerned with the complications of locoregional failure than mucositis. The panel

1861 conditionally recommends that patients with positive margins and/or ECE receiving PORT alone

1862 receive a total dose to these regions of 66 Gy in 2 Gy fractions, although the lack of high-quality

1863 data guiding this statement must be acknowledged.

1864 The evidence guiding dose and fractionation regimens in the setting of negative margins

1865 and no ECE is similarly scant. The Peters study initially randomized low risk patients to 52.2-54

1866 Gy vs 63 Gy at 1.8 Gy per fraction. Because of a higher incidence of local failure in the low dose

1867 cohort (p=0.02), the total dose subsequently was increased to 57.6 Gy. No advantages were

1868 observed with 63 Gy vs 57.6 Gy, but grade 3-4 toxicity was higher (7.7% vs 3.3%) with the

1869 higher dose. Low risk was defined as having not more than one of the following factors: oral

1870 cavity primary, mucosal margins close or positive, nerve invasion, ≥2 positive lymph nodes, NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1871 largest node > 3 cm, treatment delay greater than 6 weeks, and performance status > or = 2.

1872 Ang’s trial4 used the same clinical and pathologic criteria with low risk defined as having none

1873 of them and intermediate risk as having only one exclusive of ECE. Low-risk patients were

1874 observed; all intermediate risk patients (n=31) received 57.6 Gy and had the same local-regional

1875 control as the low risk patients.

1876 The panel considers that the dose of 57.6 Gy in 1.8 Gy fractions is approximately

1877 equivalent to 56 Gy in 2 Gy fractions. The daily dose is higher and total treatment time is shorter

1878 with 2 Gy fractions, arguing in favor of 56 Gy as the reference dose. However, because there are

1879 few data showing successful long-term control outcomes with 56 Gy, and as stated above, the

1880 true dose to involved stations using opposed lateral fields was presumably higher than the

1881 nominal dose, the panel chose a more conservative level and strongly recommends delivering 60

1882 Gy to the tumor bed and involved lymph node stations in the absence of positive surgical margin

1883 (PSM) and ECE.

1884 Currently, the most significant issue concerns the distinction between low risk and high-

1885 risk disease. It is important to recognize that most patients with high risk disease now receive

1886 adjuvant chemoradiotherapy (vide supra). Features including perineural invasion (PNI) and

1887 multiple positive lymph nodes in addition to PSM and ECE conferred eligibility into the RTOG

1888 and EORTC trials that developed this standard. The majority of the patients on these trials did

1889 not have oropharyngeal cancer, however, and secondary analyses have shown that chemotherapy

1890 did not clearly augment the effectiveness of RT for lower-risk pathologic features such as PNI

1891 and multiple nodes. Consequently, one could rationally infer in a contemporary setting that there

1892 is no benefit to be gained from dose escalation or acceleration without PSM or ECE. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1893 Consensus still exists that primary site PSM is a high risk feature. For surgically treated

1894 HPV(+) disease, however, some recent large retrospective studies suggest that minimal nodal

1895 ECE no longer constitutes high risk disease49,50 and that these patients do just as well with

1896 adjuvant RT only (60 Gy) instead of adjuvant chemoradiotherapy. The older studies of post-

1897 operative irradiation followed the traditional paradigm of dose intensification to improve

1898 outcome. Since HPV (+) driven disease has a much more favorable prognosis than its HPV(-)

1899 counterpart, current developmental strategies are focused on therapeutic deintensification. ECOG

1900 331193 is a randomized phase 2 study enrolling p16-positive oropharyngeal cancer patients who

1901 have undergone transoral resection. Patients with “intermediate-risk” pathologic factors (defined

1902 as PNI, LVI, 2-4 LN’s, <1 mm ECE, and/or margins <3 mm) will be randomized to receive 2 Gy

1903 daily either on an investigational arm of 50 Gy or a control arm of 60 Gy. The outcomes of this

1904 trial will help to redefine the standard-of-care for patients receiving adjuvant PORT in this

1905 population, but pending the mature results of this study, post-operative treatment dose and

1906 fractionation should not vary by p16 status.

1907

1908

1909 In the scenario of early T-stage tonsillar carcinoma? 1910 1911  Statement 4K. Unilateral radiotherapy should be delivered to patients with well-lateralized 1912 (confined to tonsillar fossa), T1-T2 tonsillar cancer and N0-N1 nodal category. 1913 o Recommendation strength: Strong 1914 o Quality of evidence: Moderate 1915 1916  Statement 4L. Unilateral radiotherapy may be delivered to patients with lateralized (< 1 cm of 1917 soft palate extension but without base of tongue involvement) T1-T2 N0-N2a tonsillar cancer 1918 without clinical or radiographic evidence of extra-capsular extension, after careful discussion of 1919 patient preferences and the relative benefits of unilateral treatment versus the potential for 1920 contralateral nodal recurrence and subsequent salvage treatment. 1921 o Recommendation strength: Conditional 1922 o Quality of evidence: Low 1923 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1924 1925 Narrative 1926 1927 The hypothetical advantages of ipsilateral-only radiotherapy for tonsillar cancer include

1928 improved short- and long-term functional outcomes and quality-of-life. In addition, the smaller

1929 treatment volumes may potentially lower the small but non-zero long-term risk of radiation-

1930 associated malignancies, particularly in the younger population of patients with HPV-associated

1931 disease. The key concern of restricting radiotherapy to the ipsilateral side is the risk of

1932 contralateral nodal recurrence.

1933 As suggested from both historical surgical literature on patterns of cervical metastases as

1934 well as more recent surgical series of tonsillar cancer patients undergoing bilateral neck

1935 dissections, contralateral nodal spread typically occurs either directly through soft palate or base

1936 of tongue involvement or via aberrant/retrograde flow from bulky unilateral nodal disease. In a

1937 review of 352 patients with oropharyngeal cancer (197 with tonsillar cancer) treated with

1938 surgical resection including bilateral neck dissection, investigators from Ludwig Maximilians

1939 University in Munich143 demonstrated that contralateral nodal metastasis was dramatically lower

1940 for tonsillar cancers (14.7%) as compared to cancers of the base of tongue, soft palate, or

1941 pharyngeal wall (28.8%, 25.9%, and 50.0%, respectively). Among tonsillar cancers risk for

1942 contralateral metastasis increased with higher T-category (6.3% for T1, 17.5% for T2, and 17.3%

1943 for T3-4), though local extension of tonsillar cancers to involve the base of tongue, soft palate or

1944 pharyngeal wall was not described. Additionally, risk for contralateral nodal metastases among

1945 all patients was strongly associated with N2b ipsilateral nodal category (P < 0.001). Investigators

1946 from Hallym University in Seoul144 have described a series of 76 patients with tonsillar cancer

1947 treated with surgery (including 14 having elective contralateral neck dissection for contralateral

1948 N0 status). Contralateral metastases were significantly associated with T3 category, base of NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1949 tongue involvement, soft palate involvement, and ipsilateral neck multilevel involvement, though

1950 only ipsilateral multilevel involvement was significant after multivariate adjustment. In another

1951 publication from this group on a subset of 53 patients,145 these contralateral nodal metastases

1952 were also associated with posterior pharyngeal wall involvement and extracapsular spread in the

1953 ipsilateral nodal metastases. Investigators from Yonsei University in Seoul146 reported on the

1954 pathologic findings in 43 patients with tonsillar cancer treated with surgery including elective

1955 contralateral neck dissection (contralateral N0 status). The authors did not comment on soft

1956 palate or base of tongue invasion, but there was a correlation of contralateral occult metastatic

1957 rate with T-category (0% for T1, 4.5% for T2, and 33.3% for T3-4) and N-category (0 for N0,

1958 18.2% for N1-2a, and 27.3% for N2b). Consequently, these well-documented patterns of cervical

1959 metastases from tonsillar cancer treated with surgery provide support that the concept of

1960 unilateral radiotherapy for patients with early-stage, lateralized tonsillar cancer.

1961 The evidence in the literature supporting the above consensus statement regarding

1962 ipsilateral-only radiotherapy is chiefly comprised of retrospective data with inherit selection and

1963 publication biases, variable follow-up time, limited evidence using modern radiotherapy

1964 techniques, and a lack of data on whether these findings apply specifically to oropharyngeal

1965 cancer resulting from HPV. In developing these guideline statements, we have identified 10

1966 studies which met our initial criteria for review (Table 8 and 9).

1967 One study was described as an “investigational trial” of ipsilateral-only radiotherapy, 147

1968 and this small study included 22 oropharyngeal cancers (13 of the tonsil) patients of whom only

1969 15 were treated with definitive radiotherapy (7) or chemoradiotherapy (8). Primary endpoints of

1970 this study were prospectively recorded patient recurrence in the contralateral neck and reported

1971 xerostomia score. There were 2 (9%) patients with oropharyngeal cancer (both of the tonsil) who NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1972 recurred in the contralateral neck; both were T3 N2b. No patient developed grade 2-3

1973 xerostomia. The other prospective study of ipsilateral-only radiotherapy148 included only 20

1974 patients with tonsillar cancers (none of which involved the base of tongue or soft palate) of

1975 whom only 6 received definitive chemoradiotherapy (using IMRT approximately half the time).

1976 No patient developed a contralateral nodal recurrence. Only one patient developed a grade 2

1977 xerostomia and there were no grade 3 or late toxicities; however, 11 patients did require a

1978 treatment break.

1979 Of the remaining 8 studies, all were retrospective with 3 being generally small. 149-151 In a

1980 study of 32 patients with oropharyngeal cancer (12 with soft palate primaries and 20 with tonsil

1981 primaries, 7 with soft palate extension and 5 with base of tongue extension) treated with

1982 ipsilateral-only radiotherapy (38% received concurrent carboplatin), Kagei et al.149 found no

1983 patient developing a contralateral nodal recurrence. Only 3 patients developed grade 2

1984 xerostomia, and with exception of one case of grade 3 osteoradionecrosis, there were no grade 3

1985 or late toxicities. In a study of 20 patients with tonsillar cancer, of whom 70% received post-

1986 operative ipsilateral-only radiotherapy, Koo et al150 found no patient developed a contralateral

1987 nodal recurrence. Only one patient developed grade 2 xerostomia, and there were no grade 3 or

1988 late toxicities. In a study of 58 patients with tonsillar cancer treated with unilateral radiotherapy,

1989 Liu et al.151found no patient developed a contralateral nodal recurrence. Of note, this is the only

1990 study with p16 data: of 15 patients tested, 60% were p16 positive. Only 4 patients treated

1991 unilaterally developed late toxicities, a significantly smaller percentage than the rate of late

1992 toxicity (22%) of among patients with tonsillar cancer treated with bilateral radiotherapy.

1993 While the remaining 5 studies are also retrospective, each included more than 100

1994 patients with oropharyngeal cancer and all but one with exclusively tonsillar cancer patients as NOT TO BE COPIED, DISSEMINATED OR REFERENCED

1995 well as all but one having the vast majority of patients (>90%) treated with definitive radiation

1996 therapy. Jackson et al.152 reviewed 178 patients with tonsil cancer treated with ipsilateral-only

1997 definitive radiotherapy, although the local extent of disease was not stated beyond T category.

1998 The details of contralateral nodal recurrence were detailed only for the 155 with N0-N1 disease

1999 and among these there were 4 (2.6%) contralateral recurrences (initial stage: one T2N0, one

2000 T3N0, and 2 T1-3N1). Three out of 82 patients (3.6%) of patients with stage III disease

2001 developed a contralateral recurrence, though it is possible that many patients in this older series

2002 (1975-1993) were under staged due to substandard (by today’s standards) imaging. Reported late

2003 toxicities included 2 with grade 2 soft tissue necrosis, 4 with grade 3 osteonecrosis, and 2 with

2004 grade 4 osteonecrosis.

2005 In the largest study to date, O’Sullivan et al.153 reported on 228 patients with tonsillar

2006 cancer treated with definitive ipsilateral-only radiotherapy. Eight patients (3.5%) developed

2007 contralateral nodal recurrence; however, 5 were detected synchronously with a primary site

2008 recurrence and 3 also had ipsilateral nodal recurrence. There were only 3 patients (1.3%) with

2009 exclusive contralateral nodal recurrence, and of these 3, 2 had T3 primaries (all 3 with significant

2010 [>1cm] palate involvement and 2 with significant base of tongue involvement). Of the 30

2011 patients with T3 primaries, 3 developed a contralateral recurrence. Notably, none of the 36

2012 patients with lateral palate involvement (< 1 cm) developed a contralateral recurrence, although 2

2013 out of 39 (5%) patients with lateral base of tongue invasion did, one of which was the only site of

2014 recurrence. While all contralateral nodal recurrences occurred among the 41 N1 patients and

2015 none among the 36 patients with N2 disease (28 N2a, 8 N2b), this is an older series of patients

2016 (1970-1991) in which cross-sectional imaging was “not employed for stage allocation.”

2017 Consequently, it is likely that some proportion of the N1 patients would have been classified as NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2018 N2b had contemporary imaging tools been available. The reported 5-year rate of grade 3-4

2019 toxicity was 3.4% (chiefly osteonecrosis), although the radiotherapy was 2D-based, without CT-

2020 planning.

2021 A recent series by Lynch et al.154 reported the outcomes of 136 patients with tonsillar

2022 cancer treated with primary surgical resection and post-operative ipsilateral-only radiotherapy

2023 with or without chemotherapy. None of these patients had base of tongue involvement, and only

2024 7 patients had lateral (< 1cm) soft palate extension. Eight patients (6%) recurred in the

2025 contralateral neck with 6 (4% of total) of those recurrences being the solitary site of progression.

2026 All 6 of these patients presented with the initial nodal category N2b, with 5 having extra-capsular

2027 extension and 5 being current smokers with >10 pack-years smoking history. Two of the 7

2028 patients with lateral soft palate involvement developed a contralateral recurrence. Late toxicity

2029 included grade 3 osteonecrosis in 9 (7%) and grade 3 fibrosis in 3 (2.2%), and only 8 patients

2030 (6%) required a feeding tube.

2031 The final 2 papers were large series with a majority of patients treated with definitive

2032 intensity modulated radiation therapy (IMRT) and very small proportion receiving

2033 chemotherapy. Chronowski et al.155 found a contralateral recurrence rate of 2% in a series of 102

2034 patients with tonsillar cancer treated with ipsilateral-only radiotherapy, though one of these was

2035 likely a second primary of the contralateral base of tongue. Only patients with primaries

2036 confined to the tonsillar fossa or with less than 1 cm of soft palate extension (base of tongue

2037 involvement was an exclusion) were considered for ipsilateral radiotherapy. A total of 43 N2 (21

2038 N2a, 22 N2b) patients were included in this analysis, and none of them developed a contralateral

2039 recurrence. Long term toxicity rates were not reported, though 9% of patients required feeding

2040 tubes during radiotherapy. In the study by Al-Mamgani et al,156 185 patients (47 with soft palate NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2041 primaries and 32 with N2b nodal metastases) were treated with ipsilateral-only radiotherapy,

2042 there were only 2 contralateral recurrences (1.1% of the entire cohort), and one of these was

2043 initially T1N2b (the other T2N0; both confined to the tonsillar fossa). Late grade 3 toxicity was

2044 only 2.2%, while late grade 2 toxicity was 13%.

2045 The oncologic safety of ipsilateral-only radiation therapy for HPV-associated tonsillar

2046 cancer is suggested by high control rates and low contralateral metastases rates in these relatively

2047 modern series, in which a substantial proportion of patients presumably had disease attributable

2048 to HPV. However, the propensity of small HPV-associated oropharyngeal cancers to metastasize

2049 to lymph nodes raises some concern of whether this population has a higher risk of occult

2050 contralateral lymph nodes and subsequent contralateral nodal recurrence. In a small retrospective

2051 study restricted to T1-T2 tonsillar cancers reported by Shoushtari et al.,157 25% of 28 patients

2052 with p16 positive T1-T2 tonsillar cancer presented with clinically-evident, contralateral nodal

2053 metastases, as compared to none of the 13 patients with p16 negative T1-T2 tonsillar cancer.

2054 The key issue, though, is not the incidence of contralateral clinically-evident lymph node

2055 metastases but rather the baseline risk of occult contralateral lymph node metastases and the risk

2056 of subsequent contralateral recurrence. However, to date there are no data available to inform

2057 whether HPV-associated tonsillar cancers have a higher baseline risk of occult contralateral

2058 lymph node metastases and risk of subsequent contralateral recurrence. Also, whether these risks

2059 are modified by past smoking history is unknown. Consequently, the recommendations regarding

2060 the use of ipsilateral-only radiation therapy are made independent of HPV status and smoking

2061 history.

2062 A substantial experience supports the use of ipsilateral only radiation therapy for tonsillar

2063 cancer confined to the tonsillar fossa with limited nodal metastases as effective therapy with high NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2064 survival rates, expected in-field control, and very rare contralateral recurrence. In addition, the

2065 majority (9 of 10) of N0-N1 patients with contralateral recurrence were observed in series dating

2066 back to the 1970s, with many such patients likely with under-staged N-category due to lack of

2067 modern cross-sectional imaging. Ipsilateral treatment should also reduce the risk of late

2068 toxicities, including xerostomia, dysphagia, and second primary malignancies. The panel

2069 therefore strongly recommends ipsilateral treatment in this cohort. Care in the use of ipsilateral-

2070 only radiation therapy must be taken in cases of T1/T2 category OPSCC with soft palate

2071 involvement and/or N2a categories (Table 10). There is limited clinical experience that can

2072 confirm low contralateral recurrence rates in these scenarios, although the existent data reviewed

2073 above are sufficiently encouraging that ipsilateral radiotherapy may be used in this population,

2074 provided patient preferences are fully engaged on the expected quality-of-life benefits versus the

2075 uncertain risk of contralateral recurrence.

2076 On the other hand, the vast majority of reported contralateral recurrences occurred in

2077 patients with classical contraindications to unilateral treatment (e.g. significant involvement of

2078 soft palate, base of tongue and/or T3 primary size) or tumor characteristics that suggest aberrant

2079 lymphatic flow (i.e. ECE, advanced N2b cases) (Table 9). Thus the panel does not consider

2080 ipsilateral radiotherapy reasonable in the presence of these factors. The panel did discuss whether

2081 patients with small-volume N2b disease or those with minimal base of tongue invasion should be

2082 eligible for ipsilateral radiotherapy. Because of the paucity of data suggesting a low contralateral

2083 recurrence risk in these individuals, they are not included in the conditional recommendation for

2084 the use of ipsilateral radiotherapy.

2085 Future research will hopefully provide additional guidance on this important treatment

2086 planning decision. For example, we eagerly await outcomes and contralateral recurrence risk to NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2087 be reported on series of HPV-associated T1-T2 tonsillar cancers. At present, neither the more

2088 favorable prognosis not the propensity for lymph node metastases of HPV-associated tonsillar

2089 cancer can inform the decision to spare the contralateral neck from radiation therapy. Similarly,

2090 prospective studies of patients with N2b nodal category are needed to better understand the

2091 contralateral recurrence risk with ipsilateral radiotherapy and further motivate treatment

2092 decisions in this relatively common clinical presentation.

2093

2094

2095

2096

2097

2098 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2099 Epilogue: Reflections from the patient advocate 2100 2101 ASTRO clinical practice guidelines include a patient advocate to ensure the patient

2102 perspective is considered during the guideline process and maximize the patient-centeredness of

2103 the product. We asked the patient advocate (Warren Caldwell) for this guideline to provide his

2104 “recommendations” on how medical professionals can improve the care for individuals

2105 diagnosed with oropharyngeal squamous cell carcinoma. His thoughts are below:

2106

2107 1. The sound of a silent cell phone of one awaiting biopsy results is deafening. It will be a

2108 kindness of unimaginable proportion to inform patients in an expeditious manner of the

2109 results of biopsies and scans. When it is clinically possible, avoid scheduling these tests

2110 when results will be extended over a weekend.

2111

2112 2. There is no bad way to inform a patient of good news and no good way to inform them of

2113 such existentially bad news. Make these calls when you have time to patiently respond to

2114 a hundred questions as this is not one of life’s multi-tasking moments for your patient.

2115 They will be riveted to your every word.

2116

2117 3. When the call has ended, the patient’s brain will race: I am self-employed. Will I be able

2118 to work? Should I sell my nice car and get a clunker? Why didn’t I buy life insurance?

2119 Will I see my son earn his Eagle Scout? Will I be at his wedding? Will I play with

2120 grandkids?

2121 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2122 4. How much knowledge is too much for a new patient? A new patient may self-triage with

2123 their need to know, but discuss their desire to hear details early in the conversation.

2124

2125 5. An important blade on the Swiss Army Knife of a patient’s resources to fight the fight is

2126 empowerment. The sudden discovery that so many things are out of one’s control places

2127 a premium on the things that are under one’s control. A shift away from the traditional

2128 paternalistic form of physician/patient relationship towards a teammate approach fosters a

2129 sense of self-control in a world that is madly spinning out of control.

2130

2131 6. The emerging science behind treatment options of OPSCC suggest that there are some

2132 gray areas in utilizing the tools of surgery, radiotherapy and systemic treatment. The

2133 engagement of patient preferences in determining treatment options is a preferred means

2134 of practicing medicine. Explain, Engage and Ask. Formulate a plan together.

2135

2136 7. Establish an understanding of a patient’s aversion to risk versus embrace of reward. Do

2137 they have a predisposition to be more aggressive with a possible loss of quality of life or

2138 less aggressive with the cancer more likely to recur?

2139

2140 8. If a patient elects to go less aggressively, this position should be afforded the same

2141 respect as any other. If they choose to reject some aspect of the recommended treatment,

2142 afford them the dignity and feeling of control of self-directed treatment.

2143 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2144 9. Engage the family during the new patient discussion and enlist them as part of the team.

2145 One could successfully beat it without family but it would be immeasurably more

2146 difficult.

2147

2148 10. Are we getting better and better at fighting the actual disease process but not providing a

2149 great deal of formal attention to the psychosocial aspects of the ordeal? Everyone has

2150 their own unique journey and no amount of counseling or therapy will standardize it (nor

2151 should it) but the lack of the same leaves patients to their own devices and some are

2152 better equipped than others.

2153

2154 11. There will be quite a lot of data management that most will not be used to handling.

2155 Encourage them to buy a Spandex type file in which to keep all appointment calendars,

2156 labs, reports, documents, insurance, and written prescriptions in one place. Use pill

2157 containers to help manage medications. Consider giving them these things as part of a

2158 new patient package.

2159

2160 12. The prospect of the loss of one’s teeth is terrifying to comprehend. Impossibly so. It is

2161 disquieting to be a fortunate ex-patient counseling a less fortunate ex-patient who is sad

2162 that “I used to be attractive, now I look like a different person.” Encourage patients to use

2163 their dental trays in the shower. It gives them a regular routine to use them and they are

2164 able to spit excess fluoride saliva down the drain.

2165 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2166 13. The radiation mask is a special hurdle. The loss of control by being imprisoned by

2167 something one millionth of an inch from one’s face is near complete. If a patient exhibits

2168 complete anxiety, medicate them. But if a patient must drive themselves to appointments

2169 and cannot be medicated, encourage them to leave home early for the appointment to

2170 avoid traffic stress. Listen to comforting music on the way to the hospital. Breathe slowly

2171 and deeply. Close one’s eyes during treatment. Offer them the opportunity to listen to

2172 music of their choice during the treatment.

2173

2174 14. The loss of taste is much more of a loss than it would seem on the surface. It is shocking

2175 how important the sensation of taste is in driving a person’s appetite. Indeed, when one

2176 cannot taste, the act of eating is actually a tiresome ordeal in itself.

2177

2178 15. Have patients keep Magic Mouthwash (a lidocaine mixture) right next to their chair and

2179 bedside, not in the kitchen. Keep it super-convenient and they will use it more regularly.

2180 Recommend that patients keep moisturizer in their vehicles so that they may slather it on

2181 their necks immediately after the treatment. Make the supportive care as convenient as

2182 possible for themselves.

2183

2184 16. Post-surgical and radiation mucositis can make it very difficult to sleep in a normal

2185 position. Encourage patients to purchase and utilize a reclining chair to sleep in and when

2186 they transition back to a bed, have them elevate with multiple pillows. When awake and

2187 in the chair, set up a TV tray and line up all the accoutrements of healing. Water, spit cup,

2188 prescriptions, books, TV remote. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2189

2190 17. Encourage the patient to take mastery over their mind. Have them hang a white dry erase

2191 board or the like at their home or office. Sequentially number the amount of treatments

2192 (30, 29, 28….1, DONE). Make it a ceremony to “X” out each number as it progresses.

2193 Look forward to it. Engage family and allow them to “X” out some of the days. Write

2194 slogans of encouragement on this board.

2195

2196 18. Once the miraculous “tah-dah” moment of finishing treatment is reached, the “New

2197 Normal” begins. As the treatment “gift” keeps on giving in the weeks and months

2198 following radiation, a patient may feel the need for additional sleep. Encourage naps as

2199 necessary. This “lack of productivity” can result in financial strain due to reduced income

2200 but it is a matter of aligning things in their necessary hierarchy of importance.

2201

2202 19. The effect of radiotherapy on the skin can cause the neck to look twenty years older than

2203 the rest of the patient. Don’t discount the importance of trying to reduce the long-term

2204 visible side effects of treatment.

2205

2206 20. The hiring process for staff should place a special premium on the human quality of

2207 kindness. The smallest act of kindness will be something the patient will remember for

2208 life. Begin with the valet in the parking lot. They are the first face a patient will encounter

2209 and the last to see and have an opportunity to place their positive imprint on the entire

2210 visit. The same care should be extended with check-in staff, nurses and of course, NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2211 physician assistants and physicians. By doing so, you will be treating patients as a human

2212 being instead of a disease, which is crucial to maintain dignity throughout the process.

2213

2214 21. If the patient has children, forcefully encourage (and remind) them to vaccinate their kids

2215 against HPV. Motivate them to be strong advocates for HPV vaccination for their

2216 extended family and friends.

2217 2218

2219 22. If you have read this far, I thank you from the depths of my heart. It is because of

2220 you and people like you that I backpacked 110 miles at Philmont Scout Ranch with

2221 my son last year. Remember why you do what you do!

2222

2223 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2224 TABLES 2225 2226 Table 1. Recommendation statements, including recommendation strength, quality of evidence, and percent consensus. 2227 Percent (%) agreement Strength of Quality of Guideline statement with guideline recommendation evidence statement KQ1. When is it appropriate to add systemic therapy to definitive radiotherapy in the treatment of oropharyngeal squamous cell carcinoma (OPSCC)?

In the scenario of stage IV A-B disease:

A. Concurrent high-dose intermittent cisplatin should be delivered to patients with stage IVA-B oropharyngeal squamous cell carcinoma receiving definitive radiotherapy. Strong High 100%

B. Concurrent cetuximab or carboplatin-fluorouracil should be delivered to patients with stage IVA-B oropharyngeal squamous cell carcinoma receiving definitive Strong High 88% radiotherapy who are not medically fit for high-dose cisplatin. C. Concurrent weekly cisplatin may be delivered to patients with stage IVA-B oropharyngeal squamous cell carcinoma receiving definitive radiotherapy who are not medically fit Conditional Low 94% for high-dose cisplatin, after a careful discussion of patient preferences and the limited prospective data supporting this regimen. D. Concurrent cetuximab should not be delivered in combination with chemotherapy to patients with stage IVA- Strong High 100% B oropharyngeal squamous cell carcinoma receiving definitive radiotherapy. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

E. Intra-arterial chemotherapy should not be delivered to patients with stage IVA-B oropharyngeal squamous cell Strong High 100% carcinoma receiving definitive radiotherapy.

In the scenario of stage III disease:

F. Concurrent systemic therapy should be delivered to patients with T3 N0-1 oropharyngeal squamous cell Strong Moderate 100% carcinoma receiving definitive radiotherapy. G. Concurrent systemic therapy may be delivered to patients with T1-T2 N1 oropharyngeal squamous cell carcinoma receiving definitive radiotherapy who are considered at Conditional Low 88% particularly significant risk for locoregional recurrence, after a careful discussion of patient preferences and the limited evidence supporting its use.

In the scenario of stage I-II disease:

H. Concurrent systemic therapy should not be delivered to patients with stage I-II oropharyngeal squamous cell Strong Low 100% carcinoma receiving definitive radiotherapy.

KQ2. When is it appropriate to deliver post-operative radiotherapy with and without systemic therapy following primary surgery of oropharyngeal squamous cell carcinoma (OPSCC)?

In the scenario of positive margins and/or extrascapsular nodal extension (ECE)?

A. Concurrent high-dose intermittent cisplatin should be delivered with post-operative radiotherapy to patients with positive surgical margins and/or extracapsular nodal Strong Moderate 100% extension; this high-risk population includes patients independent of HPV status or the extent of extranodal tumor. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

B. Concurrent weekly cisplatin may be delivered with post- operative radiotherapy to patients who are considered inappropriate for standard high-dose intermittent cisplatin Conditional Low 94% after a careful discussion of patient preferences and the limited evidence supporting this treatment schedule. C. For the high-risk post-operative patient unable to receive cisplatin-based concurrent chemoradiotherapy, radiotherapy alone should be routinely delivered without concurrent systemic therapy; given the limited evidence supporting Strong Moderate 94% alternative regimens, treatment with non-cisplatin systemic therapy should be accompanied by a careful discussion of the risks and unknown benefits of the combination. D. Patients treated with post-operative radiotherapy should Strong Moderate 88% not receive concurrent weekly carboplatin. E. Patients treated with post-operative radiotherapy should not receive cetuximab, either alone or in combination with Strong Low 94% chemotherapy, although such regimens are currently under investigation. F. Patients treated with post-operative radiotherapy should not routinely receive concurrent weekly docetaxel given the Strong Low 88% limited evidence supporting its use, although such regimens are currently under investigation. G. Patients treated with post-operative radiotherapy should not receive concurrent mitomycin-C, alone or with Strong Moderate 100% bleomycin, given the limited evidence and experience supporting its use. H. Post-operative chemotherapy should not be delivered Strong High 94% alone or sequentially with post-operative radiotherapy.

In the scenario of intermediate-risk pathologic factors such as lymphovascular invasion (LVI), perineural invasion (PNI), T3- 4 disease, or positive lymph nodes:

NOT TO BE COPIED, DISSEMINATED OR REFERENCED

I. Patients with intermediate-risk factors should not routinely receive concurrent systemic therapy with post-operative Strong Moderate 88% radiotherapy. J. Patients with intermediate-risk factors whose surgical procedure and/or pathologic findings imply a particularly significant risk of locoregional recurrence may receive concurrent cisplatin-based chemotherapy after a careful Conditional Low 88% discussion of patient preferences and the limited evidence supporting its use in this scenario; alternative systemic treatment regimens should only be used in the context of a clinical trial K. Post-operative radiotherapy should be delivered to Strong Low 94% patients with pathologic T3 or T4 disease. L. Post-operative radiotherapy should be delivered to patients Strong Low 100% with pathologic N2 or N3 disease. M. Post-operative radiotherapy may be delivered to patients with pathologic N1 disease after a careful discussion of Conditional Low 88% patient preferences and the limited evidence of outcomes following surgery alone in this scenario. N. Post-operative radiotherapy may be delivered to patients with LVI and/or PNI as the only risk factor(s), after a careful Conditional Low 100% discussion of patient preferences and the limited evidence of outcomes following surgery alone in this scenario.

In the scenario of no pathologic risk factors:

O. Post-operative radiotherapy may be delivered to patients without conventional adverse pathologic risk factors only if the clinical and surgical findings imply a particularly Conditional Low 88% significant risk of locoregional recurrence, after a careful discussion of patient preferences and the potential harms and benefits of radiotherapy.

NOT TO BE COPIED, DISSEMINATED OR REFERENCED

KQ3. When is it appropriate to use induction chemotherapy in the treatment of oropharyngeal squamous cell carcinoma (OPSCC)?

A. Induction chemotherapy should not be routinely delivered Strong High 100% to patients with OPSCC.

KQ4. What are the appropriate dose, fractionation, and volume regimens with and without systemic therapy in the treatment of oropharyngeal squamous cell carcinoma (OPSCC)?

In the scenario of definitive non-surgical therapy:

A. A dose of 70 Gy over 7 weeks should be delivered to gross primary and nodal disease in patients with stage III-IV Strong Moderate 100% oropharyngeal squamous cell carcinoma selected to receive standard, once-daily definitive radiotherapy. B. The biologically equivalent dose of approximately 50 Gy in 2 Gy fractions or slightly higher should be delivered Strong Low 100% electively to clinically- and radiographically-negative regions at-risk for microscopic spread of tumor. C. Altered fractionation should be used in patients with stage IVA-B oropharyngeal squamous cell carcinoma treated with Strong High 94% definitive radiotherapy who are not receiving concurrent systemic therapy. D. Either accelerated radiotherapy or hyperfractionated radiotherapy may be used in patients with oropharyngeal squamous cell carcinoma treated with altered fractionation Conditional High 100% definitive radiotherapy after a careful discussion of patient preferences and the limited evidence supporting one regimen over the other. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

E. Either standard, once-daily radiotherapy or accelerated fractionation may be used when treating oropharyngeal squamous cell carcinoma with concurrent systemic therapy, Conditional High 88% after a careful discussion of patient preferences and the risks and benefits of both approaches. F. Altered fractionation should be used in patients with T3 N0-1 oropharyngeal squamous cell carcinoma treated with Strong Moderate 94% definitive radiotherapy who do not receive concurrent systemic therapy. G. Altered fractionation may be used in patients with T1-2 N1 or T2 N0 oropharyngeal squamous cell carcinoma treated with definitive radiotherapy alone who are considered at Conditional Low 100% particularly significant risk of locoregional recurrence, after a careful discussion of patient preferences and the limited evidence supporting its use in this scenario.

In the scenario of adjuvant postoperative radiotherapy:

H. Adjuvant post-operative radiotherapy should be delivered to regions of microscopically-positive primary site surgical Strong Moderate 100% margins and extracapsular nodal extension at 2 Gy/fraction once daily to a total dose between 60 and 66 Gy I. Adjuvant post-operative radiotherapy delivered without concurrent systemic therapy should treat regions of microscopically-positive primary site surgical margins and Conditional Weak 100% extracapsular nodal extension at 2 Gy/fraction once daily to a total dose of 66 Gy, although there are limited data guiding this recommendation. J. Adjuvant post-operative radiotherapy should be delivered to the tumor bed and involved, dissected lymph node regions at 2 Gy/fraction once daily to a total dose of 60 Gy in the Strong Moderate 100% absence of primary site positive margins and extracapsular nodal extension. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

In the scenario of early T-stage tonsillar carcinoma:

K. Unilateral radiotherapy should be delivered to patients with well-lateralized (confined to tonsillar fossa), T1-T2 Strong Moderate 88% tonsillar cancer and N0-N1 nodal category. L. Unilateral radiotherapy may be delivered to patients with lateralized (< 1cm of soft palate extension but without base of tongue involvement) T1-T2 N0-N2a tonsillar cancer without clinical or radiographic evidence of extra-capsular Low Conditional 100% extension, after careful discussion of patient preferences and the relative benefits of unilateral treatment versus the potential for contralateral nodal recurrence and subsequent salvage treatment. 2228 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2229 Table 2: Summary of phase III randomized trials comparing definitive radiotherapy with definitive radiotherapy plus 2230 concurrent systemic therapy (Key Question 1) 2231 Number Chemotherapy Author and % % stage of Radiotherapy regimen regimen/ LRC benefit OS benefit year oropharynx III patients schedule YES Bolus cisplatin q3 295 total Adelstein, weeks vs. cisplatin (271 59% 4% 70 Gy in 35 fx over 7 weeks -- 3-year: 23% (RT) vs. 20039 + 5-FU with split- analyzed) 37% (RT + bolus) vs. course p=0.014 YES Ruo Redda, 164 (157 70 Gy in 35 daily fx over 7 Carboplatin q2 201011 56% 23% NO analyzed) weeks weeks 5-year: 6.9% (RT) vs.

9% (CRT), p=0.02 Mitomycin C after NO 10 Gy and on last Zakotnik, 66-70 Gy in 33 to 35 daily 64 64% 6% day of RT + -- Except 10% vs. 38% 199814 fx over 7 weeks bleomycin twice for oropharyngeal, weekly p=0.019 YES YES

Denis, 70 Gy in 35 daily fx over 7 Carboplatin and 5- 226 100% 32% 5-year: 24.7% (RT) 5-year: 15.8% (RT) 200416 weeks FU q3 weeks vs. 47.6% (CRT), vs. 22.4% (CRT), p=0.002 p=0.05 YES YES Continuous 50.4 Gy in 28 daily fx over Semrau, infusion 5-FU and 5-year survival with 240 74% 4% 38 days + 19.5 Gy boost in 5-year: 15.8% (RT) 200617 carboplatin for 5d, LRC: 12.6% (RT) 13 daily fx (total 69.9 Gy) vs. 25.6% (CRT), weeks 1 and 6 vs. 22.7% (CRT), p=0.016 p=0.01 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Number Chemotherapy Author and % % stage of Radiotherapy regimen regimen/ LRC benefit OS benefit year oropharynx III patients schedule Ghadjar, YES 201218 29% (plus Daily cisplatin 20 224 100% 5% stage 74.4 Gy in 62 fx BID mg/m2 for 5d, 10-year: 40% (CRT) NO II) weeks 1 and 5 vs. 32% (RT) p=0.049 YES YES Not 70.2 Gy in 39 fx BID over Cisplatin, 5-FU, Wendt, 298 (270 reported; 42% 51 days in 3 cycles of 23.4 and leucovorin q3 3-year: 17% (RT) 3-year: 24% (RT) vs. 199819 analyzed) 5% T1-2, Gy with 11 days in between weeks vs. 36% (CRT), 48% (CRT), 35% N0-1 p<0.004 p<0.0003 YES 64 Gy in 32 fx BID (very accelerated RT alone) vs. 5-year freedom from Bourhis, Cisplatin and 5-FU 109 62% 0% 62-64 Gy in 31-32 fx BID locoregional or NO 201120 q2 weeks with RT every other wk distant failure: (CRT) 49.1% (CRT) vs. 25% (RT), p=0.005 Budach, Continuous 16 Gy in 8 daily fx + 61.6 YES YES 200522 infusion Gy in 44 fx BID (total 77.6 fluorouracil on 384 60% 6% Gy) (RT) vs. 30 Gy in 15 5-year: 37.4% (RT) 5-year: 23.7% (RT) days 1-5 + bolus daily fx + 40.6 Gy in 29 fx vs. 49.9% (CRT), vs. 28.6% (CRT), mitomycin on days BID (total 70.6 Gy)(CRT) p=0.001 p=0.023 5 and 36 70 Gy in 35 daily fx over 7 YES YES weeks (CRT) vs. 40 Gy in Carboplatin (3

20 daily fx over 4 weeks cycles of 4 days 3-year: 58.3% Bourhis, 3-year: 42.6% (CRT) 840 66% N/R plus 30 Gy in 20 fx BID for CRT, 2 cycles (CRT) vs. 54.6% 201223 vs. 36.5% over 2 weeks (AccRT) vs. of 5 days for (AccRT) vs. 50.1% (VeryAccRT), 64.8 Gy in 36 fx BID over AccRT) + (VeryAccRT), p=0.040 3.5 weeks (VeryAccRT) NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Number Chemotherapy Author and % % stage of Radiotherapy regimen regimen/ LRC benefit OS benefit year oropharynx III patients schedule fluorouracil (q3 p=0.033 (AccRT vs. wks for CRT, q4 VeryAccRT), wks for AccRT) p=0.045 (CRT vs. VeryAccRT) YES

Ghosh- 46% 66-70 Gy in 5 fx per week Weekly cisplatin 5-year: 32% Laskar, 186 89% (stage II (CRT) vs. 66-70 Gy in 6 fx NO 30 mg/m2 (ConvRT) vs. 49% 201425 and III) per week (AccRT) (CRT) vs. 27% (AccRT), p=0.049 Fallai, 200627 66-70 Gy in 33-35 daily fx YES Olmi, (RT arm 1, CRT) vs. 64- 27% (all Carboplatin and 5- 2003158 192 100% 67.2 Gy in 40-42 fx BID, 5-year: 21% (RT 1) NO T3 N0-1) FU q4 weeks with 2 week break (RT arm vs. 18% (RT 2) vs. 2) 48% (CRT), p=0.02 Bolus cisplatin q3 72 Gy in 42 fx (3d-CRT) or weeks vs. bolus Ang, 201435 891 70% 14% 70 Gy in 35 fx (IMRT) over NO NO cisplatin + 6 weeks cetuximab Bolus cisplatin q3 Nguyen-Tan, 70 Gy in 35 fx over 7 weeks 743 (721 weeks (3 cycles for 201437 60% 22% (ConvRT) vs. 72 Gy in 42 NO NO analyzed) ConvRT, 2 cycles Ang, 20104 fx over 6 weeks (AccRT) for AccRT) Weekly intra- 46 Gy in 23 daily fx + 24 Rasch, arterial cisplatin 239 63% 5% Gy boost in 12 daily fx NO NO 201040 vs. intravenous (total 70 Gy) cisplatin q3 weeks 70 Gy in 35 daily fx over 7 Bonner, Weekly cetuximab YES YES 424 60% 25% weeks or 72-76.8 Gy in 60- 2006159,160 (+ loading dose 64 fx BID over 10 weeks or NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Number Chemotherapy Author and % % stage of Radiotherapy regimen regimen/ LRC benefit OS benefit year oropharynx III patients schedule 32.4 Gy in 18 daily fx + one week before 3-year: 34% (RT) 3-year: 45% (RT) vs. 39.6 Gy boost in 24 fx BID RT) vs. 47% (CRT), 55% (CRT), p=0.05 p<0.01 5-year: 36.4% (RT) vs. 45.6% (CRT), p=0.018

2232 Abbreviations: CRT = Chemoradiotherapy; -- = Not Reported, NS= Not significant; ConvRT: Conventional radiotherapy; AccRT: 2233 Accelerated radiotherapy; Fx: Fractions; BID: Twice daily; LRC = locoregional control; OS = overall survival. 2234

2235 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2236 Table 3. Summary of phase III randomized trials comparing postoperative radiotherapy with postoperative radiotherapy plus 2237 concurrent high-dose intermittent cisplatin (Key Question 2).

Number Median % Trial number of LRC Benefit DM Benefit DFS/PFS Benefit OS Benefit follow-up oropharynx patients YES YES YES Bernier, 5-year cumulative 200444 334 5 years 30% incidence of NO 5-year: 47% 5-year: 53% (EORTC recurrence: 18% (CRT) vs. 36% (CRT) vs. 40% 22931) CRT vs. 31% RT (RT) (p=0.04) (RT) (p=0.02) (p=0.007) YES

At median follow- YES up 45.9 months, Cooper, 200445 416 3.8 years 42% cumulative NO Hazard ratio 0.78 NO (NRG 9501) incidence of (p=0.04) favoring recurrence: 19% CRT CRT vs. 30% CRT (p=0.01) Cooper, 201246 410 9.4 years 42% NO NO NO NO (NRG 9501) 2238 Abbreviations: EORTC = European Organisation for Research and Treatment of Cancer; NRG =; LRC = locoregional control; DM = 2239 distant metastasis; DFS = disease-free survival; OS = overall survival; CRT = chemoradiotherapy; RT = radiotherapy

2240

2241

2242 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2243 Table 4. Summary of phase III randomized trials comparing postoperative radiotherapy with postoperative radiotherapy plus 2244 concurrent systemic therapy not using high-dose intermittent cisplatin (Key Question 2)

Number % Chemotherap Author Patient risk RT dose/ of oropharyn y LRC Benefit DM Benefit DFS Benefit OS Benefit and year factors schedule patients x schedule 54 Gy in 32 fx over 6.5 weeks ECE: 100% YES

Bachaud, Close/positive YES YES PSM: 35% 199654 Cisplatin margins: boost 5-yr survival

Bachaud, 83 15% weekly 50 mg to 65-70 Gy in without LRR NO 5-year: 23% 2-year: 13% ≥3 positive 199155 x 7-9 1.8-2.0 Gy fxs 55% (RT) vs. (RT) vs. 45% (RT) vs. 36% nodes or 70% (CRT), (CRT), p<0.02 (CRT), p<0.01 bulky mass: Metastatic p=0.05 33% nodes: boost to 65-74 Gy in 3 Gy fxs YES PSM At least 54 Gy 27% Mitomycin C in 27-30 fxs 5-year (crude): Rewari, 205 23% 15 mg/m2 x 1- over 5.5-6 69.9% (RT) NO -- NO 200657 ≥2 positive 2 weeks, median vs. 85.3% nodes: 34% 60 Gy (CRT), p=0.008 ECE: 53% YES YES Mitomycin C Zakotnik, PSM: 12% 15 mg/m2 56–70 Gy in 200763 5-year: 65% 5-year: 33% 114 30% Bleomycin 28-35 fxs over NO NO Smid, (RT) vs. 88% (RT) vs. 53% LVI: 18% 5 mg twice 5.5-7 weeks 200362 (CRT), (CRT), weekly p=0.026 p=0.035 PNI: 4% NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Number % Chemotherap Author Patient risk RT dose/ of oropharyn y LRC Benefit DM Benefit DFS Benefit OS Benefit and year factors schedule patients x schedule 54 Gy in 30 fx ECE: 31% over 6 weeks for negative Carboplatin 50 PSM: margins Racadot, mg/m2 twice 144 22% 49% NO -- -- NO 200864 weekly x 6-8 72 Gy in 40 fx weeks ≥2 positive over 8 weeks nodes: 36% for positive margins ECE: 75% At least 59.4 Carboplatin Gy in 33 fxs Argiris, 100 mg/m2 72 PSM: 29% 32% over 6.5 weeks NO NO NO NO 200865 weekly x 6 (boost weeks PNI: 10% allowed) Methotrexate Tobias, 253 or vincristine, 66 bleomycin, 2010 Varied by (pts who PSM: 53% 23% fluorouracil + -- -- NO NO center received methotrexate surgery) on days 1 + 14 of RT 2245 Abbreviations: RT: Radiation Therapy; CRT: Chemoradiotherapy; Pts: Patients; Fx: Fractions; Yrs = Years; -- = Not reported; DM = 2246 Distant metastases; OS = overall survival; LRC = locoregional control; LRF = locoregional failure; w/o = Without; NS = Not 2247 significant; Min = Minimum; DFS = disease-free survival; ECE = extracapsular extension; PSM = positive surgical margins; PNI = 2248 perineural invasion; LVI = lymphovascular invasion;

2249

2250 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2251 Table 5. Summary of key studies investigating intermediate-risk pathologic factors for recurrence following primary surgical 2252 resection with and without adjuvant radiotherapy (Key Question 2) 2253 Author and Number of % N stage: T stage: Prevalence LRF Impact DFS Impact OS Impact year patients oropharynx N0/N1/N2/ T1/T2/T3/T of LVI N3 4 and/or PNI YES (LVI, PNI)

801 N0: 29% 5-year: YES YES

N1: 15% T1: 12% LVI: 4% 33% (LVI present)

Langendijk, N2a: 3% T2: 24% vs. 21% (LVI 5-year: 65% (class 5-year: 67% (class 20% 200573 N2b: 37% T3: 32% PNI: 16% absent), p=0.011 I) vs. 47% (class I) vs. 50% (class II)

N2c: 13% T4: 32% II) vs. 32% (class vs. 36% (class III),

N3: 3% 31% (PNI present) III), p<0.0001* p<0.0001*

vs. 22% (PNI absent), p=0.026 P values not reported YES

3-year: 8.5% pT category: risk T1: 14% (surgery) vs. 4.7% N0: 50% ratio 1.6, pT3/4 vs. T2: 38% (surgery + RT) N1: 18% pT1/2 (p=0.0009) Ambrosch, T3: 30% 503 28% N2a: 2% -- -- 200174 T4: 18% Failure in neck N2b: 26% pN category: risk (oropharynx without RT: N2c: 5% ratio 1.9 and 2.8 for only) pN1 and pN2, pN0: 5.5% vs. 0% respectively vs. pN1: 6.3% vs. 3% pN0 (p=0.0001) pN2: 24% vs. 7% NO (+/- RT) T1: 24% Jackel, NO (+/- RT) 118 29% N1: 100% T2: 38% -- -- 200875 Any regional failure T3: 27% without RT: 9% vs. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Author and Number of % N stage: T stage: Prevalence LRF Impact DFS Impact OS Impact year patients oropharynx N0/N1/N2/ T1/T2/T3/T of LVI N3 4 and/or PNI T4: 12% 21% (10% isolated, (oropharynx no p value) only) 3-year risk of isolated regional recurrence: 11.2% (no RT) vs. 2.9% RT (p=0.09) YES (T stage) Tx: 2% T1: 7% 94.7% (T1-2) vs. Leemans, 244 13% N/R T2: 27% -- 83.8% (T3-T4), -- -- 199480 T3: 43% p=0.015 T4: 21% (primary site failure) N0: 9% -- N1: 14% T1: 41% LVI: 15%

Haughey, N2a: 18% T2: 34% 171 100% NO (T-stage) -- 201281 N2b: 42% T3: 16% PNI: 12%

N2c: 10% T4: 9%

N3: 7% YES (LVI)

50% (LVI present) LVI: 41% Ravasz, T1-T2: 15% vs. 16% (LVI 80 28% N/R -- -- 199382 T3-T4: 85% absent), p=0.005 PNI: 13% (local only failure))

(PNI not significant) NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Author and Number of % N stage: T stage: Prevalence LRF Impact DFS Impact OS Impact year patients oropharynx N0/N1/N2/ T1/T2/T3/T of LVI N3 4 and/or PNI N0: 32% T1: 5% YES (LVI) 45% Bastit, N1: 29% T2: 30% Either LVI or 420 -- -- 200183 N2: 34% T3: 55% PNI: 18% Tumor emboli RR

N3: 5% T4: 11% 1.48, p=0.061 Sydney N0: 43% Sydney N1: 15% T1: 18% N2a: 3% T2: 41% N2b: 29% T3: 30% N2c: 7% 237 42% T4: 11% YES (PNI) N3: 3% -- NO (PNI, LVI) McMahon, Sydney, 95 Sydney, -- 200384 Lanarkshir 23% Lanarkshire (on multivariable Lanarkshire e Lanarkshire T1: 17% analysis) N0: 69% T2: 63% N1: 10% T3: 17% N2a: 2% T4: 4% N2b: 13%

N2c: 4% N3: 1% YES (PNI)

T1: 8% 23% (PNI present) 47% (oral -- T2: 42% PNI: 52% vs. 9% (PNI absent), Fagan,199885 142 cavity and N/R -- T3: 26% p=0.02 oropharynx) T4: 24% (local only; oral cavity + oropharynx) NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Author and Number of % N stage: T stage: Prevalence LRF Impact DFS Impact OS Impact year patients oropharynx N0/N1/N2/ T1/T2/T3/T of LVI N3 4 and/or PNI YES (PNI)

Ipsilateral regional recurrence: 26% (PNI present) vs. 6% (PNI absent), p=0.001 N0: 37%

N1: 13% LVI: 7% T1-2: 52% 19% (LVI present) NO (PNI, LVI) NO (PNI, LVI) Lanzer, N2a: 10% 291 32% T3-4: 30% vs. 7% (LVI absent), 201386 N2b: 31% PNI: 8% Tx: 8% p=0.048 N2c: 5%

N3: 4% Contralateral regional recurrence: 9% (PNI present) vs. 1% (PNI absent), p=0.002

LVI: negative Nx: 8% Tx: 11% N0: 38% T1: 2% Peters, 240 18% N1: 18% T2: 18% PNI: 24% NO (PNI) -- -- 1993138 N2: 22% T3: 50% N3: 15% T4: 22% 2254 Abbreviations: LVI: Lymphovascular invasion; PNI: Perineural invasion; --: Not reported; LRF = locoregional failure; DFS = 2255 disease-free survival; OS = overall survival 2256 * Class I (intermediate risk): free surgical margins and no extranodal spread; class II (high risk): T1, T2, and T4 tumors with close or 2257 positive surgical margins or one lymph node metastasis with extranodal spread; class III (very high risk):T3 tumors with close or 2258 positive surgical margins or multiple lymph node metastases with extranodal spread or N3 neck. 2259 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2260 Table 6. Summary of phase III randomized trials comparing radiotherapy (or chemoradiotherapy) with induction 2261 chemotherapy followed by the same (Key Question 3) Author Number of % Induction Concurrent Radiotherapy LRC DM Benefit OS Benefit and year patients oropharynx regimen regimen regimen Benefit Induction chemotherapy followed by local therapy (surgery or radiotherapy alone) Resection and adjuvant RT YES YES YES (45-50 Gy) for 59.3% Zorat 2004, operable pts (group A), Cisplatin-5FU x 84% (group 3 years: 24% DM 5/10 years: Pacagnella 237 None and definitive 54.6% 4 cycles A) vs 72% (IC) vs. 42% DM 21%/16% vs. 199494 RT for (group B) (group B), (RT only), 8%/6%, p=0.04 for inoperable p=0.05 (p=0.04) inoperable patients (65-70 Gy) (group B) 70 Gy/35 fx YES over 7 weeks, Domenge, 318 (161pts - Cisplatin-5FU x plus 5 Gy Median 5.1 yr in 100% None NO NO 200096 no IC grp) 3 cycles boost to IC vs. 3.3 yr RT/ residual tumor; surgery alone or resection (p=0.03) Lewin,199 Cisplatin-5FU x 64-70 Gy in 2 461 55% None -- NO NO 7161 3 cycles Gy daily fx Sequential (Induction chemotherapy followed by chemotherapy) 439 pts: 155 Cisplatin-5FU (TPF-CRT), vs. Cisplatin- Bolus Hitt,201416 156 ( PF- 5FU-Docetaxel x 70 Gy/35 daily 43% cisplatin q3 NO NO NO 2 CRT ) and 3 cycles fx weeks 128 (CRT alone) Haddad, Cisplatin-5FU- Bolus CRT: 72 Gy in 2013102 145 55% Docetaxel x 3 cisplatin q3 NO NO NO 6 weeks cycles weeks (no NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Author Number of % Induction Concurrent Radiotherapy LRC DM Benefit OS Benefit and year patients oropharynx regimen regimen regimen Benefit IC); weekly (accelerate carboplatin boost) (AUC 1.5) after TPF for Post-IC: 72 Gy good in 6 weeks for responders, poor weekly responders; 70 docetaxel (20 Gy in 7 weeks mg/m2) for for good poor responders responders NO

74-75 Gy in 4 Cisplatin-5FU- Docetaxel, Except Cohen, courses of BID 285 58% Docetaxel x 2 hydroxyurea, NO ~10% (IC) vs. NO 2014103 RT, 1 week cycles 5-FU 20% (CRT) DM apart without prior LRR (p=0.043) 2262 Abbreviations: CRT = Chemoradiotherapy; RT = Radiotherapy; -- = Not Reported; IC = induction chemotherapy; NS= Not 2263 significant; Fx: Fractions; BID: Twice daily; DM = distant metastasis; Gy = Gray; LRR = locoregional recurrence. FU = fluorouracil; 2264 TPF = docetaxel, cisplatin, fluoruracil; AUC = area under curve.

2265

2266

2267

2268 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2269 Table 7: Summary of phase III randomized trials comparing radiotherapy with altered fractionation radiotherapy. (Key 2270 Question 4) 2271 Number Altered Chemotherapy Author % % stage Reference of fractionation RT regimen/ LRC Benefit OS Benefit and year oropharynx III RT regimen patients regimen schedule Carboplatin (3 YES (to CRT) 40 Gy in 20 daily cycles of 4 days YES (to CRT) fx over 4 weeks for CRT, 2 3-year: 58.3% (CRT) Bourhis, plus 30 Gy in 20 70 Gy 35 cycles of 5 days vs. 54.6% (AFX) vs. 3-year: 42.6% 201223 fx BID over 2 840 66% -- fractions over for AFX) + 50.1% (VeryAFX), (CRT) weeks (AFX), 7 weeks fluorouracil (q3 vs. 36.5% 64.8 Gy in 36 fx wks for CRT, p=0.033 (AFX vs. (VeryAFX), BID over 3.5 q4 wks for VeryAFX), p=0.045 p=0.040 weeks (VeryAFX) AFX) (CRT vs. VeryAFX) 743 (721 Bolus cisplatin AFX with boost 70 Gy in 35 Nguyen- analyzed) q3 weeks (3 plus cisplatin (72 fractions over Tan, 60% 21% cycles for SFX, NO NO Gy in 42 fxs over 7 weeks plus 201437 2 cycles for 6 weeks) cisplatin AFX) HFX (81.6 Gy NO BID); NO

AFX-continuous Not Except HFX vs. Beitler, (72 Gy in 6 70 Gy/35 fx, Except HFX vs. SFX 1076 60% reported None SFX HR 0.81, 2014112 weeks); AFX-split once-daily HR 0.79, p=0.05 (if p=0.05 (if (67.2 Gy with 2- censoring after 5 censoring after 5 wk rest after 38.4 years) years) Gy) YES Hyperfractionated

Horiot, and AFX (72 Gy 500 100% -- 70 Gy in 35 fx None 5 years: 46% SFX arm NO 1997113 in 45 fx over 5 vs. 59% AFX arm; weeks) p=0.02 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

Number Altered Chemotherapy Author % % stage Reference of fractionation RT regimen/ LRC Benefit OS Benefit and year oropharynx III RT regimen patients regimen schedule AFX (68 Gy in 23 fxs of 2 Gy and 20 Zackrisson, 68 Gy/34 fx, 750 49% 28% daily boost fxs of None NO NO 2011118 once-daily 1.1 Gy,over 4.5 weeks) YES

AFX, 66-68 66-68 Gy in Overgaard1 29% 5-year: 70% (AFX) vs 1476 21% Gy/33-34 fx, 6 fx 33-34 fx, 5 fx None NO 25 (pharynx) 60% (SFX) per week per week group p = 0.0005 YES

AFX, 66-70 66-68 Gy in Overgaard1 53% 5 years: 42% (AFX) 900 37% Gy/33-35 fx, 6 fx 33-34 fx, 5 fx None NO 26 (pharynx) vs 30% (SFX) per week per week group p = 0.004 Favorable- AFX (59.4 Gy in 70 Gy in 35 Poulsen, 12 % 343 67% 33 fxs over 24 fxs over 49 None NO NO 2001128 Unfavorab days) days le- 19% YES 80.5 Gy in 70 BID 70 Gy 35 Horiot, 325 100% 56% fractions in 7 fractions over None NO 1992129 5-year: 59% (HFX) vs. weeks 7 weeks 40% (SFX) (p=0.02) 2272 Abbreviations: OS: Overall survival; DFS: Disease free survival; AFX: Accelerated radiotherapy; Not reported; SFX: Standard 2273 fractionation; fx: Fractions; CRT: Conventional chemoradiotherapy; BID = twice per day. 2274

2275 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2276 Table 8. Studies exploring the role of ipsilateral-only radiation therapy in patients with oropharyngeal cancer squamous cell 2277 carcinoma, including years, primary site distribution, treatment, and TNM staging (Key Question 4)

Author and % OP Number of % % T- year Years of Cancers % % N-Category % Stage patients Definitiv % IMRT Category Treatment in the Chemo (#OP) e XRT Tonsil Nx: 0 Tx: 0 N0:58 I: 19 T1: 32 O’Sullivan N1:25 II: 33 1970-1991 228 (228) 100 100 0 0 T2: 52 et al110 N2a:12 III: 28 T3: 13 N2b:4 IV: 20 T4: 3 N3:1 Corvo et al147 2001-2002 30 (22) 59 68 0 53 ------Nx: 0 I: 0 Tx: 0 N0: 0 II: 0 T1: 55 Rusthoven et N1: 20 III: 20 2003-2007 20 (20) 100 30 45 95 T2: 35 al148 N2a: 15 IV: 80 T3: 10 N2b: 65 T4: 0 N3: 0 Nx: 0 Tx: 0 N0:69 T1: 21 N1:16 Kagei et al149 1989-1996 32 (32) 63 100 0 0 T2: 38 -- N2a:13 T3: 37 N2b:13 T4: 6 N3:3 Nx: 0 I: 5 Tx: 0 N0: 10 II: 45 T1: 35 N1: 40 III: 50 Koo et al150 2003-2011 20 (20) 100 30 30 30 T2: 60 N2a: 10 IV: 0 T3: 5 N2b: 40 T4: 0 N3: 0 Liu et al*151 1990-2002 58 (58) 100 100 0 -37 Tx: 2 Nx: 0 I: 10 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

T1: 15 N0: 43 II: 17 T2: 52 N1: 24 III: 40 T3: 29 N2a: 18 IV: 33 T4: 2 N2b: 7 N3: 9 Nx: 0 Tx: 0 N0:57 I: 13 T1: 21 N1:30 II: 24 Jackson et al152 1975-1993 178 (178) 100 100 0 0 T2: 45 N2a: 4 III: 46 T3: 29 N2b: 4 IV: 17 T4: 5 N3: 9 Nx: 0 Tx: 0 N0: 20 T1: 42 N1: 15 Lynch et al**154 1995-2011 136 (136) 100 43 0 -63 T2: 54 -- N2a: 23 T3: 4 N2b: 40 T4: 0 N3: 2 Tx: 17 Nx: 2 T1: 51 N0: 32 Chronowski et T2: 33 N1: 22 1970-2007 102 (102) 100 92 67 5 -- al155 T3: 0 N2a: 21 T4: 0 N2b: 22 N3: 0 Nx: 0 Tx: 0 N0: 50 T1: 27 Al-Mamgani et N1: 23 2000-2011 185 (185) 70 100 100 6 T2: 66 -- al156 N2a: 10 T3: 7 N2b: 17 T4: 0 N3: 0 2278 *Smoking status available for 53 patients with 38% current smokers and 17% never smokers; P16 available for 15 patients with 60% 2279 P16 positive 2280 **Smoking status available for 113 patients with 54% having >10 pack/years and 37% never smokers NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2281 Table 9. Studies exploring the role of ipsilateral-only radiation therapy in patients with oropharyngeal cancer with details of 2282 survival and contralateral nodal metastases (Key Question 4) 2283 Author and year % Grade 5yr % 5yr % 5yr % 5yr % Crude Isolated TNM of Those 3+ Overall Disease Local Regional Contralater (no other with Isolated Late Survival Specific Recurrence Recurrenc al Nodal site of Contralateral Toxicity Survival e Recurrence recurrence) Recurrence Risk (%) Contralater al Recurrence Risk (%)

T2N1 (SP-BOT O’Sullivan et al110 3.4 N/R 76 (3yr) 23 (3yr) 19 (3yr) 3.5 3 ext.)

Corvo et al147 N/R N/R N/R N/R N/R 6.7 6.7 T3N2b

80 (2yr Rusthoven et al148 0 80 (2yr) 21 (2yr) 0 0 0 N/A DFS)

Kagei et al149 3.1 64 (3yr) 79 (3yr) 26 19 0 0 N.A.

Koo et al150 0 95 95 (PFS) 5 0 0 0 N/A

83 (crude Liu et al*151 7 N/R 9 4 0 0 N/A DFS) NOT TO BE COPIED, DISSEMINATED OR REFERENCED

T2N0 T3N0 Jackson et al152 3.4 56 69 25 14 2.6 N/R T1-3N1 T1-3N1 T2N2b T2N2b (ECE) T2N2b (ECE) Lynch et al**154 9 89 86 2 (crude) 7 (crude) 5.9 4.4 T2N2b (ECE) T2N2b (ECE) T2N2b (ECE) Chronowski et al155 N/R 95 96 (DFS) 0 2 2 1 T2N0 (spm) T1N2b Al-Mamgani et al156 2.2 70 84 (DFS) 9 4 1.1 1.1 T2N0 2284 Abbreviations: YR: year; N/R: Not Reported; DFS: Disease-free survival; PFS: Progression-free survival; N/A: Not applicable; SP- 2285 BOT ext.: Soft palate and/or base of tongue extension; spm: most likely a second primary malignancy of contralateral base of tongue 2286 with nodal metastasis; ECE: Extra-capsular nodal extension

2287

2288

2289

2290

2291

2292

2293

2294 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2295 Table 10. Consensus statement for use of ipsilateral-only radiation therapy in patients with oropharyngeal squamous cell 2296 carcinoma (Key Question 4). 2297 Strongly Conditionally recommended recommended (if all of the (based on extension or N following) category) Primary Site Tonsil Tonsil Soft Palate involvement None <1cm T Category T1 or T2 T1 or T2 N Category N0 or N1 N2a Extra-Capsular Nodal Extension None None Human Papillomavirus Status Unknown impact Unknown impact 2298 NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2299 References:

2300 1. Gillison ML, Koch WM, Capone RB, et al. Evidence for a causal association between human 2301 papillomavirus and a subset of head and neck cancers. Journal of the National Cancer Institute. 2302 May 3 2000;92(9):709-720. 2303 2. Chaturvedi AK, Engels EA, Pfeiffer RM, et al. Human papillomavirus and rising oropharyngeal 2304 cancer incidence in the United States. Journal of clinical oncology : official journal of the 2305 American Society of Clinical Oncology. Nov 10 2011;29(32):4294-4301. 2306 3. Gillison ML, D'Souza G, Westra W, et al. Distinct risk factor profiles for human papillomavirus 2307 type 16-positive and human papillomavirus type 16-negative head and neck cancers. Journal of 2308 the National Cancer Institute. Mar 19 2008;100(6):407-420. 2309 4. Ang KK, Harris J, Wheeler R, et al. Human papillomavirus and survival of patients with 2310 oropharyngeal cancer. N Engl J Med. Jul 1 2010;363(1):24-35. 2311 5. Graham R, Mancher M, Wolman DM, Greenfield S, Steinberg E., eds. Clinical Practice Guidelines 2312 We Can Trust. Washington, DC: The National Academies Press;2011. 2313 6. Balshem H, Helfand M, Schunemann HJ, et al. GRADE guidelines: 3. Rating the quality of 2314 evidence. Journal of clinical epidemiology. Apr 2011;64(4):401-406. 2315 7. Andrews J, Guyatt G, Oxman AD, et al. GRADE guidelines: 14. Going from evidence to 2316 recommendations: the significance and presentation of recommendations. Journal of clinical 2317 epidemiology. Jul 2013;66(7):719-725. 2318 8. Loblaw DA, Prestrud AA, Somerfield MR, et al. American Society of Clinical Oncology Clinical 2319 Practice Guidelines: formal systematic review-based consensus methodology. J Clin Oncol. 2320 2012;30(25):3136-3140. 2321 9. Adelstein DJ, Li Y, Adams GL, et al. An intergroup phase III comparison of standard radiation 2322 therapy and two schedules of concurrent chemoradiotherapy in patients with unresectable 2323 squamous cell head and neck cancer. J Clin Oncol. Jan 1 2003;21(1):92-98. 2324 10. Calais G, Alfonsi M, Bardet E, et al. Randomized trial of radiation therapy versus concomitant 2325 chemotherapy and radiation therapy for advanced-stage oropharynx carcinoma. Journal of the 2326 National Cancer Institute. Dec 15 1999;91(24):2081-2086. 2327 11. Ruo Redda MG, Ragona R, Ricardi U, et al. Radiotherapy alone or with concomitant daily low- 2328 dose carboplatin in locally advanced, unresectable head and neck cancer: definitive results of a 2329 phase III study with a follow-up period of up to ten years. Tumori. Mar-Apr 2010;96(2):246-253. 2330 12. Zakotnik B, Smid L, Budihna M, et al. Concomitant radiotherapy with mitomycin C and 2331 bleomycin compared with radiotherapy alone in inoperable head and neck cancer: final report. 2332 International journal of radiation oncology, biology, physics. Jul 15 1998;41(5):1121-1127. 2333 13. Ghosh-Laskar S, Kalyani N, Gupta T, et al. Conventional radiotherapy versus concurrent 2334 chemoradiotherapy versus accelerated radiotherapy in locoregionally advanced carcinoma of 2335 head and neck: Results of a prospective randomized trial. Head & neck. Sep 15 2014. 2336 14. Zakotnik B, Smid L, Budihna M, et al. Concomitant radiotherapy with mitomycin C and 2337 bleomycin compared with radiotherapy alone in inoperable head and neck cancer: final report. 2338 Int J Radiat Oncol Biol Phys. Jul 15 1998;41(5):1121-1127. 2339 15. Olmi P, Crispino S, Fallai C, et al. Locoregionally advanced carcinoma of the oropharynx: 2340 conventional radiotherapy vs. accelerated hyperfractionated radiotherapy vs. concomitant 2341 radiotherapy and chemotherapy--a multicenter randomized trial. International journal of 2342 radiation oncology, biology, physics. Jan 1 2003;55(1):78-92. 2343 16. Denis F, Garaud P, Bardet E, et al. Final results of the 94-01 French Head and Neck Oncology and 2344 Radiotherapy Group randomized trial comparing radiotherapy alone with concomitant NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2345 radiochemotherapy in advanced-stage oropharynx carcinoma. J Clin Oncol. Jan 1 2004;22(1):69- 2346 76. 2347 17. Staar S, Rudat V, Stuetzer H, et al. Intensified hyperfractionated accelerated radiotherapy limits 2348 the additional benefit of simultaneous chemotherapy--results of a multicentric randomized 2349 German trial in advanced head-and-neck cancer. Int J Radiat Oncol Biol Phys. Aug 1 2350 2001;50(5):1161-1171. 2351 18. Ghadjar P, Simcock M, Studer G, et al. Concomitant cisplatin and hyperfractionated radiotherapy 2352 in locally advanced head and neck cancer: 10-year follow-up of a randomized phase III trial 2353 (SAKK 10/94). Int J Radiat Oncol Biol Phys. Feb 1 2012;82(2):524-531. 2354 19. Wendt TG, Grabenbauer GG, Rodel CM, et al. Simultaneous radiochemotherapy versus 2355 radiotherapy alone in advanced head and neck cancer: a randomized multicenter study. J Clin 2356 Oncol. Apr 1998;16(4):1318-1324. 2357 20. Bourhis J, Lapeyre M, Tortochaux J, et al. Accelerated radiotherapy and concomitant high dose 2358 chemotherapy in non resectable stage IV locally advanced HNSCC: results of a GORTEC 2359 randomized trial. Radiother Oncol. Jul 2011;100(1):56-61. 2360 21. Budach V, Stromberger C, Poettgen C, et al. Hyperfractionated accelerated radiation therapy 2361 (HART) of 70.6 Gy with concurrent 5-FU/Mitomycin C is superior to HART of 77.6 Gy alone in 2362 locally advanced head and neck cancer: long-term results of the ARO 95-06 randomized phase III 2363 trial. International journal of radiation oncology, biology, physics. Apr 1 2015;91(5):916-924. 2364 22. Budach V, Stuschke M, Budach W, et al. Hyperfractionated accelerated chemoradiation with 2365 concurrent fluorouracil-mitomycin is more effective than dose-escalated hyperfractionated 2366 accelerated radiation therapy alone in locally advanced head and neck cancer: final results of 2367 the radiotherapy cooperative clinical trials group of the German Cancer Society 95-06 2368 Prospective Randomized Trial. J Clin Oncol. Feb 20 2005;23(6):1125-1135. 2369 23. Bourhis J, Sire C, Graff P, et al. Concomitant chemoradiotherapy versus acceleration of 2370 radiotherapy with or without concomitant chemotherapy in locally advanced head and neck 2371 carcinoma (GORTEC 99-02): an open-label phase 3 randomised trial. The Lancet. Oncology. Feb 2372 2012;13(2):145-153. 2373 24. Adelstein DJ. Oropharyngeal cancer: the role of chemotherapy. Current treatment options in 2374 oncology. Feb 2003;4(1):3-13. 2375 25. Ghosh-Laskar S, Kalyani N, Gupta T, et al. Conventional radiotherapy versus concurrent 2376 chemoradiotherapy versus accelerated radiotherapy in locoregionally advanced carcinoma of 2377 head and neck: Results of a prospective randomized trial. Head Neck. Sep 15 2014. 2378 26. Denis F, Garaud P, Bardet E, et al. Late toxicity results of the GORTEC 94-01 randomized trial 2379 comparing radiotherapy with concomitant radiochemotherapy for advanced-stage oropharynx 2380 carcinoma: comparison of LENT/SOMA, RTOG/EORTC, and NCI-CTC scoring systems. 2381 International journal of radiation oncology, biology, physics. Jan 1 2003;55(1):93-98. 2382 27. Fallai C, Bolner A, Signor M, et al. Long-term results of conventional radiotherapy versus 2383 accelerated hyperfractionated radiotherapy versus concomitant radiotherapy and 2384 chemotherapy in locoregionally advanced carcinoma of the oropharynx. Tumori. 2006;92(1):41- 2385 54. 2386 28. Semrau R, Mueller RP, Stuetzer H, et al. Efficacy of intensified hyperfractionated and accelerated 2387 radiotherapy and concurrent chemotherapy with carboplatin and 5-fluorouracil: updated results 2388 of a randomized multicentric trial in advanced head-and-neck cancer. International journal of 2389 radiation oncology, biology, physics. Apr 1 2006;64(5):1308-1316. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2390 29. Blanchard P, Baujat B, Holostenco V, et al. Meta-analysis of chemotherapy in head and neck 2391 cancer (MACH-NC): a comprehensive analysis by tumour site. Radiotherapy and oncology : 2392 journal of the European Society for Therapeutic Radiology and Oncology. Jul 2011;100(1):33-40. 2393 30. Blanchard P, Hill C, Guihenneuc-Jouyaux C, Baey C, Bourhis J, Pignon JP. Mixed treatment 2394 comparison meta-analysis of altered fractionated radiotherapy and chemotherapy in head and 2395 neck cancer. Journal of clinical epidemiology. Sep 2011;64(9):985-992. 2396 31. Pignon JP, le Maitre A, Maillard E, Bourhis J. Meta-analysis of chemotherapy in head and neck 2397 cancer (MACH-NC): an update on 93 randomised trials and 17,346 patients. Radiotherapy and 2398 oncology : journal of the European Society for Therapeutic Radiology and Oncology. Jul 2399 2009;92(1):4-14. 2400 32. Xiao C, Hanlon A, Zhang Q, et al. Risk factors for clinician-reported symptom clusters in patients 2401 with advanced head and neck cancer in a phase 3 randomized clinical trial: RTOG 0129. Cancer. 2402 Mar 15 2014;120(6):848-854. 2403 33. Bourhis J, Calais G, Lapeyre M, et al. Concomitant radiochemotherapy or accelerated 2404 radiotherapy: analysis of two randomized trials of the French Head and Neck Cancer Group 2405 (GORTEC). Seminars in oncology. Dec 2004;31(6):822-826. 2406 34. Garden AS, Harris J, Vokes EE, et al. Preliminary results of Radiation Therapy Oncology Group 97- 2407 03: a randomized phase ii trial of concurrent radiation and chemotherapy for advanced 2408 squamous cell carcinomas of the head and neck. Journal of clinical oncology : official journal of 2409 the American Society of Clinical Oncology. Jul 15 2004;22(14):2856-2864. 2410 35. Ang KK, Zhang Q, Rosenthal DI, et al. Randomized phase III trial of concurrent accelerated 2411 radiation plus cisplatin with or without cetuximab for stage III to IV head and neck carcinoma: 2412 RTOG 0522. J Clin Oncol. Sep 20 2014;32(27):2940-2950. 2413 36. Martins RG, Parvathaneni U, Bauman JE, et al. Cisplatin and radiotherapy with or without 2414 erlotinib in locally advanced squamous cell carcinoma of the head and neck: a randomized 2415 phase II trial. Journal of clinical oncology : official journal of the American Society of Clinical 2416 Oncology. Apr 10 2013;31(11):1415-1421. 2417 37. Nguyen-Tan PF, Zhang Q, Ang KK, et al. Randomized Phase III Trial to Test Accelerated Versus 2418 Standard Fractionation in Combination With Concurrent Cisplatin for Head and Neck Carcinomas 2419 in the Radiation Therapy Oncology Group 0129 Trial: Long-Term Report of Efficacy and Toxicity. J 2420 Clin Oncol. Dec 1 2014;32(34):3858-3867. 2421 38. Alkureishi LW, de Bree R, Ross GL. RADPLAT: an alternative to surgery? The oncologist. May 2422 2006;11(5):469-480. 2423 39. Kumar P RK, Harris J, McCulloch T, Cmelak A, Sofferman R, Levine P, Fu K. Mature results of 2424 Radiation Therapy Oncology Group (RTOG) Trial 9615 using intra-arterial cisplatin (IA-P) and 2425 concurrent radiation therapy (RT) for stage IV-T4 head/neck (H/N) squamous cell carcinoma 2426 (SCCa). Journal of Clinical Oncology. 2005;Vol 23, No 16S (June 1 Supplement), 2005: 5579. 2427 40. Rasch CR, Hauptmann M, Schornagel J, et al. Intra-arterial versus intravenous chemoradiation 2428 for advanced head and neck cancer: Results of a randomized phase 3 trial. Cancer. May 1 2429 2010;116(9):2159-2165. 2430 41. Mesia R, Rueda A, Vera R, et al. Adjuvant therapy with cetuximab for locally advanced squamous 2431 cell carcinoma of the oropharynx: results from a randomized, phase II prospective trial. Annals 2432 of oncology : official journal of the European Society for Medical Oncology / ESMO. Feb 2433 2013;24(2):448-453. 2434 42. Lok BH, Setton J, Caria N, et al. Intensity-modulated radiation therapy in oropharyngeal 2435 carcinoma: effect of tumor volume on clinical outcomes. International journal of radiation 2436 oncology, biology, physics. Apr 1 2012;82(5):1851-1857. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2437 43. Garden AS, Asper JA, Morrison WH, et al. Is concurrent chemoradiation the treatment of choice 2438 for all patients with Stage III or IV head and neck carcinoma? Cancer. Mar 15 2004;100(6):1171- 2439 1178. 2440 44. Bernier J, Domenge C, Ozsahin M, et al. Postoperative irradiation with or without concomitant 2441 chemotherapy for locally advanced head and neck cancer. The New England journal of medicine. 2442 May 6 2004;350(19):1945-1952. 2443 45. Cooper JS, Pajak TF, Forastiere AA, et al. Postoperative concurrent radiotherapy and 2444 chemotherapy for high-risk squamous-cell carcinoma of the head and neck. The New England 2445 journal of medicine. May 6 2004;350(19):1937-1944. 2446 46. Cooper JS, Zhang Q, Pajak TF, et al. Long-term follow-up of the RTOG 9501/intergroup phase III 2447 trial: postoperative concurrent radiation therapy and chemotherapy in high-risk squamous cell 2448 carcinoma of the head and neck. International journal of radiation oncology, biology, physics. 2449 Dec 1 2012;84(5):1198-1205. 2450 47. Bernier J, Cooper JS, Pajak TF, et al. Defining risk levels in locally advanced head and neck 2451 cancers: a comparative analysis of concurrent postoperative radiation plus chemotherapy trials 2452 of the EORTC (#22931) and RTOG (# 9501). Head & neck. Oct 2005;27(10):843-850. 2453 48. Sinha P, Piccirillo JF, Kallogjeri D, Spitznagel EL, Haughey BH. The role of postoperative 2454 chemoradiation for oropharynx carcinoma: A critical appraisal of the published literature and 2455 National Comprehensive Cancer Network guidelines. Cancer. Jun 1 2015;121(11):1747-1754. 2456 49. Sinha P, Lewis JS, Jr., Piccirillo JF, Kallogjeri D, Haughey BH. Extracapsular spread and adjuvant 2457 therapy in human papillomavirus-related, p16-positive oropharyngeal carcinoma. Cancer. Jul 15 2458 2012;118(14):3519-3530. 2459 50. Maxwell JH, Ferris RL, Gooding W, et al. Extracapsular spread in head and neck carcinoma: 2460 impact of site and human papillomavirus status. Cancer. Sep 15 2013;119(18):3302-3308. 2461 51. Klozar J, Koslabova E, Kratochvil V, Salakova M, Tachezy R. Nodal status is not a prognostic factor 2462 in patients with HPV-positive oral/oropharyngeal tumors. Journal of surgical oncology. May 2463 2013;107(6):625-633. 2464 52. Iyer NG, Dogan S, Palmer F, et al. Detailed Analysis of Clinicopathologic Factors Demonstrate 2465 Distinct Difference in Outcome and Prognostic Factors Between Surgically Treated HPV-Positive 2466 and Negative Oropharyngeal Cancer. Annals of surgical oncology. Mar 24 2015. 2467 53. Lewis JS, Jr., Carpenter DH, Thorstad WL, Zhang Q, Haughey BH. Extracapsular extension is a 2468 poor predictor of disease recurrence in surgically treated oropharyngeal squamous cell 2469 carcinoma. Modern pathology : an official journal of the United States and Canadian Academy of 2470 Pathology, Inc. Nov 2011;24(11):1413-1420. 2471 54. Bachaud JM, Cohen-Jonathan E, Alzieu C, David JM, Serrano E, Daly-Schveitzer N. Combined 2472 postoperative radiotherapy and weekly cisplatin infusion for locally advanced head and neck 2473 carcinoma: final report of a randomized trial. Int J Radiat Oncol Biol Phys. Dec 1 1996;36(5):999- 2474 1004. 2475 55. Bachaud JM, David JM, Boussin G, Daly N. Combined postoperative radiotherapy and weekly 2476 cisplatin infusion for locally advanced squamous cell carcinoma of the head and neck: 2477 preliminary report of a randomized trial. International journal of radiation oncology, biology, 2478 physics. Feb 1991;20(2):243-246. 2479 56. Geiger JL, Lazim AF, Walsh FJ, et al. Adjuvant chemoradiation therapy with high-dose versus 2480 weekly cisplatin for resected, locally-advanced HPV/p16-positive and negative head and neck 2481 squamous cell carcinoma. Oral oncology. Apr 2014;50(4):311-318. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2482 57. Rewari AN, Haffty BG, Wilson LD, et al. Postoperative concurrent chemoradiotherapy with 2483 mitomycin in advanced squamous cell carcinoma of the head and neck: results from three 2484 prospective randomized trials. Cancer journal (Sudbury, Mass.). Mar-Apr 2006;12(2):123-129. 2485 58. Weissberg JB, Son YH, Papac RJ, et al. Randomized clinical trial of mitomycin C as an adjunct to 2486 radiotherapy in head and neck cancer. International journal of radiation oncology, biology, 2487 physics. Jul 1989;17(1):3-9. 2488 59. Haffty BG, Son YH, Sasaki CT, et al. Mitomycin C as an adjunct to postoperative radiation therapy 2489 in squamous cell carcinoma of the head and neck: results from two randomized clinical trials. 2490 International journal of radiation oncology, biology, physics. Sep 30 1993;27(2):241-250. 2491 60. Haffty BG, Son YH, Wilson LD, et al. Bioreductive alkylating agent porfiromycin in combination 2492 with radiation therapy for the management of squamous cell carcinoma of the head and neck. 2493 Radiation oncology investigations. 1997;5(5):235-245. 2494 61. Haffty BG, Wilson LD, Son YH, et al. Concurrent chemo-radiotherapy with mitomycin C 2495 compared with porfiromycin in squamous cell cancer of the head and neck: final results of a 2496 randomized clinical trial. International journal of radiation oncology, biology, physics. Jan 1 2497 2005;61(1):119-128. 2498 62. Smid L, Budihna M, Zakotnik B, et al. Postoperative concomitant irradiation and chemotherapy 2499 with mitomycin C and bleomycin for advanced head-and-neck carcinoma. International journal 2500 of radiation oncology, biology, physics. Jul 15 2003;56(4):1055-1062. 2501 63. Zakotnik B, Budihna M, Smid L, et al. Patterns of failure in patients with locally advanced head 2502 and neck cancer treated postoperatively with irradiation or concomitant irradiation with 2503 Mitomycin C and Bleomycin. International journal of radiation oncology, biology, physics. Mar 1 2504 2007;67(3):685-690. 2505 64. Racadot S, Mercier M, Dussart S, et al. Randomized clinical trial of post-operative radiotherapy 2506 versus concomitant carboplatin and radiotherapy for head and neck cancers with lymph node 2507 involvement. Radiotherapy and oncology : journal of the European Society for Therapeutic 2508 Radiology and Oncology. May 2008;87(2):164-172. 2509 65. Argiris A, Karamouzis MV, Johnson JT, et al. Long-term results of a phase III randomized trial of 2510 postoperative radiotherapy with or without carboplatin in patients with high-risk head and neck 2511 cancer. The Laryngoscope. Mar 2008;118(3):444-449. 2512 66. Tobias JS, Monson K, Gupta N, et al. Chemoradiotherapy for locally advanced head and neck 2513 cancer: 10-year follow-up of the UK Head and Neck (UKHAN1) trial. The Lancet. Oncology. Jan 2514 2010;11(1):66-74. 2515 67. Machtay Mitchell M. A Phase III Study of Postoperative Radiation Therapy (IMRT) +/- 2516 Cetuximab for Locally-Advanced Resected Head and Neck Cancer. 2517 68. Rosenthal DI, Harris J, Forastiere AA, et al. Early postoperative paclitaxel followed by concurrent 2518 paclitaxel and cisplatin with radiation therapy for patients with resected high-risk head and neck 2519 squamous cell carcinoma: report of the phase II trial RTOG 0024. Journal of clinical oncology : 2520 official journal of the American Society of Clinical Oncology. Oct 1 2009;27(28):4727-4732. 2521 69. Harari PM, Harris J, Kies MS, et al. Postoperative chemoradiotherapy and cetuximab for high-risk 2522 squamous cell carcinoma of the head and neck: Radiation Therapy Oncology Group RTOG-0234. 2523 Journal of clinical oncology : official journal of the American Society of Clinical Oncology. Aug 10 2524 2014;32(23):2486-2495. 2525 70. Harari PM, Paul M, Rosenthal David I. Randomized Phase II/III Trial of Surgery and Postoperative 2526 Radiation Delivered with Concurrent Cisplatin versus Docetaxel versus Docetaxel and Cetuximab 2527 for High-Risk Squamous Cell Cancer of the Head and Neck. 3/5/2015. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2528 71. Pignon JP, Bourhis J, Domenge C, Designe L. Chemotherapy added to locoregional treatment for 2529 head and neck squamous-cell carcinoma: three meta-analyses of updated individual data. 2530 MACH-NC Collaborative Group. Meta-Analysis of Chemotherapy on Head and Neck Cancer. 2531 Lancet. Mar 18 2000;355(9208):949-955. 2532 72. Laramore GE, Scott CB, al-Sarraf M, et al. Adjuvant chemotherapy for resectable squamous cell 2533 carcinomas of the head and neck: report on Intergroup Study 0034. International journal of 2534 radiation oncology, biology, physics. 1992;23(4):705-713. 2535 73. Langendijk JA, Slotman BJ, van der Waal I, Doornaert P, Berkof J, Leemans CR. Risk-group 2536 definition by recursive partitioning analysis of patients with squamous cell head and neck 2537 carcinoma treated with surgery and postoperative radiotherapy. Cancer. Oct 1 2538 2005;104(7):1408-1417. 2539 74. Ambrosch P, Kron M, Pradier O, Steiner W. Efficacy of selective neck dissection: a review of 503 2540 cases of elective and therapeutic treatment of the neck in squamous cell carcinoma of the upper 2541 aerodigestive tract. Otolaryngology--head and neck surgery : official journal of American 2542 Academy of Otolaryngology-Head and Neck Surgery. Feb 2001;124(2):180-187. 2543 75. Jackel MC, Ambrosch P, Christiansen H, Martin A, Steiner W. Value of postoperative 2544 radiotherapy in patients with pathologic N1 neck disease. Head & neck. Jul 2008;30(7):875-882. 2545 76. Shimizu K, Inoue H, Saitoh M, et al. Distribution and impact of lymph node metastases in 2546 oropharyngeal cancer. Acta oto-laryngologica. Aug 2006;126(8):872-877. 2547 77. McLaughlin MP, Mendenhall WM, Mancuso AA, et al. Retropharyngeal adenopathy as a 2548 predictor of outcome in squamous cell carcinoma of the head and neck. Head & neck. May-Jun 2549 1995;17(3):190-198. 2550 78. Weinstein GS, Quon H, Newman HJ, et al. Transoral robotic surgery alone for oropharyngeal 2551 cancer: an analysis of local control. Archives of otolaryngology--head & neck surgery. Jul 2552 2012;138(7):628-634. 2553 79. Kao J, Lavaf A, Teng MS, Huang D, Genden EM. Adjuvant radiotherapy and survival for patients 2554 with node-positive head and neck cancer: an analysis by primary site and nodal stage. 2555 International journal of radiation oncology, biology, physics. Jun 1 2008;71(2):362-370. 2556 80. Leemans CR, Tiwari R, Nauta JJ, van der Waal I, Snow GB. Recurrence at the primary site in head 2557 and neck cancer and the significance of neck lymph node metastases as a prognostic factor. 2558 Cancer. Jan 1 1994;73(1):187-190. 2559 81. Haughey BH, Sinha P. Prognostic factors and survival unique to surgically treated p16+ 2560 oropharyngeal cancer. The Laryngoscope. Sep 2012;122 Suppl 2:S13-33. 2561 82. Ravasz LA, Hordijk GJ, Slootweg PJ, Smit F, Tweel IV. Uni- and multivariate analysis of eight 2562 indications for post-operative radiotherapy and their significance for local-regional cure in 2563 advanced head and neck cancer. The Journal of laryngology and otology. May 1993;107(5):437- 2564 440. 2565 83. Bastit L, Blot E, Debourdeau P, Menard J, Bastit P, Le Fur R. Influence of the delay of adjuvant 2566 postoperative radiation therapy on relapse and survival in oropharyngeal and hypopharyngeal 2567 cancers. International journal of radiation oncology, biology, physics. Jan 1 2001;49(1):139-146. 2568 84. McMahon J, O'Brien CJ, Pathak I, et al. Influence of condition of surgical margins on local 2569 recurrence and disease-specific survival in oral and oropharyngeal cancer. The British journal of 2570 oral & maxillofacial surgery. Aug 2003;41(4):224-231. 2571 85. Fagan JJ, Collins B, Barnes L, D'Amico F, Myers EN, Johnson JT. Perineural invasion in squamous 2572 cell carcinoma of the head and neck. Archives of otolaryngology--head & neck surgery. Jun 2573 1998;124(6):637-640. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2574 86. Lanzer M, Gander T, Kruse A, Luebbers HT, Reinisch S. Influence of histopathologic factors on 2575 pattern of metastasis in squamous cell carcinoma of the head and neck. The Laryngoscope. May 2576 2014;124(5):E160-166. 2577 87. Liao CT, Chang JT, Wang HM, et al. Does adjuvant radiation therapy improve outcomes in pT1- 2578 3N0 oral cavity cancer with tumor-free margins and perineural invasion? International journal of 2579 radiation oncology, biology, physics. Jun 1 2008;71(2):371-376. 2580 88. Chinn SB, Spector ME, Bellile EL, et al. Impact of perineural invasion in the pathologically N0 2581 neck in oral cavity squamous cell carcinoma. Otolaryngology--head and neck surgery : official 2582 journal of American Academy of Otolaryngology-Head and Neck Surgery. Dec 2013;149(6):893- 2583 899. 2584 89. Ang KK, Trotti A, Brown BW, et al. Randomized trial addressing risk features and time factors of 2585 surgery plus radiotherapy in advanced head-and-neck cancer. International journal of radiation 2586 oncology, biology, physics. Nov 1 2001;51(3):571-578. 2587 90. Psychogios G, Mantsopoulos K, Kuenzel J, et al. Primary surgical treatment of T2 oropharyngeal 2588 carcinoma. Journal of surgical oncology. Jun 1 2012;105(7):719-723. 2589 91. Grant DG, Hinni ML, Salassa JR, Perry WC, Hayden RE, Casler JD. Oropharyngeal cancer: a case 2590 for single modality treatment with transoral laser microsurgery. Archives of otolaryngology-- 2591 head & neck surgery. Dec 2009;135(12):1225-1230. 2592 92. Alicandri-Ciufelli M, Bonali M, Piccinini A, et al. Surgical margins in head and neck squamous cell 2593 carcinoma: what is 'close'? European archives of oto-rhino-laryngology : official journal of the 2594 European Federation of Oto-Rhino-Laryngological Societies (EUFOS) : affiliated with the German 2595 Society for Oto-Rhino-Laryngology - Head and Neck Surgery. Sep 2013;270(10):2603-2609. 2596 93. Group ECO. Transoral Surgery Followed By Low-Dose or Standard-Dose Radiation Therapy With 2597 or Without Chemotherapy in Treating Patients With HPV Positive Stage III-IVA Oropharyngeal 2598 Cancer. 2599 94. Paccagnella A, Orlando A, Marchiori C, et al. Phase III trial of initial chemotherapy in stage III or 2600 IV head and neck cancers: a study by the Gruppo di Studio sui Tumori della Testa e del Collo. 2601 Journal of the National Cancer Institute. 1994;86(4):265-272. 2602 95. Zorat PL, Paccagnella A, Cavaniglia G, et al. Randomized phase III trial of neoadjuvant 2603 chemotherapy in head and neck cancer: 10-year follow-up. Journal of the National Cancer 2604 Institute. 2004;96(22):1714-1717. 2605 96. Domenge C, Hill C, Lefebvre JL, et al. Randomized trial of neoadjuvant chemotherapy in 2606 oropharyngeal carcinoma. French Groupe d'Etude des Tumeurs de la Tete et du Cou (GETTEC). 2607 Br J Cancer. 2000;83(12):1594-1598. 2608 97. Posner MR, Hershock DM, Blajman CR, et al. Cisplatin and fluorouracil alone or with docetaxel in 2609 head and neck cancer. N Engl J Med. 2007;357(17):1705-1715. 2610 98. Vermorken JB, Remenar E, van Herpen C, et al. Cisplatin, fluorouracil, and docetaxel in 2611 unresectable head and neck cancer. The New England journal of medicine. Oct 25 2612 2007;357(17):1695-1704. 2613 99. Lorch JH, Goloubeva O, Haddad RI, et al. Induction chemotherapy with cisplatin and fluorouracil 2614 alone or in combination with docetaxel in locally advanced squamous-cell cancer of the head 2615 and neck: long-term results of the TAX 324 randomised phase 3 trial. The Lancet. Oncology. 2616 2011;12(2):153-159. 2617 100. van Herpen CM, Mauer ME, Mesia R, et al. Short-term health-related quality of life and 2618 symptom control with docetaxel, cisplatin, 5-fluorouracil and cisplatin (TPF), 5-fluorouracil (PF) 2619 for induction in unresectable locoregionally advanced head and neck cancer patients (EORTC 2620 24971/TAX 323). British journal of cancer. Oct 12 2010;103(8):1173-1181. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2621 101. Gupta D, Shukla P, Bisht SS, et al. A prospective comparision of sequential chemoradiation vs 2622 concurrent chemoradiation in locally advanced oropharyngeal carcinomas. Cancer Biol Ther. 2623 2009;8(3):213-217. 2624 102. Haddad R, O'Neill A, Rabinowits G, et al. Induction chemotherapy followed by concurrent 2625 chemoradiotherapy (sequential chemoradiotherapy) versus concurrent chemoradiotherapy 2626 alone in locally advanced head and neck cancer (PARADIGM): a randomised phase 3 trial. The 2627 Lancet. Oncology. Mar 2013;14(3):257-264. 2628 103. Cohen EE, Karrison TG, Kocherginsky M, et al. Phase III randomized trial of induction 2629 chemotherapy in patients with N2 or N3 locally advanced head and neck cancer. J Clin Oncol. 2630 2014;32(25):2735-2743. 2631 104. Hitt R, Grau JJ, Lopez-Pousa A, et al. A randomized phase III trial comparing induction 2632 chemotherapy followed by chemoradiotherapy versus chemoradiotherapy alone as treatment 2633 of unresectable head and neck cancer. Ann Oncol. 2014;25(1):216-225. 2634 105. Brockstein B, Haraf DJ, Rademaker AW, et al. Patterns of failure, prognostic factors and survival 2635 in locoregionally advanced head and neck cancer treated with concomitant chemoradiotherapy: 2636 a 9-year, 337-patient, multi-institutional experience. Annals of oncology : official journal of the 2637 European Society for Medical Oncology / ESMO. Aug 2004;15(8):1179-1186. 2638 106. O'Sullivan B, Huang SH, Siu LL, et al. Deintensification candidate subgroups in human 2639 papillomavirus-related oropharyngeal cancer according to minimal risk of distant metastasis. 2640 Journal of clinical oncology : official journal of the American Society of Clinical Oncology. Feb 10 2641 2013;31(5):543-550. 2642 107. Huang SH, Xu W, Waldron J, et al. Refining American Joint Committee on Cancer/Union for 2643 International Cancer Control TNM stage and prognostic groups for human papillomavirus- 2644 related oropharyngeal carcinomas. Journal of clinical oncology : official journal of the American 2645 Society of Clinical Oncology. Mar 10 2015;33(8):836-845. 2646 108. Cmelak A, et al. Reduced-dose IMRT in human papilloma virus (HPV)-associated resectable 2647 oropharyngeal squamous carcinomas (OPSCC) after clinical complete response (cCR) to 2648 induction chemotherapy (IC). . Journal of Clinical Oncology. 2014;Vol 32, No 18_suppl 2649 109. Pignon JP, le Maitre A, Maillard E, Bourhis J. Meta-analysis of chemotherapy in head and neck 2650 cancer (MACH-NC): an update on 93 randomised trials and 17,346 patients. Radiother Oncol. 2651 2009;92(1):4-14. 2652 110. O'Sullivan B, Warde P, Grice B, et al. The benefits and pitfalls of ipsilateral radiotherapy in 2653 carcinoma of the tonsillar region. International journal of radiation oncology, biology, physics. 2654 Oct 1 2001;51(2):332-343. 2655 111. Withers HR, Peters LJ, Taylor JM, et al. Local control of carcinoma of the tonsil by radiation 2656 therapy: an analysis of patterns of fractionation in nine institutions. International journal of 2657 radiation oncology, biology, physics. Oct 15 1995;33(3):549-562. 2658 112. Beitler JJ, Zhang Q, Fu KK, et al. Final results of local-regional control and late toxicity of RTOG 2659 9003: a randomized trial of altered fractionation radiation for locally advanced head and neck 2660 cancer. Int J Radiat Oncol Biol Phys. 2014;89(1):13-20. 2661 113. Horiot JC, Bontemps P, van den Bogaert W, et al. Accelerated fractionation (AF) compared to 2662 conventional fractionation (CF) improves loco-regional control in the radiotherapy of advanced 2663 head and neck cancers: results of the EORTC 22851 randomized trial. Radiotherapy and 2664 oncology : journal of the European Society for Therapeutic Radiology and Oncology. Aug 2665 1997;44(2):111-121. 2666 114. Gregoire V, Ang K, Budach W, et al. Delineation of the neck node levels for head and neck 2667 tumors: a 2013 update. DAHANCA, EORTC, HKNPCSG, NCIC CTG, NCRI, RTOG, TROG consensus NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2668 guidelines. Radiotherapy and oncology : journal of the European Society for Therapeutic 2669 Radiology and Oncology. Jan 2014;110(1):172-181. 2670 115. Fletcher GH. Elective irradiation of subclinical disease in cancers of the head and neck. Cancer. 2671 Jun 1972;29(6):1450-1454. 2672 116. Overgaard J, Alsner J, Eriksen J, Horsman MR, Grau C. Importance of overall treatment time for 2673 the response to radiotherapy in patients with squamous cell carcinoma of the head and neck. 2674 Rays. Jul-Sep 2000;25(3):313-319. 2675 117. Ghoshal S, Goda JS, Mallick I, Kehwar TS, Sharma SC. Concomitant boost radiotherapy compared 2676 with conventional radiotherapy in squamous cell carcinoma of the head and neck--a phase III 2677 trial from a single institution in India. Clinical oncology (Royal College of Radiologists (Great 2678 Britain)). Apr 2008;20(3):212-220. 2679 118. Zackrisson B, Nilsson P, Kjellen E, et al. Two-year results from a Swedish study on conventional 2680 versus accelerated radiotherapy in head and neck squamous cell carcinoma--the ARTSCAN 2681 study. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology 2682 and Oncology. Jul 2011;100(1):41-48. 2683 119. Lambrecht M, Dirix P, Van den Bogaert W, Nuyts S. Incidence of isolated regional recurrence 2684 after definitive (chemo-) radiotherapy for head and neck squamous cell carcinoma. 2685 Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and 2686 Oncology. Dec 2009;93(3):498-502. 2687 120. Garden AS, Dong L, Morrison WH, et al. Patterns of disease recurrence following treatment of 2688 oropharyngeal cancer with intensity modulated radiation therapy. International journal of 2689 radiation oncology, biology, physics. Mar 15 2013;85(4):941-947. 2690 121. Duprez F, Bonte K, De Neve W, Boterberg T, De Gersem W, Madani I. Regional relapse after 2691 intensity-modulated radiotherapy for head-and-neck cancer. International journal of radiation 2692 oncology, biology, physics. Feb 1 2011;79(2):450-458. 2693 122. Kasibhatla M, Kirkpatrick JP, Brizel DM. How much radiation is the chemotherapy worth in 2694 advanced head and neck cancer? International journal of radiation oncology, biology, physics. 2695 Aug 1 2007;68(5):1491-1495. 2696 123. Nuyts S, Lambrecht M, Duprez F, et al. Reduction of the dose to the elective neck in head and 2697 neck squamous cell carcinoma, a randomized clinical trial using intensity modulated 2698 radiotherapy (IMRT). Dosimetrical analysis and effect on acute toxicity. Radiotherapy and 2699 oncology : journal of the European Society for Therapeutic Radiology and Oncology. Nov 2700 2013;109(2):323-329. 2701 124. Fu KK, Pajak TF, Trotti A, et al. A Radiation Therapy Oncology Group (RTOG) phase III randomized 2702 study to compare hyperfractionation and two variants of accelerated fractionation to standard 2703 fractionation radiotherapy for head and neck squamous cell carcinomas: first report of RTOG 2704 9003. International journal of radiation oncology, biology, physics. Aug 1 2000;48(1):7-16. 2705 125. Overgaard J, Hansen HS, Specht L, et al. Five compared with six fractions per week of 2706 conventional radiotherapy of squamous-cell carcinoma of head and neck: DAHANCA 6 and 7 2707 randomised controlled trial. Lancet. Sep 20 2003;362(9388):933-940. 2708 126. Overgaard J, Mohanti BK, Begum N, et al. Five versus six fractions of radiotherapy per week for 2709 squamous-cell carcinoma of the head and neck (IAEA-ACC study): a randomised, multicentre 2710 trial. The Lancet. Oncology. Jun 2010;11(6):553-560. 2711 127. Lassen P, Eriksen JG, Krogdahl A, et al. The influence of HPV-associated p16-expression on 2712 accelerated fractionated radiotherapy in head and neck cancer: evaluation of the randomised 2713 DAHANCA 6&7 trial. Radiother Oncol. 2011;100(1):49-55. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2714 128. Poulsen MG, Denham JW, Peters LJ, et al. A randomised trial of accelerated and conventional 2715 radiotherapy for stage III and IV squamous carcinoma of the head and neck: a Trans-Tasman 2716 Radiation Oncology Group Study. Radiotherapy and oncology : journal of the European Society 2717 for Therapeutic Radiology and Oncology. Aug 2001;60(2):113-122. 2718 129. Horiot JC, Le Fur R, N'Guyen T, et al. Hyperfractionation versus conventional fractionation in 2719 oropharyngeal carcinoma: final analysis of a randomized trial of the EORTC cooperative group of 2720 radiotherapy. Radiotherapy and oncology : journal of the European Society for Therapeutic 2721 Radiology and Oncology. Dec 1992;25(4):231-241. 2722 130. Bourhis J, Lapeyre M, Tortochaux J, et al. Phase III randomized trial of very accelerated radiation 2723 therapy compared with conventional radiation therapy in squamous cell head and neck cancer: 2724 a GORTEC trial. Journal of clinical oncology : official journal of the American Society of Clinical 2725 Oncology. Jun 20 2006;24(18):2873-2878. 2726 131. Baujat B, Bourhis J, Blanchard P, et al. Hyperfractionated or accelerated radiotherapy for head 2727 and neck cancer. The Cochrane database of systematic reviews. 2010(12):Cd002026. 2728 132. Setton J, Lee NY, Riaz N, et al. A multi-institution pooled analysis of gastrostomy tube 2729 dependence in patients with oropharyngeal cancer treated with definitive intensity-modulated 2730 radiotherapy. Cancer. Jan 15 2015;121(2):294-301. 2731 133. Eisbruch A, Harris J, Garden AS, et al. Multi-institutional trial of accelerated hypofractionated 2732 intensity-modulated radiation therapy for early-stage oropharyngeal cancer (RTOG 00-22). 2733 International journal of radiation oncology, biology, physics. Apr 2010;76(5):1333-1338. 2734 134. Brizel DM. Radiotherapy and concurrent chemotherapy for the treatment of locally advanced 2735 head and neck squamous cell carcinoma. Seminars in radiation oncology. Oct 1998;8(4):237-246. 2736 135. Bensadoun RJ, Benezery K, Dassonville O, et al. French multicenter phase III randomized study 2737 testing concurrent twice-a-day radiotherapy and cisplatin/5-fluorouracil chemotherapy (BiRCF) 2738 in unresectable pharyngeal carcinoma: Results at 2 years (FNCLCC-GORTEC). Int J Radiat Oncol 2739 Biol Phys. 2006;64(4):983-994. 2740 136. Huguenin P, Beer KT, Allal A, et al. Concomitant cisplatin significantly improves locoregional 2741 control in advanced head and neck cancers treated with hyperfractionated radiotherapy. J Clin 2742 Oncol. Dec 1 2004;22(23):4665-4673. 2743 137. Heijstek MW, Kamphuis S, Armbrust W, et al. Effects of the live attenuated measles-mumps- 2744 rubella booster vaccination on disease activity in patients with juvenile idiopathic arthritis: a 2745 randomized trial. Jama. Jun 19 2013;309(23):2449-2456. 2746 138. Peters LJ, Goepfert H, Ang KK, et al. Evaluation of the dose for postoperative radiation therapy 2747 of head and neck cancer: first report of a prospective randomized trial. Int J Radiat Oncol Biol 2748 Phys. 1993;26(1):3-11. 2749 139. Zelefsky MJ, Harrison LB, Fass DE, Armstrong JG, Shah JP, Strong EW. Postoperative radiation 2750 therapy for squamous cell carcinomas of the oral cavity and oropharynx: impact of therapy on 2751 patients with positive surgical margins. International journal of radiation oncology, biology, 2752 physics. Jan 1993;25(1):17-21. 2753 140. Sanguineti G, Richetti A, Bignardi M, et al. Accelerated versus conventional fractionated 2754 postoperative radiotherapy for advanced head and neck cancer: results of a multicenter Phase 2755 III study. International journal of radiation oncology, biology, physics. Mar 1 2005;61(3):762-771. 2756 141. Suwinski R, Bankowska-Wozniak M, Majewski W, et al. Randomized clinical trial on 7-days-a- 2757 week postoperative radiotherapy for high-risk squamous cell head and neck cancer. 2758 Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and 2759 Oncology. May 2008;87(2):155-163. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2760 142. Lukens JN, Lin A, Gamerman V, et al. Late consequential surgical bed soft tissue necrosis in 2761 advanced oropharyngeal squamous cell carcinomas treated with transoral robotic surgery and 2762 postoperative radiation therapy. International journal of radiation oncology, biology, physics. 2763 Aug 1 2014;89(5):981-988. 2764 143. Olzowy B, Tsalemchuk Y, Schotten KJ, Reichel O, Harreus U. Frequency of bilateral cervical 2765 metastases in oropharyngeal squamous cell carcinoma: a retrospective analysis of 352 cases 2766 after bilateral neck dissection. Head & neck. Feb 2011;33(2):239-243. 2767 144. Chung EJ, Oh JI, Choi KY, et al. Pattern of cervical lymph node metastasis in tonsil cancer: 2768 predictive factor analysis of contralateral and retropharyngeal lymph node metastasis. Oral 2769 oncology. Aug 2011;47(8):758-762. 2770 145. Lee DJ, Kwon MJ, Nam ES, et al. Histopathologic predictors of lymph node metastasis and 2771 prognosis in tonsillar squamous cell carcinoma. Korean journal of pathology. Jun 2772 2013;47(3):203-210. 2773 146. Lim YC, Lee SY, Lim JY, et al. Management of contralateral N0 neck in tonsillar squamous cell 2774 carcinoma. The Laryngoscope. Sep 2005;115(9):1672-1675. 2775 147. Corvo R, Foppiano F, Bacigalupo A, Berretta L, Benasso M, Vitale V. Contralateral parotid-sparing 2776 radiotherapy in patients with unilateral squamous cell carcinoma of the head and neck: 2777 technical methodology and preliminary results. Tumori. Jan-Feb 2004;90(1):66-72. 2778 148. Rusthoven KE, Raben D, Schneider C, Witt R, Sammons S, Raben A. Freedom from local and 2779 regional failure of contralateral neck with ipsilateral neck radiotherapy for node-positive tonsil 2780 cancer: results of a prospective management approach. International journal of radiation 2781 oncology, biology, physics. Aug 1 2009;74(5):1365-1370. 2782 149. Kagei K, Shirato H, Nishioka T, et al. Ipsilateral irradiation for carcinomas of tonsillar region and 2783 soft palate based on computed tomographic simulation. Radiotherapy and oncology : journal of 2784 the European Society for Therapeutic Radiology and Oncology. Feb 2000;54(2):117-121. 2785 150. Koo TR, Wu HG. Long-term results of ipsilateral radiotherapy for tonsil cancer. Radiation 2786 oncology journal. Jun 2013;31(2):66-71. 2787 151. Liu C, Dutu G, Peters LJ, Rischin D, Corry J. Tonsillar cancer: the Peter MacCallum experience 2788 with unilateral and bilateral irradiation. Head & neck. Mar 2014;36(3):317-322. 2789 152. Jackson SM, Hay JH, Flores AD, et al. Cancer of the tonsil: the results of ipsilateral radiation 2790 treatment. Radiotherapy and oncology : journal of the European Society for Therapeutic 2791 Radiology and Oncology. May 1999;51(2):123-128. 2792 153. O'Sullivan B, Huang SH, Perez-Ordonez B, et al. Outcomes of HPV-related oropharyngeal cancer 2793 patients treated by radiotherapy alone using altered fractionation. Radiotherapy and oncology : 2794 journal of the European Society for Therapeutic Radiology and Oncology. Apr 2012;103(1):49-56. 2795 154. Lynch J, Lal P, Schick U, et al. Multiple cervical lymph node involvement and extra-capsular 2796 extension predict for contralateral nodal recurrence after ipsilateral radiotherapy for squamous 2797 cell carcinoma of the tonsil. Oral oncology. Sep 2014;50(9):901-906. 2798 155. Chronowski GM, Garden AS, Morrison WH, et al. Unilateral radiotherapy for the treatment of 2799 tonsil cancer. International journal of radiation oncology, biology, physics. May 1 2800 2012;83(1):204-209. 2801 156. Al-Mamgani A, van Rooij P, Fransen D, Levendag P. Unilateral neck irradiation for well- 2802 lateralized oropharyngeal cancer. Radiotherapy and oncology : journal of the European Society 2803 for Therapeutic Radiology and Oncology. Jan 2013;106(1):69-73. 2804 157. Shoushtari AN, Meeneghan M, Treharne GC, et al. Clinical nodal staging of T1-2 tonsillar 2805 squamous cell carcinoma stratified by p16 status and implications for ipsilateral neck irradiation. 2806 Cancer journal (Sudbury, Mass.). May-Jun 2010;16(3):284-287. NOT TO BE COPIED, DISSEMINATED OR REFERENCED

2807 158. Olmi P, Crispino S, Fallai C, et al. Locoregionally advanced carcinoma of the oropharynx: 2808 conventional radiotherapy vs. accelerated hyperfractionated radiotherapy vs. concomitant 2809 radiotherapy and chemotherapy--a multicenter randomized trial. Int J Radiat Oncol Biol Phys. 2810 Jan 1 2003;55(1):78-92. 2811 159. Bonner JA, Harari PM, Giralt J, et al. Radiotherapy plus cetuximab for squamous-cell carcinoma 2812 of the head and neck. N Engl J Med. Feb 9 2006;354(6):567-578. 2813 160. Bonner JA, Harari PM, Giralt J, et al. Radiotherapy plus cetuximab for locoregionally advanced 2814 head and neck cancer: 5-year survival data from a phase 3 randomised trial, and relation 2815 between cetuximab-induced rash and survival. The Lancet. Oncology. Jan 2010;11(1):21-28. 2816 161. Lewin F, Damber L, Jonsson H, et al. Neoadjuvant chemotherapy with cisplatin and 5-fluorouracil 2817 in advanced squamous cell carcinoma of the head and neck: a randomized phase III study. 2818 Radiother Oncol. 1997;43(1):23-28. 2819 162. Hitt R, Grau JJ, Lopez-Pousa A, et al. A randomized phase III trial comparing induction 2820 chemotherapy followed by chemoradiotherapy versus chemoradiotherapy alone as treatment 2821 of unresectable head and neck cancer. Annals of oncology : official journal of the European 2822 Society for Medical Oncology / ESMO. Jan 2014;25(1):216-225. 2823

2824