Non‐Metastatic Squamous Cell Carcinoma of the Oropharynx: Primary Surgery or (Chemo)radiotherapy? Authors: Primož Strojan, Avraham Eisbruch, June Corry, Sandra Nuyts, Pim de Graaf, Luiz P Kowalski, Jonathan J Beitler, Sweet Ping Ng, William M Mendenhall, M P Sreeram, Karthik N Rao, Anna Luíza Damaceno Araújo, Alfio Ferlito Published: 2026 ABSTRACT ABSTRACTBackgroundIn patients with locally advanced, non‐metastatic oropharyngeal squamous cell carcinoma (OPSCC), surgery‐based and nonsurgical treatment approaches have not been adequately compared in randomized Phase III clinical trials. Current clinical guidelines recommend various surgical and nonsurgical strategies across disease stages for both p16‐negative and p16‐positive OPSCC, without clearly favoring a single approach.ObjectiveThis review summarizes the available evidence relevant to clinical decision‐making between primary surgical and primary nonsurgical treatment approaches in patients with non‐metastatic OPSCC.MethodsWe reviewed studies evaluating tumor control across different treatment modalities, with emphasis on the highest‐quality available evidence and post‐treatment quality of life (QoL). Outcomes were discussed in relation to disease extent (early‐stage vs. locally advanced) and treatment approaches. Emerging therapeutic strategies with the potential to improve the efficacy‐to‐toxicity ratio were also reviewed.ResultsIn the absence of Phase III randomized trials in non‐metastatic OPSCC, the most informative comparative data derive from studies employing propensity score matching to reduce selection bias. These analyses suggest that survival outcomes are largely comparable between surgery‐based and nonsurgical approaches in both p16‐positive and p16‐negative OPSCC. With respect to post‐treatment QoL, surgery alone is generally associated with the lowest toxicity burden, whereas surgery followed by adjuvant (chemo)radiotherapy and definitive (chemo)radiotherapy demonstrate broadly comparable, albeit modality‐specific, toxicity and QoL profiles. p16‐positive status is consistently associated with more favorable post‐treatment QoL. De‐escalation strategies in p16/HPV‐associated OPSCC, as well as emerging modalities such as immune checkpoint inhibitors, proton therapy, and adaptive radiotherapy, may influence future treatment paradigms.ConclusionsIn the absence of clear differences in oncologic efficacy between primary surgical and nonsurgical approaches, treatment selection should be guided primarily by anticipated toxicity and its impact on QoL, within the context of multidisciplinary assessment and patient preferences. The influence of treatment costs in the decision‐making process varies across healthcare systems, reflecting differences in the valuation and financing of health interventions. The role of emerging treatment strategies in OPSCC will need to be defined through prospective, disease‐specific clinical trials.Keywords: local control, oropharyngeal carcinoma, quality of life, radiotherapy, surgery, treatment decision Introduction Oropharyngeal squamous cell carcinoma (OPSCC) presents substantial management challenges. Oncological and functional outcomes following both primary surgical and nonsurgical treatments are dependent on disease stage, tumor subsite within the oropharynx, growth pattern (exophytic vs. infiltrative), and patient‐related factors such as sex, age, performance status, and comorbidities. Treatment decision‐making is further complicated by tumor classification based on its association with human papillomavirus (HPV) infection. The p16 protein status, an immunohistochemically determined surrogate marker of HPV tumor cell infection, is now incorporated into the UICC/AJCC TNM staging system (since the 8th edition), where it serves as a key stratification factor for prognosis [1]. Moreover, the global incidence of HPV‐related OPSCC is increasing, further underscoring the clinical significance of this issue [2]. In general, primary therapeutic options for non‐metastatic OPSCC can be broadly divided into surgery‐based and nonsurgical treatments [3, 4]. Within each option, various treatment scenarios exist, differing in technical execution, extent of intervention, sequencing of therapeutic modalities, treatment intensity, and other factors. However, the range of treatment approaches available within individual institutions is often constrained by local technological capabilities, funding systems for specific treatments—particularly systemic therapies—and the availability and development of supportive services. As a result, decisions made by multidisciplinary teams are based not only on patient and disease‐related considerations but also on institutional practices, resources, prior experience with specific treatment strategies, and the level of technical expertise [5]. The aim of this review is to summarize the available evidence and arguments relevant to decision‐making between primary surgical and primary nonsurgical treatment approaches in patients with non‐metastatic OPSCC. Comparison of the Tumor Control Existing clinical guidelines, such as those issued by the National Comprehensive Cancer Network and the European Society for Medical Oncology, recommend one or more surgical and nonsurgical treatment options for different stages of both p16‐negative and p16‐positive disease, without favoring any particular approach [3, 4]. This reflects the fact that, in patients with locally advanced, non‐metastatic OPSCC, these two treatment strategies have never been adequately compared in a large‐scale, prospective, randomized Phase III clinical trial. Consequently, a single optimal treatment approach for these tumors remains undefined, and responsibility for selecting a strategy rests with local multidisciplinary treatment teams. Their role is therefore not simply to choose between first‐ and second‐best alternatives, but to identify the most appropriate personalized treatment strategy for each patient. Achieving this requires judicious clinical judgment that incorporates patient and caregiver needs and preferences. In practice, in the absence of definitive evidence, decisions between surgical and organ‐preserving approaches should be based on an individualized assessment of multiple variables influencing outcomes, including patient‐ and disease‐related factors, as well as critical appraisal of published evidence on survival outcomes, treatment‐related toxicity, functional impairment across different modalities, and patient‐reported quality of life (QoL). The only published randomized comparison of treatment strategies for Stage III/IV non‐metastatic head and neck cancer (excluding the nasopharynx and salivary glands) was conducted in Singapore between 1996 and 2002 [6]. However, the study was terminated early due to poor patient accrual. Over a 5.5‐year recruitment period, only 119 of the planned 200 patients were enrolled; only three were HPV‐positive, and the treatment regime used is now considered outdated by current standards. No differences in 5‐year disease‐specific survival and overall survival rates were observed between patients treated with primary surgery and those treated with primary chemoradiation, either in the overall study population or within the OPSCC subgroup, which comprised only 21% of the enrolled patients [6, 7]. Higher‐quality evidence regarding the therapeutic potential of different treatment strategies is available only for patients with early‐stage OPSCC (Table 1). Nichols et al. conducted a Phase II randomized (1:1) ORATOR trial between 2012 and 2017, enrolling 68 patients with T1–T2, N0–N2 disease (≤ 4 cm, AJCC 7th edition), 88% of whom had p16‐positive tumors [8]. Patients were randomized to receive either radiotherapy (70 Gy, with concomitant chemotherapy for N1–N2 disease) or transoral robotic surgery (TORS) with neck dissection, with or without adjuvant (chemo)radiotherapy based on pathological findings. After a median follow‐up of 5.1 years, progression‐free survival and overall survival did not differ between the two groups [8, 9]. In the randomized Phase II ORATOR2 trial, Palma et al. compared surgery and radiotherapy as primary treatment modalities within a de‐escalation approach [10]. Of the 61 enrolled patients with AJCC 8th edition T1–2, N0–2 HPV‐related disease (out of the planned 140; accrual was prematurely halted because of excessive toxicity in the surgical arm), 31 were randomized to transoral surgery with neck dissection and 30 to the radiotherapy arm. A reduced dose of adjuvant radiotherapy (50 Gy) was administered to 78% of surgically treated patients, all without concurrent chemotherapy, whereas patients in the radiotherapy arm received 60 Gy, with 72% receiving concurrent cisplatin. After a median follow‐up of 3.7 years, 2‐year progression‐free survival (86% vs. 100%, p = 0.012) and overall survival (90% vs. 100%, p = 0.041) were significantly worse in the surgical arm [10, 11]. The results of another randomized study, the EORTC 1420 Phase 3 trial (“Best‐of‐trial”), are eagerly awaited, in part because of several important distinctions from the ORATOR study [12]. EORTC 1420 permits both transoral laser microsurgery (TLM) and conventional transoral surgery and restricts eligibility to N1 disease, a design that may reduce the need for treatment intensification. In contrast, ORATOR included patients with N2 tumors, resulting in the use of chemoradiotherapy in 24% and 72% of patients in the TORS and radiotherapy arms, respectively. In addition, EORTC 1420 includes also supraglottic and T1 N0 hypopharyngeal primaries, with stratification by tumor site to better define indications for each treatment modality. Surgical procedures in EORTC 1420 are less invasive and more consistent with European practice, including no mandatory tracheostomy, allowance for margin revision, and safeguards to avoid resection of T3 disease [12]. The study also incorporates rigorous quality assurance programs for both radiotherapy and surgery [12, 13]. TABLE 1Primary non‐surgical vs. surgical treatment of oropharyngeal cancer: Randomized trials.Authors (ref.), studyNo. of patientsInclusion criteriaCompared therapiesFUP, medianOutcomes (RT arm vs. S arm)Functional and QOLAdverse effectsSurvival Nichols et al. [8, 9] ORATOR Phase II 2012–2017 68 T1‐2 N0‐2 (≤ 4 cm) AJCC 7th ed. p16+, 88% RT arm, n = 34: RT 70 Gy, n = 9 cCRT, in N+, n = 23 Withdrew from the study, n = 2 S arm, n = 34: TORS+ND, n = 10 TORS+ND + RT 60 Gy b , n = 16 TORS+ND + cCRT 64 Gy b , n = 8 5.1 years MDADI, mean score, at 1 year: 86.9 vs. 80.1, p = 0.042 at 5 years: 85.3 vs. 84.6, p > 0.05 EORTC QLQ‐C30 and H&N35, p < 0.05: RT worse: dry mouth, p = 0.032 S worse: pain, p = 0.002 FOIS: 95% vs. 95% Grade 2–5: 91% vs. 97%, p = 0.61 RT more common: hearing loss neutropenia S more common: dysphagia, other pain Grades 3–5: 56% vs. 47%, p = 0.47 Grade 5: 0/34 vs. 1/34 OS, at 5 years: 84% vs. 85.1%, p = 0.92 PFS, at 5 years: 84% vs. 82.2%, p = 0.83 Palma et al. [10, 11] ORATOR2 Phase II 2018–2020 61/170 a T1‐2 N0‐2 AJCC 8th ed. HPV+, all Arm 1, n = 30: RT 60 Gy, n = 8 cCRT, in N1 > 3 cm and N2, n = 21 Not received protocol‐specified treatment, n = 1 Arm 2, n = 31: TOS + ND (n = 4) or TOS + ND + RT 50 or 60 Gy b , n = 23 Not received protocol‐specified treatment, n = 4 3.7 years MDADI, mean score, at 2 years: 89 vs. 83, p = 0.02 EORTC H&N3, p < 0.05: RT worse: taste, smell S worse: social eating, coughing, pain killer use Feeding tube in place beyond 1 year: none Grade 2–5: 70% vs. 77%, p = 0.51 RT more common: anorexia, dysgeusia Grade 5: 0/30 vs. 2/31 OS, at 2 years: 100% vs. 90%. p = 0.041 PFS, at 2 years: 100% vs. 86%, p = 0.012 Abbreviations: cCRT, concurrent chemoradiotherapy; FOIS, functional oral intake score; FUP, follow‐up; MDADI, MD Anderson Dysphagia Invetory; ND, neck dissection; OS, overall survival; PFS, progression‐free survival; QOL, quality of life; RT, radiotherapy; S, surgery; TORS, transoral robotic surgery. a Accrual was discontinued early due to toxicity in the surgical arm. b Based on pathology. Primary non‐surgical vs. surgical treatment of oropharyngeal cancer: Randomized trials. Nichols et al. [8, 9] ORATOR Phase II 2012–2017 T1‐2 N0‐2 (≤ 4 cm) AJCC 7th ed. p16+, 88% RT arm, n = 34: RT 70 Gy, n = 9 cCRT, in N+, n = 23 Withdrew from the study, n = 2 S arm, n = 34: TORS+ND, n = 10 TORS+ND + RT 60 Gy b , n = 16 TORS+ND + cCRT 64 Gy b , n = 8 MDADI, mean score, at 1 year: 86.9 vs. 80.1, p = 0.042 at 5 years: 85.3 vs. 84.6, p > 0.05 EORTC QLQ‐C30 and H&N35, p < 0.05: RT worse: dry mouth, p = 0.032 S worse: pain, p = 0.002 FOIS: 95% vs. 95% Grade 2–5: 91% vs. 97%, p = 0.61 RT more common: hearing loss neutropenia S more common: dysphagia, other pain Grades 3–5: 56% vs. 47%, p = 0.47 Grade 5: 0/34 vs. 1/34 OS, at 5 years: 84% vs. 85.1%, p = 0.92 PFS, at 5 years: 84% vs. 82.2%, p = 0.83 Palma et al. [10, 11] ORATOR2 Phase II 2018–2020 61/170 a T1‐2 N0‐2 AJCC 8th ed. HPV+, all Arm 1, n = 30: RT 60 Gy, n = 8 cCRT, in N1 > 3 cm and N2, n = 21 Not received protocol‐specified treatment, n = 1 Arm 2, n = 31: TOS + ND (n = 4) or TOS + ND + RT 50 or 60 Gy b , n = 23 Not received protocol‐specified treatment, n = 4 MDADI, mean score, at 2 years: 89 vs. 83, p = 0.02 EORTC H&N3, p < 0.05: RT worse: taste, smell S worse: social eating, coughing, pain killer use Feeding tube in place beyond 1 year: none Grade 2–5: 70% vs. 77%, p = 0.51 RT more common: anorexia, dysgeusia Grade 5: 0/30 vs. 2/31 OS, at 2 years: 100% vs. 90%. p = 0.041 PFS, at 2 years: 100% vs. 86%, p = 0.012 Abbreviations: cCRT, concurrent chemoradiotherapy; FOIS, functional oral intake score; FUP, follow‐up; MDADI, MD Anderson Dysphagia Invetory; ND, neck dissection; OS, overall survival; PFS, progression‐free survival; QOL, quality of life; RT, radiotherapy; S, surgery; TORS, transoral robotic surgery. Accrual was discontinued early due to toxicity in the surgical arm. Based on pathology. The remaining evidence from the literature is largely retrospective and therefore inherently prone to selection bias. Additionally, differences in the demographic characteristics of treated patient populations across healthcare settings and world regions, the accuracy of diagnostic procedures, access to modern medical equipment and systemic therapies, the experience and expertise of diagnostic and therapeutic teams, the availability of supportive care services, and social and healthcare policies within specific environments must all be considered. Along with methodological inconsistencies between studies, such as unmeasured confounders (e.g., comorbidities and socioeconomic status), incomplete information on tumor p16/HPV status, and the use of different editions of the TNM staging system, these factors contribute to inconclusive findings. Accordingly, the results of recent meta‐analyses should be interpreted in the context of these limitations [14, 15, 16, 17, 18]. To mitigate selection bias and balance covariates in analytical comparisons, several studies have employed propensity score matching to enhance the relevance and interpretability of their findings [19]. Although this approach cannot replace evidence from randomized Phase III studies, the results of these studies are more informative than unmatched retrospective comparisons. Overall, they suggest that survival outcomes in patients with OPSCC are likely comparable regardless of therapeutic approach, in both cohorts with p16/HPV‐positive tumors and those with p16/HPV‐negative tumors ([20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34], Table 2). However, substantial heterogeneity exists across these studies, including differences in sample sizes and disease stages evaluated (restricted to early T1–T2 tumors with or without nodal involvement versus all stages, including M1), as well as in the survival endpoints reported. Further variability arises from differences in HPV status assessment (p16 immunohistochemistry, in situ hybridization, or unknown status) and in the treatments compared, including open surgery, TORS, various radiotherapy techniques, post‐treatment management and follow‐up, and the indications for and type of systemic therapy. Nevertheless, apart from the ORATOR trial, studies using this analytical approach provide the most robust evidence currently available to evaluate the comparative effectiveness of the two treatment strategies in non‐metastatic OPSCC, although even the most sophisticated statistical methodologies cannot fully substitute for the evidentiary rigor of a well‐designed, large‐scale randomized trial. TABLE 2Propensity score–matched comparisons of primary surgery and primary (chemo)radiotherapy: A literature review.Author (ref.)Source periodNo of patientsStagep16/HPV statusCompared therapies a Measured outcomeFUP, medianResult (S vs. RT) a AllPSMAll patientsPSM cohortEarly‐stage tumorsKelly et al. [20] NCDB 2010–2012 1044822 T1‐2 N1‐2b, M0 Negative S vs. RT OS30 monthsNo differenceNo differenceBaliga et al. [21] NCDB 2010–2014 17 1501584 T1‐2 N0‐2b, M0 Positive, 77% Negative, 23% TORS vs. RTOS31.4 monthsn.s.HPV+ & HPV−, no differenceKelly et al. [22] NCDB 2010–2013 30632526 T1‐2 N1‐2, M0 PositiveS vs. RTOS32 monthsNo differenceNo differenceThakar et al. [25] Tertiary‐care center 2011–2019 14648 T1‐3 N0‐2b, M0 NegativeTORS/TOUSS/TOLM vs. RTDFS, OS57DFS/OS: S betterOS/DFS: no differenceZhou et al. [27] SEER 2010–2015 1004444 T1‐2 N0, M0 Positive, 58% Negative, 42% S vs. RTCSS, OS31 monthsHPV+ & HPV−: CSS/OS, no differenceHPV+ & HPV−: CSS/OS, no differenceAmin et al. [28] TriNetX platform 2002–2022 3352726 T1‐2 N all, M0 UnknownTORS vs. RTOSn.s.n.s.TORS better b Peng et al. [30] SEER 2010–2015 18771172 T1‐2 N0‐1, M0 Positive, 65% Negative, 35% S vs. RT CSM, nCSM 32 months HPV+: CSM, no difference nCSM, S better HPV−: CSM/nCSM, S better HPV+: CSM/nCSM, no difference HPV−: CSM/nCSM, S better Constantino et al. [33] NCDB 2010–2020 2566, ≥ 70 years 1094, ≥ 70 years T1‐2 N0‐3, M0 PositiveTOS vs. RTOS48.2 monthsTOS betterTOS betterAll stagesCulie et al. [23] 7 French centers 2009–2014 474180All, M0PositiveS vs. RTRFS, DSS, OS41.9 monthsRFS/DSS/OS: S better RFS: S better DSS/OS: no difference Culie et al. [24] 7 French centers 2009–2014 382250All, M0NegativeS vs. RTRFS, DSS, OS42.7 months RFS/DSS; S better OS: no difference RFS: S better DSS/OS: no difference Tsai et al. [26] Taiwan Cancer Registry 2019–2020 1180972III‐IVaUnknownS vs. RTL‐RPFS, DMFS, PFS, OS43.4 months L‐RPFS: no difference DMFS: RT better PFS/OS: S better OS: no difference c Lau et al. [29] São Paulo Cancer Center 2004–2014 68403818All, M0UnknownS vs. RTDFS, OSn.s.n.s. OS: S better c DFS: no difference Lee et al. [31] Taipei Veterans General Hospital 2011–2019 7646 T1‐4 N1‐3, M0 NegativeS vs. RTRRFS, DSS, OS S: 53.3 RT: 45.6 months RRFS: S better DSS/OS: no difference RRFS: SURG better DSS/OS: no difference Peng et al. [32] SEER 2010–2015 80574444All, M0 Positive, 75% Negative, 25% S vs. RTCSM, nCSM HPV+: 31 months HPV−: 27 months HPV+: CSM/nCMS; S better HPV−: CSM/nCSM, S better HPV+: CSM, S better nCSM, no difference HPV−: CSM/nCSM, S better Vu et al. [34] TriNetX platform n.s. 66 9504826All, including M1UnknownS vs. RTOS S: 28.5 RT: 30.3 months n.s.S better b Abbreviations: CSM, cancer specific mortality; CSS, cancer‐specific survival; DFS, disease‐free survival; DSS, disease‐specific survival; n.s., not specified; NCDB, National Cancer Database; nCSM, noncancer specific mortality; OS, overall survival; PFS, progression‐free survival; RFS, recurrence‐free survival; RRFS, regional recurrence‐free survival; RT, Radiotherapy; S, surgery; SEER, surveillance, epidemiology, and end results; TOLM, transoral laser microsurgery; TORS, transoral robotic surgery; TOS, transoral surgery; TOUSS, transoral ultrasonic surgery; TriNetX, US Collaborative Network. a Primary surgery with/without adjuvant (chemo)radiotherapy vs. primary (chemo)radiotherapy. b Cohorts were unmatched for cancer stage and HPV status. c Cohorts were unmatched for HPV status. Propensity score–matched comparisons of primary surgery and primary (chemo)radiotherapy: A literature review. NCDB 2010–2012 T1‐2 N1‐2b, M0 S vs. RT NCDB 2010–2014 T1‐2 N0‐2b, M0 Positive, 77% Negative, 23% NCDB 2010–2013 T1‐2 N1‐2, M0 Tertiary‐care center 2011–2019 T1‐3 N0‐2b, M0 SEER 2010–2015 T1‐2 N0, M0 Positive, 58% Negative, 42% TriNetX platform 2002–2022 T1‐2 N all, M0 SEER 2010–2015 T1‐2 N0‐1, M0 Positive, 65% Negative, 35% CSM, nCSM HPV+: CSM, no difference nCSM, S better HPV−: CSM/nCSM, S better HPV+: CSM/nCSM, no difference HPV−: CSM/nCSM, S better NCDB 2010–2020 2566, ≥ 70 years T1‐2 N0‐3, M0 7 French centers 2009–2014 RFS: S better DSS/OS: no difference 7 French centers 2009–2014 RFS/DSS; S better OS: no difference RFS: S better DSS/OS: no difference Taiwan Cancer Registry 2019–2020 L‐RPFS: no difference DMFS: RT better PFS/OS: S better São Paulo Cancer Center 2004–2014 OS: S better c DFS: no difference Taipei Veterans General Hospital 2011–2019 T1‐4 N1‐3, M0 S: 53.3 RT: 45.6 months RRFS: S better DSS/OS: no difference RRFS: SURG better DSS/OS: no difference SEER 2010–2015 Positive, 75% Negative, 25% HPV+: 31 months HPV−: 27 months HPV+: CSM/nCMS; S better HPV−: CSM/nCSM, S better HPV+: CSM, S better nCSM, no difference HPV−: CSM/nCSM, S better TriNetX platform n.s. S: 28.5 RT: 30.3 months Abbreviations: CSM, cancer specific mortality; CSS, cancer‐specific survival; DFS, disease‐free survival; DSS, disease‐specific survival; n.s., not specified; NCDB, National Cancer Database; nCSM, noncancer specific mortality; OS, overall survival; PFS, progression‐free survival; RFS, recurrence‐free survival; RRFS, regional recurrence‐free survival; RT, Radiotherapy; S, surgery; SEER, surveillance, epidemiology, and end results; TOLM, transoral laser microsurgery; TORS, transoral robotic surgery; TOS, transoral surgery; TOUSS, transoral ultrasonic surgery; TriNetX, US Collaborative Network. Primary surgery with/without adjuvant (chemo)radiotherapy vs. primary (chemo)radiotherapy. Cohorts were unmatched for cancer stage and HPV status. Cohorts were unmatched for HPV status. Quality of Life Treatment‐related adverse effects and post‐treatment QoL are central to therapeutic decision‐making. While surgery alone appears to confer the lowest toxicity burden, its use is limited to early‐stage tumors not requiring adjuvant therapy. In this context, selection of the surgical modality is critical, as TORS and non‐robotic transoral surgery yield similar survival in early‐stage OPSCC, whereas TORS is associated with fewer positive margins and reduced need for adjuvant therapy, potentially improving functional outcomes [35, 36, 37]. Although comparative analyses of TORS and TLM indicate that TLM for HPV‐positive OPSCC may result in moderately less impairment of swallowing function and QoL in the early postoperative period, surgeons generally prefer TORS because of its superior robotic visualization and access, greater margin clearance around the tumor, and ergonomic advantages that reduce tremor and enhance surgeon comfort [38]. In a retrospective analysis of 76 patients with HPV‐positive OPSCC staged cT0–2, N0–1, Xu et al. demonstrated that surgery alone (i.e., TORS) was associated with significantly better salivary, taste, and oral function, as well as less pain and fewer financial, oral/dental, and sexual problems compared with surgery followed by adjuvant therapy or definitive (chemo)radiotherapy [39]. The minimal impact of surgery alone on post‐treatment QoL was further supported by a prospective cohort study by the same authors [40] and by Kim et al. [41]. This observation underscores the need for improved patient selection strategies, as illustrated by the Phase II randomized ORATOR trial, in which only 10 of 34 patients (29%) in the TORS arm underwent single‐modality surgery, while the remainder required adjuvant therapy [8]. Another important observation from these studies was that, in patients with early‐stage tumors, surgery followed by adjuvant (chemo)radiotherapy and definitive (chemo)radiotherapy were generally associated with similar QoL outcomes [39, 40]. A comparable conclusion was reached in the ORATOR trial, which identified distinct toxicity and QoL profiles in certain domains between TORS (± adjuvant [chemo]radiotherapy) and (chemo)radiotherapy, with more xerostomia reported in the radiotherapy arm and more pain and dysphagia in the TORS arm, though demonstrated comparable long‐term swallowing and dietary outcomes [9]. Specifically, the ORATOR trial reported mean 1‐year MD Anderson Dysphagia Inventory (MDAI) total scores of 86.9 for radiotherapy versus 80.1 for TORS (p = 0.042), though these differences were not considered clinically meaningful and converged by 5‐years [8, 9]. However, in the Phase II randomized ORATOR2 study, patient accrual was halted early because of excessive toxicity in the surgical arm [11]. Two treatment‐related deaths occurred in this arm (7% Grade 5 toxic effect rate). Although the overall rate of Grade 2 or higher toxic effects was comparable between the surgical and radiotherapy arms (71% vs. 67%), the radiotherapy arm was associated with higher rates of anorexia and dysgeusia. Despite a 10 Gy lower radiotherapy dose in the surgical group (50 Gy in 78% of patients) than in the radiotherapy arm (60 Gy, with concurrent cisplatin in 72% of patients), 2‐year MDADI scores were similar between treatment arms (83 vs. 89, p = 0.11), with no clinically meaningful differences in other QoL domains except for worse coughing and greater weight loss in the surgical group. No patient in either arm required enteral feeding at 1 year [11]. Meta‐analyses by Campo et al. and Gupta et al., focusing on swallowing‐related QoL, further support the equivalence of up‐front TORS (± adjuvant therapy) and up‐front radiotherapy (± chemotherapy) with respect to functional outcomes in patients with T1–T2, N0–2 OPSCC; however, both treatment strategies were associated with impaired swallowing function [42, 43]. Importantly, among patients with early‐stage tumors, Kim et al. found that HPV status influences post‐treatment QoL, with HPV‐positive patients experiencing better QoL than HPV‐negative patients [41]. Higher tumor stages require more extensive surgical procedures, larger low‐ and high‐dose radiation targets, and more frequent concomitant administration of systemic therapy with radiotherapy, resulting in greater normal tissue toxicity. In this context, Plonowska et al. reported the highest symptom burden following surgery combined with adjuvant chemoradiotherapy, including greater difficulties with swallowing, social eating, and role and physical functioning compared with definitive chemoradiotherapy [40]. Moreover, a meta‐analysis by Quan et al. demonstrated no statistically significant difference in gastrostomy dependence at 12 months between surgery with adjuvant (chemo)radiotherapy and cisplatin‐based chemoradiotherapy (8.3% vs. 4.2%, p = 0.37), though poorer performance on selected patient‐reported measures of swallowing and overall function at 12 months post‐therapy was found after primary surgery [44]. Patients with HPV‐associated OPSCC typically exhibit higher pretreatment QoL scores than HPV‐negative patients [45, 46]. Although the HPV‐positive disease is associated with more favorable QoL outcomes overall [47, 48], chemoradiotherapy studies indicate a greater treatment‐related decline in QoL, with QoL recovery to baseline and re‐emergence of initial advantage within 6–12 months [45, 46, 49]. Importantly, post‐treatment QoL is influenced not only by treatment modality and HPV status but also by sociodemographic and clinicopathological characteristics [41]. In a retrospective analysis, Kim et al. identified older age, female sex, lower educational attainment and income, unemployment, fewer household members, advanced disease stage, shorter time since diagnosis, and disease recurrence as additional independent predictors of poorer QoL outcomes. Notably, among patients with advanced disease, QoL did not differ according to treatment modality or HPV status [41]. Emerging Treatment Options Over the past several years, the approximately two‐decade‐old treatment paradigm for OPSCC has undergone intensive modification efforts. This applies to both primary surgery‐based and primary nonsurgical treatments. De‐Escalation Studies in HPV‐Related OPSCC Patients with early‐stage HPV‐related OPSCC and no or minimal smoking history are typically younger, have a longer life expectancy, and experience a substantially more favorable prognosis than patients with HPV‐unrelated disease. Consequently, multiple clinical trials over the past decade have investigated treatment de‐escalation strategies in this population with the aim of reducing treatment‐related toxicity while maintaining a high cure rate [50]. These de‐escalation efforts have encompassed both surgical and definitive (chemo)radiotherapy‐based approaches. Surgical strategies aimed at toxicity reduction have evolved with the adoption of TORS, which enables resection of early‐stage OPSCC with significantly lower morbidity compared with traditional open surgical techniques, while facilitating safer achievement of tighter tumor margins. However, data from a high‐volume surgical center demonstrated that among 541 patients with T1–2 N1–2 OPSCC treated with TORS, only 17% were managed with surgery alone, whereas 46% required postoperative radiotherapy and 38% required postoperative chemoradiotherapy [51]. As discussed above, the use of trimodality therapy is associated with increased cumulative toxicity. These findings highlight the need to refine patient selection prior to surgery and support consideration of adjuvant treatment de‐escalation strategies following TORS in patients with favorable‐risk HPV‐positive OPSCC, including postoperative radiation dose reduction and omission of chemotherapy when appropriate. The Phase II randomized ECOG‐ACRIN 3311 trial demonstrated that, among patients classified as intermediate risk—defined by surgical margins < 3 mm, involvement of 2–4 pathologic lymph nodes, ≤ 1 mm extranodal extension, and/or the presence of perineural or lymphovascular invasion—postoperative radiation dose reduction to 50 Gy resulted in tumor control comparable to that achieved with the standard 60 Gy regimen [52]. Similarly, in a randomized trial conducted at the Mayo Clinic, patients were assigned either to standard‐of‐care adjuvant therapy (60–66 Gy with concurrent cisplatin) or to a de‐escalated regimen consisting of 30–36 Gy delivered twice daily over 2 weeks with concurrent docetaxel [53]. Both treatment arms achieved similarly high rates of locoregional control, while the more aggressive de‐escalation regimen was associated with significantly lower rates of chronic toxicity [53]. Beyond dose reduction, early de‐escalation trials have also explored omission of radiotherapy to the primary tumor bed in cases where the primary lesion was completely resected without adverse pathologic features, with radiation instead directed solely to the neck based on nodal involvement. This approach resulted in local recurrence in only one of 60 treated patients and was associated with excellent functional outcomes [54]. The administration of neoadjuvant docetaxel and cisplatin chemotherapy prior to TORS and neck dissection in patients with stage III and IVA tumors (AJCC 7th edition), with adjuvant therapy reserved for patients with unresectable positive margins, more than three pathological lymph nodes, or extracapsular invasion > 2 mm on postoperative histopathology (i.e., the NECTORS protocol), was associated with a lower rate of distant metastasis compared with a historical cohort treated with concurrent chemoradiotherapy [55]. Hazard ratios of 9.98 for distant recurrence and 3.21 for failure at any site were reported, together with excellent QoL outcomes in the NECTORS cohort [55, 56]. Recent Phase II–III studies investigating de‐escalation strategies in definitive (chemo)radiotherapy for HPV‐related OPSCC have evaluated multiple approaches, including reduced radiation doses to the primary tumor, omission of concurrent cisplatin in selected or all patients, substitution of cisplatin with cetuximab, reduction of prophylactically irradiated elective target volumes (e.g., unilateral radiotherapy), and decreases in both the extent and dose of radiation delivered to these elective regions [57]. Another strategy involves response‐adapted (chemo)radiotherapy following neoadjuvant chemotherapy [58]. Many of these studies reported excellent locoregional control rates and favorable functional outcomes when compared with historical controls [50]. However, some of the trials directly comparing certain de‐escalation approaches with standard‐of‐care definitive therapy have demonstrated inferior oncologic outcomes in the de‐escalated arms [59, 60, 61, 62]. Consequently, de‐escalated chemoradiotherapy for patients with good‐prognosis HPV‐positive OPSCC should not currently be adopted outside the context of rigorously designed and carefully conducted clinical trials. Immune Checkpoint Inhibitors Two pivotal randomized Phase 3 trials, KEYNOTE‐689 and NIVOPOSTOP, have recently reported positive outcomes for perioperative immunotherapy in patients with resectable stage III or IV head and neck cancer ([63, 64]; Table 3). In KEYNOTE‐689, 714 patients were randomized, of whom only 10.2% had OPSCC, with p16 positivity documented in 37% of these tumors [63]. The study found that induction therapy with two cycles of pembrolizumab before surgery, followed by risk‐adapted adjuvant (chemo)radiotherapy with three concomitant cycles of pembrolizumab and 12 additional maintenance cycles, resulted in significantly improved 3‐year event‐free survival compared with standard treatment consisting of surgery followed by (chemo)radiotherapy [63]. Similarly, the NIVOPOSTOP (GORTEC 2018‐01) study randomized 666 patients treated primarily with surgery; OPSCC accounted for only 17.6% of tumors, and the p16 positivity rate was 28.2%. Eligibility was restricted to patients with high‐risk disease. The addition of a single pre‐adjuvant cycle of nivolumab, followed by chemoradiotherapy with three concomitant cycles of nivolumab and six subsequent maintenance doses, significantly improved 3‐year disease‐free survival compared with chemoradiotherapy alone after surgery [64]. In KEYNOTE‐689, the therapeutic benefit was predominantly driven by a reduction in distant metastatic events, whereas in NIVOPOSTOP, fewer locoregional recurrences were observed in the experimental arm, likely reflecting differences in the baseline balance between distant and locoregional failure risk across the two study populations [63, 64, 65]. TABLE 3Neoadjuvant immune checkpoint inhibitors in operable locally advanced oropharyngeal cancer: Phase III clinical trials.Authors (ref.), studyNo. of patientsInclusion criteriaCompared therapiesTreatment compliance a Toxicity a Survival, at 36 monthsPathological response a EFS/DFSOS Uppaluri et al. [63] KEYNOTE‐689 Phase III 2018–2023 714 OP: 73 (10.2%) p16‐positive: 27/73 (37%) Stage III‐IVA OC, OP p16−, Lx, HP Stage III OP p16+: only if T4 N0‐2 ECOG PS 0–1 Control arm, n = 351 S → (cC)RT Study arm, n = 363: Pembrolizumab* (2 cycles) → S → (cC)RT+ pembrolizumab* (3 + 12 cycles) *200 mg IV every 3 weeks ≥ 1 pembrolizumab, neoadjuvant: 99.2% Pembrolizumab no. of cycles, median: 15 Surgery performed: 88.7% vs. 88.7% RT dose, median: 60 Gy vs. 66 Gy Concurrent cisplatin: 38.9% vs. 50.5% patients Grades 3–5: 44.6% vs. 42.9% p = 0.47 Grade 5: 1.1% vs. 0.3% Serious: 19.1 vs. 10.5% Immune‐mediated, Grades 3–5: 10% vs. 0.6% After neoadjuvant pebrolizumab: mPR, all: 9.4 vs. 0% pCR: 3% vs. 0% All: 57.6 vs. 46.4% p = 0.0008 CPS ≥ 1: 58.2 vs. 44.9% p = 0.003 CPS ≥ 10: 59.8 vs. 45.9% p = 0.004 All: 68.4 vs. 61.1% CPS ≥ 1: 69 vs. 60.2% CPS ≥ 10: 68.2 vs. 59.2% p = 0.04 Bourhis et al. [64] NIVOPOST‐OP Phase III 2018–2024 666 OP: 117 (17.6%) p16‐positive: 33/117 (28.2%) After macroscopic complete resection, high‐risk b Stage pIII‐IV OC, OP p16−, Lx, HP Stage pII OP p16+: only if pT3‐4 N1 and ≥ 20 pack‐years ECOG PS 0–1 Control arm, n = 334: S → cCRT Study arm, n = 332: S → nivolumab* (1 cycle) → cCRT+nivolumab (3** + 6*** cycles) *240 mg IV **360 mg IV every 3 weeks (with cCRT) ***480 mg IV every 4 weeks Nivolumab no. of cycles, median: 10 RT dose of 66 Gy: 85% vs. 89% Total cisplatin dose ≥ 200 mg/m2: 77% vs. 80% Grade 3–5: 74% vs. 69% Grade 5: 3% vs. 3% Serious: 33% vs. 19% Immune‐mediated, Grades 3–5: 21% vs. 0% n.a. 63.1 vs. 52.5% p = 0.034 Immature dataAbbreviations: cCRT, concurrent chemoradiotherapy; CPS, composite prognostic score; DFS, DISEASE‐free survival; EFS, Event‐free survival; HP, hypopharynx; Lx, larynx; mPR, major pathological response; n.a., not applicable; OC, oral cavity; OP, oropharynx; OS, overall survival; pCR, pathological complete response; PS, performance status. a Study arm vs. control arm. b Extracapsula r nodal extension, microscopically positive margins or a close margin of ≤ 1 mm, ≥ 4 cervical nodal involvements without extracapsular extension, and multiple perineural invasions. Neoadjuvant immune checkpoint inhibitors in operable locally advanced oropharyngeal cancer: Phase III clinical trials. Uppaluri et al. [63] KEYNOTE‐689 Phase III 2018–2023 714 OP: 73 (10.2%) p16‐positive: 27/73 (37%) Stage III‐IVA OC, OP p16−, Lx, HP Stage III OP p16+: only if T4 N0‐2 ECOG PS 0–1 Control arm, n = 351 S → (cC)RT Study arm, n = 363: Pembrolizumab* (2 cycles) → S → (cC)RT+ pembrolizumab* (3 + 12 cycles) *200 mg IV every 3 weeks ≥ 1 pembrolizumab, neoadjuvant: 99.2% Pembrolizumab no. of cycles, median: 15 Surgery performed: 88.7% vs. 88.7% RT dose, median: 60 Gy vs. 66 Gy Concurrent cisplatin: 38.9% vs. 50.5% patients Grades 3–5: 44.6% vs. 42.9% p = 0.47 Grade 5: 1.1% vs. 0.3% Serious: 19.1 vs. 10.5% Immune‐mediated, Grades 3–5: 10% vs. 0.6% After neoadjuvant pebrolizumab: mPR, all: 9.4 vs. 0% pCR: 3% vs. 0% All: 57.6 vs. 46.4% p = 0.0008 CPS ≥ 1: 58.2 vs. 44.9% p = 0.003 CPS ≥ 10: 59.8 vs. 45.9% p = 0.004 All: 68.4 vs. 61.1% CPS ≥ 1: 69 vs. 60.2% CPS ≥ 10: 68.2 vs. 59.2% p = 0.04 Bourhis et al. [64] NIVOPOST‐OP Phase III 2018–2024 666 OP: 117 (17.6%) p16‐positive: 33/117 (28.2%) After macroscopic complete resection, high‐risk b Stage pIII‐IV OC, OP p16−, Lx, HP Stage pII OP p16+: only if pT3‐4 N1 and ≥ 20 pack‐years ECOG PS 0–1 Control arm, n = 334: S → cCRT Study arm, n = 332: S → nivolumab* (1 cycle) → cCRT+nivolumab (3** + 6*** cycles) *240 mg IV **360 mg IV every 3 weeks (with cCRT) ***480 mg IV every 4 weeks Nivolumab no. of cycles, median: 10 RT dose of 66 Gy: 85% vs. 89% Total cisplatin dose ≥ 200 mg/m2: 77% vs. 80% Grade 3–5: 74% vs. 69% Grade 5: 3% vs. 3% Serious: 33% vs. 19% Immune‐mediated, Grades 3–5: 21% vs. 0% 63.1 vs. 52.5% p = 0.034 Abbreviations: cCRT, concurrent chemoradiotherapy; CPS, composite prognostic score; DFS, DISEASE‐free survival; EFS, Event‐free survival; HP, hypopharynx; Lx, larynx; mPR, major pathological response; n.a., not applicable; OC, oral cavity; OP, oropharynx; OS, overall survival; pCR, pathological complete response; PS, performance status. Study arm vs. control arm. Extracapsula r nodal extension, microscopically positive margins or a close margin of ≤ 1 mm, ≥ 4 cervical nodal involvements without extracapsular extension, and multiple perineural invasions. However, the limited representation of OPSCC, particularly p16‐positive disease, in both trials restricts robust subgroup analyses. Moreover, overall survival data are not yet mature, despite favorable emerging trends. Consequently, although both studies demonstrated a significant event‐free survival benefit with perioperative immunotherapy, caution is warranted when extrapolating these findings to the OPSCC population. Several design‐related concerns also merit consideration. Treatment adherence was suboptimal: in KEYNOTE‐689, only 68.3% of patients received adjuvant immunotherapy, and just 42.7% completed the full treatment protocol, with a median treatment duration of 9.1 months compared to 2.9 months in the control arm. Adverse events led to treatment discontinuation in 20.4% of patients receiving pembrolizumab, and serious adverse events occurred in 19.1%, compared to 10.5% in the control group [63, 65]. In NIVOPOSTOP, treatment‐related serious adverse events were reported in 33.0% of patients versus 19.3% in the control arm, likely reflecting the combined intensity of adjuvant cisplatin and immunotherapy [64, 65]. These findings highlight the need to reduce the intensity of adjuvant therapy by strengthening the neoadjuvant phase through the combination of immunotherapeutics with chemotherapy, which has already demonstrated efficacy in head and neck cancer [66]. Importantly, neither trial addressed key unresolved questions, including the optimal timing and duration of immune checkpoint inhibitor therapy or the criteria for response‐adapted de‐escalation of adjuvant treatment—strategies essential to improve treatment adherence, particularly as highlighted by KEYNOTE‐689. In addition, clinically relevant subgroups, such as patients with PD‐L1 composite prognostic score (CPS) < 1 tumors, elderly patients, and those with poor performance status, remain insufficiently explored. Finally, given the substantial financial consequences associated with these regimens, rigorous cost‐effectiveness analyses are imperative to support sustainable implementation in real‐world clinical practice [65]. Despite these inconsistencies and unanswered questions, a treatment strategy incorporating neoadjuvant pembrolizumab for patients with non‐metastatic, locally advanced (T1–2N1 or higher) p16‐negative OPSCC with CPS PD‐L1 ≥ 1 has already been incorporated into NCCN clinical practice guidelines [3]. Compared with the successful addition of immunotherapy to surgical approaches for advanced head and neck cancer, studies attempting to improve outcomes by combining definitive chemoradiotherapy with PD‐1 or PD‐L1 blockade in locoregionally advanced head and neck cancer—with about one‐third of the patient population having p16‐positive OPSCC—show disappointing results [67, 68, 69, 70, 71]. One explanation for this failure is the immunosuppressive effect of radiotherapy to neck lymph nodes delivered concurrently with immunotherapy, which may have diminished the efficacy of immune checkpoint inhibitors [72]. The positive results of similarly designed Phase 3 studies in nasopharyngeal carcinoma are likely due to its association with Epstein–Barr virus and consequent differences in tumor immunobiology, characterized by high PD‐L1 expression, an inflammatory tumor microenvironment, and dense immune cell infiltration—features known to increase sensitivity to immune checkpoint inhibitors [73, 74, 75]. To overcome this resistance in non‐nasopharyngeal tumors, which exhibit moderate immunesensitivity, ongoing trials are using induction and/or sequential immunotherapy before radiotherapy begins [76, 77]. Other Neoadjuvant Treatments To achieve tumor downstaging, improve margin clearance, and facilitate early eradication of micrometastases—thereby potentially improving functional outcomes and survival—various neoadjuvant approaches prior to surgery have been investigated. In a meta‐analysis, Pedroso et al. reviewed six prospective randomized clinical trials including a total of 739 patients, predominantly or exclusively with oral cavity carcinoma, comparing neoadjuvant chemotherapy with standard treatment alone [78]. Various chemotherapy regimens were used, most commonly cisplatin‐based combinations. No significant differences were observed between treatment arms in postoperative morbidity, and overall survival was likewise not substantially different [78]. With the advent of immunotherapy, various combinations of chemotherapeutic and immunotherapeutic agents have been evaluated in the neoadjuvant setting prior to surgery. Cao et al. analyzed 13 prospective single‐arm and retrospective studies including 458 patients, of whom only 17.5% had OPSCC [66]. A pooled pathological complete response rate of 37% at the primary site and a major pathological response rate of 61% were reported. In addition, the odds ratio for achieving a major pathological response was 2.09 in patients with CPS ≥ 20 compared with those with CPS < 20, while nodal pathological remission rates ranged from 26.2% to 86.7% across studies, and pooled disease‐free survival was 91% [66]. Using propensity score matching, Wu et al. reported that, in a cohort of 51 patients (33.3% with OPSCC), neoadjuvant chemoimmunotherapy was associated with shorter operative time, reduced postoperative hospitalization, lower intraoperative blood loss, fewer free flap reconstructions, and significantly better postoperative quality of life compared with upfront surgery [79]. A meta‐analysis by Zhu et al., comparing studies of neoadjuvant chemotherapy versus chemoimmunotherapy in predominantly operable tumors with a variable proportion of OPSCC, demonstrated promising short‐term efficacy, including higher pathological complete response and disease control rates, as well as comparable safety for neoadjuvant chemoimmunotherapy, although no significant differences in progression‐free or overall survival were observed [80]. To enhance tumor antigen release and presentation, promote T‐cell infiltration, and stimulate a systemic immune response, neoadjuvant immunotherapy has been combined with preoperative stereotactic body radiotherapy (SBRT) in Phase I/Ib and retrospective studies [81, 82, 83]. Irradiation of macroscopic disease was administered 3–6 weeks prior to surgery, with doses ranging from 2 × 6 Gy to 5 × 8 Gy. Major pathological response rates ranged from 60% to 86%, complete pathological response rates from 33.3% to 75%, and clinical‐to‐pathological downstaging rates from 83.3% to 90%. Notably, neoadjuvant immunoradiotherapy did not cause delays in surgical treatment or severe adverse events [81, 82, 83]. In the definitive chemoradiotherapy setting, the role of neoadjuvant taxane‐based chemotherapy has been extensively evaluated in several Phase III randomized trials and subsequent meta‐analyses, which did not demonstrate improvements in progression‐free survival or overall survival in patients with locally advanced head and neck carcinomas compared with chemoradiotherapy alone [84, 85]. In contrast, intensification of neoadjuvant chemotherapy through the addition of immunotherapy has shown more encouraging results in retrospective studies with variable proportions of OPSCC patients [86]. Recently, Rosenberg et al. reported results from the OPTIMA II non‐randomized Phase II clinical trial, which evaluated three cycles of neoadjuvant nivolumab plus carboplatin and paclitaxel followed by risk‐ and response‐adapted therapy and six cycles of adjuvant nivolumab in 77 patients with locoregionally advanced HPV‐positive OPSCC [87]. Only 9 of 19 low‐risk patients who achieved ≥ 50% tumor reduction per RECIST criteria underwent TORS, while the remainder received (chemo)radiotherapy. The study reported 2‐year progression‐free and overall survival rates of 90% and 91.4%, respectively, along with improved functional outcomes, including in patients with high‐risk disease [87]. Similarly, encouraging results were reported in the DEPEND Phase II non‐randomized clinical trial using the same neoadjuvant chemoimmunotherapy regimen in 36 patients with stage IVA/B HPV‐negative head and neck cancer, 19.4% of whom had OPSCC [76]. After neoadjuvant therapy, 53% of patients achieved ≥ 50% tumor shrinkage per RECIST 1.1 and received de‐escalated chemoradiotherapy to 66 Gy with omission of elective nodal volumes, while the remaining patients received standard‐dose chemoradiotherapy. After a median follow‐up of 20 months, no differences in progression‐free survival, overall survival, or locoregional control were observed between groups, but selectively reduced treatment‐related morbidity [76]. Proton Therapy In radiotherapy, several randomized trials are currently evaluating the comparative effectiveness of proton versus photon irradiation, either broadly in head and neck squamous cell carcinoma or specifically in OPSCC, primarily in locally advanced disease but also in early‐stage tumors [88]. To date, results from two randomized Phase 3 studies have been reported (Table 4). The first is a multicenter US study by Frank et al., which included 440 patients with stage III or IV OPSCC, 95% of whom had p16‐positive tumors [89]. The study was designed as a non‐inferiority comparison with a margin of 9 percentile points of intensity‐modulated proton therapy (IMPT) versus intensity‐modulated (photon) radiotherapy (IMRT), with induction chemotherapy and/or concurrent chemotherapy administered to 14% and all patients, respectively. At a median follow‐up of 3.2 years, no difference in 3‐year progression‐free survival was observed between the treatment arms; however, overall survival was significantly improved in the IMPT arm, showing 42% reduction of the hazard risk of death compared with IMRT. In addition, IMPT was associated with lower rates of severe (Grade ≥ 3) lymphopenia, dysphagia, xerostomia, and gastrostomy tube dependence at 60 days after treatment initiation. On the basis of reduced treatment‐related toxicity and improved overall survival, the authors proposed IMPT as a potential new standard‐of‐care option for patients with OPSCC [89]. Nevertheless, several limitations warrant consideration. Substantial treatment crossover occurred, with 23% of patients switching between IMPT and IMRT after randomization due to insurance authorization issues for proton therapy which may have introduced socioeconomic imbalance. In the United States, patients treated with proton therapy are more likely to have higher socioeconomic status and better baseline health [91]. Socioeconomic status is a prognostic factor in OPSCC and has been associated with a 10%–15% difference in overall survival, as well as differences in self‐reported late adverse events between patients with higher and lower socioeconomic status [92]. Moreover, gastrostomy tube dependence at 60 days after treatment initiation may be influenced by institutional practices regarding prophylactic versus reactive tube placement. Despite a median follow up time 3.2 years, no statistical analysis of late toxicity was reported although it could have contributed to the observed difference in overall survival 3 years after treatment completion, particularly given the absence of differences in disease control. Additional factors, including heterogeneity in radiation delivery systems across participating centers and potential differences in post‐progression salvage and palliative therapies, may also have influenced the observed overall survival outcome [93]. TABLE 4Proton therapy in locally advanced oropharyngeal cancer: Phase III clinical trials.Authors (ref.), studyNo. of patientsInclusion criteriaCompared therapiesToxicity a SurvivalPFS [89]/FLR [90]OS Frank et al. [89] Phase III 2013–2022 Non‐inferiority trial 440 p16+, 95% Stage III‐IVB AJCC 7th ed. ECOG PS 0–2 Study arm, n = 221: IMPT iCT, 11% cCT, 100% Crossover to IMRT b , 18% Control arm, n = 219: IMRT iCT, 12% cCT, 100% Crossover to IMPT b , 33% Grades 3–4: Lymphopenia, 76% vs. 89% Grade 3: Dysphagia, 34% vs. 49% Weight loss, 46% vs. 55% Xerostomia, 33% vs. 45% Gastrostomy‐tube dependence, at 60 days after treatment start: 26.8% vs. 40.2%, p = 0.018 and at 1 year: 0 vs. 4 patients Intention‐to‐treat, n = 440: At 3 years: 82.5% vs. 83% At 5 years: 81.3% vs. 76.2%, p = 0.005 Per protocol, n = 296 At 3 years: 82.7% vs. 83.9% At 5 years: 81.2% vs. 76.8%, p = 0.011 Intention‐to‐treat, n = 440: At 3 years: 92.2% vs. 90% At 5 years: 90.9% vs. 81%, p = 0.045 Per protocol, n = 296 At 3 years: 93.4% vs. 92% At 5 years: 91.8% vs. 82.2%, p = 0.057 Thompson et al. [90] TORPEdO Phase III 2020–2023 205 p16+, 96% p16+: T1‐2 N1‐2 (excluding T1 with a single ipsilateral node < 3 cm), T3‐4 N0‐2 p16−: T1N2, T2N1‐2, T3‐4N0‐2 WHO PS 0‐1 Study arm, n = 136: IMPT cCT, 100% Control arm, n = 69: IMRT cCT, 100% Gastrostomy‐tube dependance or Grade 3 weight loss, at 12 months: 18% vs. 6% p = 0.079 UW‐QoL mean physical composite score, at 12 months: 78.3 vs. 77.1, p = 0.56 MDADI, mean score, at 12 months: 79.5 vs. 79.7, p = 0.93 EORTC QoL‐C30 mean score, at 12 months: 76.3 vs. 74.7, p = 0.62 At 2 years: 94% vs. 97%, p = 0.24At 2 years: 94.6% vs. 95.3%, p = 0.47Abbreviations: cCT, concurrent chemotherapy; FLR, freedom from loco‐regional recurrence; IMPT, intensity‐modulated proton therapy; IMRT, intensity‐modulated (photon) radiation therapy; MDADI, MD Anderson Dysphagia Inventory; OS, overall survival; PFS, progression‐free survival; PS, performance status; UW‐QoL, University of Washington Quality of Life. a Study arm vs. control arm. b After randomization and before treatment due to proton therapy insurance denial in the IMPT cohort or proton therapy insurance approval in the IMRT cohort and subsequent patient refusal of IMRT. Proton therapy in locally advanced oropharyngeal cancer: Phase III clinical trials. Frank et al. [89] Phase III 2013–2022 Non‐inferiority trial 440 p16+, 95% Stage III‐IVB AJCC 7th ed. ECOG PS 0–2 Study arm, n = 221: IMPT iCT, 11% cCT, 100% Crossover to IMRT b , 18% Control arm, n = 219: IMRT iCT, 12% cCT, 100% Crossover to IMPT b , 33% Grades 3–4: Lymphopenia, 76% vs. 89% Grade 3: Dysphagia, 34% vs. 49% Weight loss, 46% vs. 55% Xerostomia, 33% vs. 45% Gastrostomy‐tube dependence, at 60 days after treatment start: 26.8% vs. 40.2%, p = 0.018 and at 1 year: 0 vs. 4 patients Intention‐to‐treat, n = 440: At 3 years: 82.5% vs. 83% At 5 years: 81.3% vs. 76.2%, p = 0.005 Per protocol, n = 296 At 3 years: 82.7% vs. 83.9% At 5 years: 81.2% vs. 76.8%, p = 0.011 Intention‐to‐treat, n = 440: At 3 years: 92.2% vs. 90% At 5 years: 90.9% vs. 81%, p = 0.045 Per protocol, n = 296 At 3 years: 93.4% vs. 92% At 5 years: 91.8% vs. 82.2%, p = 0.057 Thompson et al. [90] TORPEdO Phase III 2020–2023 205 p16+, 96% p16+: T1‐2 N1‐2 (excluding T1 with a single ipsilateral node < 3 cm), T3‐4 N0‐2 p16−: T1N2, T2N1‐2, T3‐4N0‐2 WHO PS 0‐1 Study arm, n = 136: IMPT cCT, 100% Control arm, n = 69: IMRT cCT, 100% Gastrostomy‐tube dependance or Grade 3 weight loss, at 12 months: 18% vs. 6% p = 0.079 UW‐QoL mean physical composite score, at 12 months: 78.3 vs. 77.1, p = 0.56 MDADI, mean score, at 12 months: 79.5 vs. 79.7, p = 0.93 EORTC QoL‐C30 mean score, at 12 months: 76.3 vs. 74.7, p = 0.62 Abbreviations: cCT, concurrent chemotherapy; FLR, freedom from loco‐regional recurrence; IMPT, intensity‐modulated proton therapy; IMRT, intensity‐modulated (photon) radiation therapy; MDADI, MD Anderson Dysphagia Inventory; OS, overall survival; PFS, progression‐free survival; PS, performance status; UW‐QoL, University of Washington Quality of Life. Study arm vs. control arm. After randomization and before treatment due to proton therapy insurance denial in the IMPT cohort or proton therapy insurance approval in the IMRT cohort and subsequent patient refusal of IMRT. The results of a second study were recently reported by Thomson et al. [90]. The multicenter UK TORPEdO trial enrolled 205 patients with locally advanced OPSCC requiring bilateral chemoradiotherapy and randomized them in a 2:1 ratio to receive either IMPT or IMRT. At 12 months after treatment, no differences were observed between the two groups across a range of clinician‐assessed and patient‐reported outcomes. After a median follow‐up of 28.3 months, there were also no differences in 2‐year locoregional control or overall survival. In the absence of measurable differences in efficacy and toxicity endpoints, the authors concluded that IMPT does not represent an alternative to contemporary high‐quality IMRT. Notably, 2‐year locoregional control rates were exceptionally high in both treatment arms (95% in IMPT and 97% in IMRT), substantially exceeding expectations based on historical outcomes in locally advanced OPSCC with adverse prognostic features (cT3–4: 48%, cN2c: 21%, ≥ 10 pack‐year smoking history: 31%), despite the high proportion (96%) of p16‐positive tumors in the treated cohort [90]. Taken together, the findings of both the US trial and the TORPEdO study underscore the need for longer‐term follow‐up analyses of late toxicity, dose–response relationships and survival data, as well as for additional randomized evidence. Along with forthcoming data from the Danish DAHANCA35 trial and the Swedish randomized Phase 2 trial, planned to complete enrolment in coming year(s), these results will help to more robustly define the role of IMPT in the management of OPSCC [88, 93]. Adaptive Radiotherapy Another potential strategy for treatment optimization is incorporating adaptive radiotherapy (ART) into standard radiotherapy protocols. However, available evidence does not demonstrate a clear clinical benefit. A Phase III randomized trial in patients with stage III–IVB OPSCC showed no reduction in xerostomia or improvement in patient‐reported outcomes, toxicity, or survival with weekly ART compared with standard IMRT, except for improved parotid gland excretory function [94]. Similarly, preliminary data on daily ART delivered using a 1.5‐T MR‐linac showed no significant reductions in Grade ≥ 2 acute toxicity, xerostomia, or dysgeusia in the early post‐treatment period compared with a matched cohort treated with helical TomoTherapy, although lower hospitalization rates during treatment and higher performance status 3 months after radiotherapy were observed in the ART group, with comparable local control [95]. Consistent with these findings, a systematic review by Lindegaard et al. concluded that while ART may provide modest dosimetric advantages, its impact on toxicity and disease control remains uncertain [96]. Finally, the ART strategy was not found to be cost‐effective compared with standard IMRT in locally advanced oropharyngeal cancer [97]. Conclusion Based on the currently available evidence, largely retrospective and limited by institutional bias, heterogenous reporting practices, and variable scoring systems, and acknowledging the obvious absence of high‐quality data from large randomized trials, there appears to be no clinically meaningful difference in effectiveness between primary surgical and primary non‐surgical treatment approaches. This conclusion applies particularly to p16/HPV‐positive tumors and likely also to p16/HPV‐negative disease. Consequently, treatment selection for an individual patient should continue to be guided by multidisciplinary team assessment, integrating clinical, radiological, and histopathological factors, as well as careful estimation of the anticipated toxicities associated with each approach and their potential impact on QoL. In this context, particularly for patients with locally advanced disease requiring more intensive multimodal and therefore more toxic treatment, assessment of the availability and capacity of supportive services is also important. In addition, the patient's socioeconomic circumstances, preference and expectations—both oncologic and functional—remain critical considerations and should be incorporated in an equitable, shared decision‐making process involving the patient and their family. The influence of treatment costs on these decisions may vary substantially across countries, reflecting differences in the valuation of health interventions and in the structure and financing of national health systems [98, 99, 100]. At present, treatment of early‐stage tumors, whether HPV‐related or HPV‐unrelated, should generally be limited to single‐modality therapy (surgery or radiotherapy) to minimize unnecessary toxicity, underscoring the importance of refined pretreatment patient selection. In contrast, advanced‐stage disease typically requires combined‐modality treatment, with careful consideration of the full spectrum of clinical, oncological, and patient‐related factors discussed above when determining the optimal therapeutic approach. The extent to which emerging de‐escalation strategies and novel treatment modalities, such as immunotherapy, proton radiotherapy, and adaptive radiotherapy, will reshape current treatment paradigms remains to be determined and will become clear over time as these approaches are systematically evaluated in disease‐specific clinical trials and real‐world OPSCC settings. Ethics Statement The authors have nothing to report. Strojan P., Eisbruch A., Corry J., et al., “Non‐Metastatic Squamous Cell Carcinoma of the Oropharynx: Primary Surgery or (Chemo)radiotherapy?,” Head & Neck 48, no. 9 (2026): 2541–2556, 10.1002/hed.70356. This paper was written by members of the International Head and Neck Scientific Group (Padua). Data Availability Statement Data sharing not applicable to this article as no datasets were generated or analysed during the current study. References Associated Data Data Availability Statement Data sharing not applicable to this article as no datasets were generated or analysed during the current study.