Randomized Controlled Trials Comparing the Standard of Care to Alternative Treatments in Patients with Resectable Locally Advanced Head and Neck Squamous Cell Carcinoma: A Systematic Literature Review

Abstract

Introduction

Standard of care (SoC) for resectable locally advanced head and neck squamous cell carcinoma (LA-HNSCC) is surgery with adjuvant radiotherapy (RT) or chemoradiotherapy (for tumors at high risk of recurrence). As long-term prognosis is suboptimal with SoC, this study aimed to summarize findings from recent studies comparing alternative treatments to SoC.

Methods

A broad systematic literature review (search date: December 1, 2025) searched Embase, MEDLINE, and CENTRAL to identify randomized controlled trials evaluating surgery with RT and/or systemic treatments in the neoadjuvant and/or adjuvant setting in LA-HNSCC. Trials published since 2004 were included in this report if they compared interventions to SoC in terms of event-free survival (EFS) and overall survival (OS) or reported pathological response following neoadjuvant therapy.

Results

Fifty-six trials were included in the broader review, of which 25 were included in this report. Trials reporting EFS counted at least recurrence/progression and death as events, except one that did not provide a definition. Improvement in EFS (regardless of definition) was reported with the addition of perioperative pembrolizumab or addition of perioperative camrelizumab with neoadjuvant nab-paclitaxel + carboplatin to surgery + RT ± cisplatin, addition of adjuvant cetuximab or adjuvant cisplatin to surgery + RT, and addition of adjuvant nivolumab to surgery + RT + cisplatin. EFS with other interventions was comparable to SoC. OS was generally comparable between interventions and SoC. Across treatment arms with at least one neoadjuvant chemotherapy agent, pathological complete response rates ranged from 10.5% to 42.6% while it ranged from 0 to 12.9% in treatment arms without any neoadjuvant chemotherapy agent. Some trials reported an association between achieving pathological response and improved survival outcomes in exploratory analyses.

Conclusion

Immunotherapy shows promise as an addition to the SoC in LA-HNSCC. OS results are generally comparable and may still be immature in some trials. Evidence on pathological response suggests potential prognostic value for long-term survival outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12325-026-03604-5.

Plain Language Summary

People with head and neck squamous cell carcinoma that is locally advanced but still able to be removed with surgery are usually treated with surgery followed by radiotherapy. If the tumor has a high risk of coming back, doctors also add chemotherapy to the radiotherapy. Because long-term outcomes remain disappointing with this usual treatment (the standard of care), we reviewed recent randomized controlled trials that tested other options. We searched major databases up to December 1, 2025 to identify trials published since 2004 that compared new approaches with the standard of care and reported event-free survival (time until the cancer returns, worsens, or the patient dies) and overall survival (time alive), or reported tumor response following treatments given before surgery. Our review identified 25 trials that were suitable for inclusion in this report. Some treatments improved event-free survival, including pembrolizumab and camrelizumab (both help the immune system attack cancer) given before and after surgery with radiotherapy and chemotherapy, cetuximab or cisplatin given after surgery with radiotherapy, and nivolumab given after surgery with radiotherapy and chemotherapy. Overall survival was similar between the tested treatments and the standard of care. When chemotherapy was given before surgery, complete disappearance of tumor cells in the removed tissue occurred more often (10.5–42.6%) than when it was not (0–12.9%). Some studies also suggested that better pathological response may result in better long-term outcomes. Overall, immunotherapy appears to be a promising addition to the standard of care, and pathological response may help predict long-term survival.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12325-026-03604-5.

Key Summary Points

Why carry out this study?

The standard of care for locally advanced resectable head and neck squamous cell carcinoma (LA-HNSCC) is a risk-adapted approach consisting of surgery followed by adjuvant radiotherapy (surgery + RT), with adjuvant chemotherapy added for tumors at higher risk of recurrence.

Since long-term outcomes remain suboptimal with the current standard of care, novel systemic treatments such as immunotherapy and targeted therapy are being investigated in the neoadjuvant and/or adjuvant setting.

This systematic literature review was conducted to provide an up-to-date understanding of how novel treatments compare to the current standard of care in terms of event-free survival (EFS) and overall survival (OS) in patients with LA-HNSCC.

What was learned from the study?

Improvement in EFS (or analogous outcomes) was reported with the addition of perioperative pembrolizumab or perioperative camrelizumab with neoadjuvant nab-paclitaxel + carboplatin to surgery + RT ± cisplatin, addition of adjuvant cetuximab or adjuvant cisplatin to surgery + RT, as well as addition of adjuvant nivolumab to surgery + RT + cisplatin, while OS data were occasionally immature and generally comparable between interventions and the standard of care.

Immunotherapy with inhibitors of programmed cell death protein 1 has emerged as a promising addition to the standard of care in patients with LA-HNSCC, while most other evaluated interventions have not demonstrated consistent EFS or OS benefits.

Ongoing and future randomized controlled trials will be critical to refining the role of immunotherapy and immunochemotherapy combination regimens in this population.

Introduction

Head and neck squamous cell carcinoma (HNSCC) is the seventh most common malignancy globally, representing 4.5% of all diagnosed cancer cases [1, 2]. In the USA, an estimated 71,100 new cases were diagnosed with cancer of oral cavity, pharynx, or larynx in 2024, with over 16,000 estimated deaths in this population [3]. These patients are often diagnosed in the locoregionally advanced (LA) stages, defined as stage III, IVa, or IVb by the eighth edition of American Joint Committee on Cancer (AJCC) [4, 5]. For tumors located outside the oropharynx and for human papillomavirus (HPV)-negative tumors in the oropharynx, LA-HNSCC is defined as any T3–T4 or N2a–N3 tumors (with no distant metastasis) [6, 7]. For HPV-positive lesions in the oropharynx, which are associated with more favorable prognosis, LA-HNSCC can be defined as any T4 or N3 tumors (with no distant metastasis) [6, 7].

Management of patients with LA-HNSCC is planned by multidisciplinary committees including surgeons, radiation oncologists, medical oncologists, and other specialists to ensure all aspects of patient care are considered [8]. The standard of care (SoC) for patients with resectable tumors without high risk factors for locoregional recurrence (e.g., extranodal extension [ENE], positive surgical margins [PSM]) is surgery followed by adjuvant radiotherapy (RT). For tumors that are at high risk of postoperative recurrence, clinical practice guidelines recommend adding concomitant cisplatin chemotherapy to RT [9, 10]. This approach was established based on the results of the phase III EORTC 22931 (NCT00002555) and RTOG 9501 (NCT00002670) trials, published in 2004, which showed that adjuvant chemoradiation therapy improved locoregional control and survival-related outcomes compared to adjuvant RT alone in patients with resectable HNSCC at high risk of recurrence [11, 12]. A retrospective exploratory meta-analysis, published in 2005, further confirmed that patients with ENE and/or PSM who were enrolled in these trials achieved improved survival outcomes with the addition of adjuvant cisplatin to surgery + RT [13]. Lastly, for patients with non-nasopharyngeal LA-HNSCC who are ineligible for postoperative cisplatin, adjuvant docetaxel or docetaxel + cetuximab (in the presence of ENE and/or PSM) has been recommended by the National Comprehensive Cancer Network^® (NCCN) as treatment options to be concurrently administered with adjuvant RT [14] (NCCN makes no warranties of any kind whatsoever regarding their content, use, or application and disclaims any responsibility for their application or use in any way).

Although induction chemotherapy prior to surgery (i.e., in the neoadjuvant setting) can be beneficial in certain situations (e.g., for laryngeal preservation), it is usually associated with increased toxicity and lower patient compliance [15, 16]. As such, alternative interventions such as immune checkpoint inhibitors have been evaluated in the neoadjuvant setting in this population. Recent trials evaluating inhibitors of programmed cell death protein 1 in the neoadjuvant and/or adjuvant settings have demonstrated favorable absolute outcomes in terms of response and survival, with an acceptable safety profile in this population [17–20]. Moreover, when compared to SoC in randomized controlled trials, regimens containing these agents have also shown improved relative treatment effects [21, 22].

With growing evidence for novel systemic interventions in the treatment of patients with resectable LA-HNSCC, it is of interest to understand how these interventions improve survival-related outcomes in this population when compared to SoC. Furthermore, with pathological response outcomes following neoadjuvant treatment increasingly recognized as potential indicators of treatment efficacy, it is also of interest to describe these outcomes in trials where patients received neoadjuvant systemic therapy prior to surgery. To that end, a comprehensive systematic literature review of recent studies was conducted to achieve two complementary objectives: (1) systematically identify and qualitatively summarize randomized evidence on event-free survival (EFS) and overall survival (OS) for perioperative, neoadjuvant, and adjuvant systemic regimens relative to the current SoC; and (2) qualitative synthesize available data on pathological response following neoadjuvant systemic therapy. This review aims to provide clinically actionable insights into both the comparative efficacy of systemic therapies and the evolving role of pathological response as an important clinical endpoint in LA-HNSCC.

Methods

The methods used to conduct the systematic review were consistent with guidelines from the Preferred Reporting Items of Systematic Review and Meta-analyses (PRISMA) 2020 statement and Cochrane Handbook for Systematic Reviews [23, 24]. A protocol was not prospectively registered for this systematic review. This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Study Eligibility

Searches were conducted on December 1, 2025 to identify a broad set of randomized controlled trials evaluating various combinations of curative surgery accompanied by RT and/or systemic treatments in the neoadjuvant and/or adjuvant setting in patients with LA-HNSCC. The study eligibility criteria of the systematic review were pre-specified in terms of population, interventions, comparators, outcomes, time, and study designs (PICOTS) of interest (Supplementary Appendix A). The population of interest was newly diagnosed adult patients with LA-HNSCC that were eligible for surgery. Since tumor staging criteria were expected to vary slightly across the different editions of the AJCC guidelines, no restriction was applied to the population criteria in terms of the disease T and N stages and HPV/p16 status, as long as patients did not have distant metastasis (i.e., were M0) at trial entry. Interventions of interest were curative surgical resection of tumors with adjuvant therapy (RT and/or systemic treatment), neoadjuvant therapy (RT and/or systemic treatment), or both. Comparator treatments included any treatment meeting the aforementioned intervention criteria as well as surgery alone. Phase II and phase III randomized controlled trials published in English were eligible.

Of the studies identified in the broader systematic review, two clinically relevant subsets of trials were included in this report. The first subset of studies was randomized controlled trials comparing their experimental intervention arms to SoC, defined as surgery + RT, surgery + RT + cisplatin, surgery + RT + docetaxel, or surgery + RT + docetaxel + cetuximab. Experimental regimens of interest included at least one systemic treatment. For this subset of trials, the focus was studies that reported comparisons (i.e., hazard ratios) of EFS (defined as time from randomization to disease progression/recurrence or death due to any cause) or OS (defined as time from randomization to death due to any cause). Since definitions (and labels) for EFS were expected to vary across the trials, analogous time-to-event outcomes (e.g., disease-free survival) were also of interest if they considered both disease progression (or recurrence) and death due to any cause as events in their definitions. If a trial did not provide any definition for EFS, it was assumed to have included these events in the definition and was therefore eligible for inclusion. Trials’ original endpoint label and definition were retained as they were reported in their respective publications. Differences in event components (e.g., inclusion of second primary tumors) were described and flagged as sources of heterogeneity. The second subset of studies was trials reporting pathological response (major pathological response [mPR] and/or pathological complete response [pCR]) following neoadjuvant therapy.

The year 2004 marks a major inflection point in the management of resectable LA-HNSCC, as the EORTC 22931 and RTOG 9501 trials (both published in 2004) demonstrated that adding concurrent cisplatin to adjuvant RT improves outcomes in patients with high-risk tumor features. These findings established the contemporary SoC and shaped current clinical guidelines. To maintain relevance to modern practice and reduce heterogeneity arising from outdated treatment paradigms, only trials published in or after 2004 were included.

Databases Searched

Embase, MEDLINE, and Cochrane Central Registry of Controlled Trials (CENTRAL) were searched via the Ovid platform. All search strategies included a combination of free-text terms and medical subject headings (MeSH) terms as appropriate for each database (Supplementary Appendix B). Recent proceedings of relevant conferences were systematically searched, using similar population terms, in the Northern Light Life Sciences Conference Abstracts Database (Supplementary Appendix B) on the Ovid platform. Furthermore, manual searches of clinicaltrials.gov and clinicaltrialsregister.eu were conducted, using the term “head and neck squamous cell carcinoma”.

Study Selection

Two investigators working independently reviewed all titles and abstracts identified in the literature search. Citations deemed eligible for inclusion by both investigators were advanced to the full-text screening stage. The same investigators independently assessed the eligibility of the full-text articles for final inclusion in the systematic review. Following reconciliation of disagreements between the two investigators, a third investigator was included to reach consensus on any remaining conflicts.

Data Extraction and Quality Assessment

The two investigators independently extracted data for the final list of included studies. Any discrepancies between the data extracted were resolved by involving the third investigator and reaching a consensus. In a similar fashion, the investigators independently assessed the risk of bias of the included studies using the Cochrane Collaboration’s Risk of Bias tool (Version 2) [25]. Data were extracted for key trial characteristics, population characteristics, and outcomes.

Results

Study Selection

A total of 30,884 citations were identified through searches of Embase, MEDLINE, and CENTRAL. After removal of the 3,598 duplicates, 27,286 citations were screened, leading to the exclusion of a further 27,058 citations. Of the 228 citations that were reviewed at the full-text screening stage, 64 full-text publications were included. An additional 40 citations were included via searches of conference proceedings and other gray literature sources. This resulted in a total of 104 citations, representing 56 unique randomized controlled trials, which were included in the broader systematic review. Of these, 25 trials met the additional eligibility criteria of the current report and have been described herein (study selection PRISMA flow diagram presented in Fig. 1) [11–13, 21, 22, 26–50]. These trials are summarized in Tables 1 and 2 in terms of trial characteristics, evaluated interventions, baseline patient characteristics, and reported outcomes and their definitions. A summary of the remaining 31 trials is presented in Supplementary Appendix C.

Fig. 1.

Study selection PRISMA flow diagram. PRISMA Preferred Reporting Items for Systematic reviews and Meta-Analyses, SLR systematic literature review

Table 1.

Summary of trials comparing survival outcomes with alternative regimens versus standard of care

Trial	Trial characteristics	Treatment arms	N	Population characteristicsa	Tumor location and HPV status (%)a,b	High-risk features, %	Median follow-up duration, months	Overall survival HR (95% CI)	Event-free survival HR (95% CI), event definition
Trials evaluating perioperative treatments
He et al., 2025 (CAMORAL) (ChiCTR2000037980) [26]	Phase II Open-label China (multicenter)	Camrelizumab + nab-paclitaxel + carboplatin + surgery + RT + camrelizumab ± cisplatin	63	Age: 54.2 (mean) (SD 10.6) Male: 73 White: – Current/former smoker: – ECOG 0/1:100	OC: 79.4 OP: 19 HP/L: 1.6 HPV: 4.8	–	27.5	0.25; nominal p value < 0.05	EFS 0.34 (0.17–0.68) Progression, recurrence, or death
		Surgery + RT ± cisplatin	62	Age: 58.4 (mean) (SD 12.5) Male: 85.5 White: – Current/former smoker: – ECOG 0/1: 100	OC: 85.5 OP: 14.5 HP/L: 0 HPV: 3.2	–	22.1	Reference	Reference
Uppaluri et al., 2025 (KEYNOTE-689) (NCT03765918) [22]	Phase III Open-label Multinational	Pembrolizumab + surgery + pembrolizumab + RT ± cisplatin	363	Age: 60 (29–82) Male: 78.8 White: 78.2 Current/former smoker: 80.7 ECOG 0/1:100	OC: 60.3 OP: 9.6 HP: 7.7 L: 22.3 HPV: 3.3	ENE, PSM < 1 mm: 32.5	30	0.76 (0.59–0.98)	EFS 0.73 (0.58–0.92) Progression, recurrence, or death
		Surgery + RT ± cisplatin	351	Age: 61 (22–87) Male: 78.9 White: 76.9 Current/former smoker: 76.1 ECOG 0/1: 100	OC: 60.7 OP: 10.8 HP: 7.4 L: 20.8 HPV: 4.3	ENE, PSM < 1 mm: 44.4	23.4	Reference	Reference
Trials evaluating neoadjuvant treatments
Chaukar et al., 2022[29]	Phase II Open-label India (single center)	Docetaxel + cisplatin + 5-FU + surgery + RT ± cisplatin	34	Age: 46 (27–62) Male: 97.1 White: – Current/former smoker: 79.4 ECOG 0/1: 94.1	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE: 40 PNI: 13.3 VE: 6.6	49.2	0.90 (0.51–1.59)	DFS 0.91 (0.52–1.61) Recurrence or death
		Surgery + RT ± cisplatin	34	Age: 49.5 (27–68) Male: 85.3 White: – Current/former smoker: 85.3 ECOG 0/1: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE: 41.1 PNI: 17.6 VE: 0	40.8	Reference	Reference
Talor et al., 2022 (IT-MATTERS) (NCT01265849) [30–32]	Phase III Open-label Multinational	Leukocyte interleukin + cyclophosphamide + indomethacin + zinc + surgery + RT ± cisplatin	395	Age: 56.5 (mean) Male: 79 White: 78.7 Current/former smoker: 79.2 KPS > 70: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE, PSM, ≥ 2 metastatic lymph nodes: 50.6	56	1.09 (0.89–1.32)	PFS 1.09 (0.9–1.31) Recurrence, new disease above clavicle, distant metastases or death
		Leukocyte interleukin + surgery + RT ± cisplatin	134	Age: 55.9 (mean) Male: 78.4 White: 80.6 Current/former smoker: 82.8 KPS > 70: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE, PSM, ≥ 2 metastatic lymph nodes: 51.5	55.6	1.07 (0.81–1.42)	PFS 1.1 (0.84–1.43) Recurrence, new disease above clavicle, distant metastases or death
		Surgery + RT ± cisplatin	394	Age: 57 (mean) Male: 79.9 White: 80.5 Current/former smoker: 78.2 KPS > 70: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE, PSM, ≥ 2 metastatic lymph nodes: 50.3	55.9	Reference	Reference
Zhong et al., 2013 (NCT01542931) [33]	Phase III Open-label China (single center)	Docetaxel + cisplatin + 5-FU + surgery + RT	128	Age: 55 (29–74) Male: 71.1 White: – Current/former smoker: 53.9 KPS > 60: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE: 13.2	30c	0.98 (0.63–1.51)	DFS 0.97 (0.65–1.45) Recurrence or death
		Surgery + RT	128	Age: 56 (26–75) Male: 68.8 White: – Current/former smoker: 44.5 KPS > 60: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE: 16.5	30c	Reference	Reference
Zhong, 2022 (EAGLE) (NCT01434394) [34]	Phase III Open-label China (multicenter)	Cetuximab + docetaxel + cisplatin + surgery + RT	134	Age: – Male: – White: – Current/former smoker: – ECOG 0/1: –	OC/OP: 100 HP: 0 L: 0 HPV: –	–	100c	1.09 (0.79–1.52)	DFS 1.12 (0.82–1.52) –
		Surgery + RT	136	Age: – Male: – White: – Current/former smoker: – ECOG 0/1: –	OC/OP: 100 HP: 0 L: 0 HPV: –	–	100c	Reference	Reference
Trials evaluating adjuvant treatments
Bernier et al., 2004 (EORTC 22931) (NCT00002555) [11, 13]	Phase III Europe (multicenter)	Surgery + RT + cisplatin	167	Age: 55 Male: 91.6 White: – Current/former smoker: – ECOG 0/1: –	OC: 24.6 OP: 32.3 HP: 20.4 L: 22.2 HPV: –	ENE: 61.1 PSM ≤ 5 mm: 31.1 PNI: 12.8 VE: 21.0	61	0.7 (0.52–0.95)	PFS 0.75 (0.56–0.99) Progression or death
		Surgery + RT	167	Age: 53 Male: 92.8 White: – Current/former smoker: – ECOG 0/1: –	OC: 27.5 OP: 28.1 HP: 20.4 L: 22.8 HPV: –	ENE: 53.3 PSM ≤ 5 mm: 25.7 PNI: 14.4 VE: 18.6	58	Reference	Reference
Bourhis et al., 2025 (NIVOPOST-OP) (NCT03576417) [21]	Phase III Open-label Multinational	Surgery + RT + cisplatin + nivolumab	332	Age: 59 (18–74) Male: 75.3 White: – Current/former smoker: 86.4 ECOG 0/1: 100	OC: 57.8 OP: 17.2 HP: 13.0 L: 12.0 HPV: 4.8	ENE, microscopic PSM, ≥ 4 metastatic cervical nodes without ENE, multiple PNIs: 100% Microscopic PSM: 61.1 ENE: 61.7 Multiple PNIs: 54.8	30.3c	–	DFS 0.76 (0.60–0.98) Local, regional, or distant failure or death
		Surgery + RT + cisplatin	334	Age: 59 (22–74) Male: 76.7 White: – Current/former smoker: 80.8 ECOG 0/1: 100	OC: 57.8 OP: 18.0 HP: 12.0 L: 12.3 HPV: 5.1	ENE, microscopic PSM, ≥ 4 metastatic cervical nodes without ENE, multiple PNIs: 100% Microscopic PSM: 54.8 ENE: 63.8 Multiple PNIs: 56.3	30.3c	–	Reference
Cooper et al., 2004 (RTOG 9501) (NCT00002670) [12, 13, 27]	Phase III Open-label South Africa, USA	Surgery + RT + cisplatin	206	Age: 56 (24–80) Male: 85.9 White: 75.7 Current/former smoker: – KPS ≥ 70: 99.5	OC: 24.3 OP: 48.1 HP: 7.3 L: 20.4 HPV: –	ENE, microscopic PSM, ≥ 2 metastatic lymph nodes: 100 ENE: 53c Microscopic PSM: 10c	45.9c	0.84 (0.65–1.09)	DFS 0.78 (0.61–0.99)
							112.8c	0.89 (0.7–1.12)	0.88 (0.71–1.09) Any tumor (local, regional, metastatic, or second primary) or death
		Surgery + RT	210	Age: 55 (28–79) Male: 86.2 White: 73.3 Current/former smoker: – KPS ≥ 70: 97.1	OC: 29.5 OP: 37.1 HP: 12.4 L: 21.0 HPV: –	ENE, microscopic PSM, ≥ 2 metastatic lymph nodes: 100 ENE: 53c Microscopic PSM: 10c	45.9c	Reference	Reference
							112.8c
Dietz et al., 2025 (ADRISK) (NCT03480672) [28]	Phase IIb Open-label Germany (multicenter)	Surgery + RT + cisplatin + pembrolizumab	102	Age: 61.5 (36–84) Male: 79.4 White: – Current/former smoker: 78.4 ECOG 0/1: 100	OC: 23.5 OP: 65.7 HP: 3.9 L: 6.9 HPV: 45.1	High risk (microscopic PSM, ENE): 63.7 Intermediate risk (R0 < 5 mm, ≥ 2 metastatic lymph nodes): 36.3	30c	0.83 (0.45–1.53)	EFS 0.81 (0.49–1.35) Locoregional or distant recurrence, further malignant tumor, death, or initiation of new anti-cancer treatment without a previous event
		Surgery + RT + cisplatin	102	Age: 59.8 (33–80) Male: 83.3 White: – Current/former smoker: 76.5 ECOG 0/1: 100	OC: 18.6 OP: 65.7 HP: 5.9 L: 9.8 HPV: 46.1	High risk (microscopic PSM, ENE): 63.7 Intermediate risk (R0 < 5 mm, ≥ 2 metastatic lymph nodes): 36.3	30c	Reference	Reference
Haddad et al., 2025 (IMvoke010) (NCT03452137) [35]	Phase III Quadruple-blinded Multinational	Surgery ± RT + atezolizumab	203	Age: 59.4 (mean) Male: 82.8 White: 63.4 Current/former smoker: 81.8 ECOG 0/1: 99.5	OC: 26.6 OP: 35.0 HP: 22.2 L: 16.3 HPV: 17.2	–	46.5c	0.87 (0.47–1.59)	EFS 0.79 (0.48–1.3) Recurrence, new second primary HNSCC lesion, distant metastasis or death
		Surgery ± RT	203	Age: 57.7 (mean) Male: 85.7 White: 68.5 Current/former smoker: 84.2 ECOG 0/1: 100	OC: 28.1 OP: 38.4 HP: 15.3 L: 18.2 HPV: 17.7	–	46.5c	Reference	Reference
Harrington et al., 2015 (EGF102988) (NCT00424255) [36]	Phase III Double-blinded Multinational	Surgery + lapatinib + RT + cisplatin + maintenance lapatinib	346	Age: 54 (27–74) Male: 82.7 White: 64.2 Current/former smoker: – ECOG 0/1: –	OC: 40.5 OP: 18.8 HP: 15.3 L: 22.3 HPV: 6.6	ENE, PSM ≤ 5 mm: 100	35.3c	0.96 (0.73–1.25)	DFS 1.1 (0.85–1.43) Recurrence or death
		Surgery + RT + cisplatin	342	Age: 55 (24–74) Male: 83.9 White: 64.3 Current/former smoker: – ECOG 0/1: –	OC: 42.4 OP: 19.9 HP: 9.6 L: 23.4 HPV: 6.1	ENE, PSM ≤ 5 mm: 100	35.3c	Reference	Reference
Iyer, 2025 (IHN01) (NCT00957086) [37]	Phase III Double-blinded Multinational	Surgery + RT + cisplatin + nimotuzumab	208	Age: 54 Male: – White: – Current/former smoker: - ECOG 0/1: 100	OC: 67.8 OP: 11.5 HP: 7.2 L: 13.5 HPV: –	ENE, PSM: 73	60.5c	0.84 (0.60–1.18)	DFS 0.94 (0.70–1.25) Progression or death
		Surgery + RT + cisplatin	214	Age: 54 Male: – White: – Current/former smoker: – ECOG 0/1: 100	OC: 65.4 OP: 11.7 HP: 7.5 L: 15.4 HPV: –	ENE, PSM: 73	60.5c	Reference	Reference
Jiang et al., 2024 (NCT02923258) [38]	Phase II Open-label China (single center)	Surgery + RT + docetaxel	112	Age: 59.5 (37–73) Male: 60.7 White: – Current/former smoker: 42.0 ECOG 0/1: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE, PSM, ≥ 2 metastatic lymph nodes: 100	28.8c	0.87 (0.55–1.38)	DFS 0.92 (0.61–1.40) Recurrence or death
		Surgery + RT + cisplatin	112	Age: 56 (24–72) Male: 76.8 White: – Current/former smoker: 51.8 ECOG 0/1: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE, PSM, ≥ 2 metastatic lymph nodes: 100	28.8c	Reference	Reference
Laskar et al., 2023 (NCT00193843) [39]	Phase III Open-label India (single center)	Surgery + RT (5fr/wk) + cisplatin	300	Age: 45 (IQR: 39–53) Male: 87.0 White: – Current/former smoker: – KPS ≥ 70: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	Microscopic PSM: 0.7 ENE: 53.3 PNI: 24.7 VE: 3.7	95.9c	0.95 (0.76–1.18)	DFS 0.93 (0.75–1.17) Any disease progression at locoregional or distant sites or death
		Surgery + RT (6fr/wk)	301	Age: 45 (IQR 38–52) Male: 86.0 White: – Current/former smoker: – KPS ≥ 70: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	Microscopic PSM: 0.3 ENE: 53.5 PNI: 27.9 VE: 2.3	95.9c	0.86 (0.69–1.08)	0.82 (0.66–1.09) Any disease progression at locoregional or distant sites or death
		Surgery + RT (5fr/wk)	299	Age: 46 (IQR 38–54) Male: 87.0 White: – Current/former smoker: – KPS ≥ 70: 99.7	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	Microscopic PSM: 0.3 ENE: 55.9 PNI: 28.4 VE: 3.3	95.9c	Reference	Reference
Ma et al., 2025 (MC1675) (NCT02908477) [40]	Phase III Open-label USA (multicenter)	Surgery + RT + docetaxel	130	Age: 59.4 (38–82) Male: 88.5 White: 93.8 Current/former smoker: 28.5 ECOG 0/1: 100	OC: 0 OP: 100 HP: 0 L: 0 HPV: –	High risk (ENE): 59.2 Intermediate risk (LVI, PNI, ≥ 2 metastatic lymph nodes, lymph node ≥ 3 cm, pT3): 40.8	37.4	1.68 (0.36–7.95)	PFS 4.76 (1.11–20.4) Progression or death
		Surgery + RT + cisplatin	64	Age: 59.2 (48–73) Male: 90.1 White: 95.3 Current/former smoker: 28.1 ECOG 0/1: 100	OC: 0 OP: 100 HP: 0 L: 0 HPV: –	High risk (ENE): 59.4 Intermediate risk (LVI, PNI, ≥ 2 metastatic lymph nodes, lymph node ≥ 3 cm, pT3): 40.6	32.1	Reference	Reference
Machtay et al., 2025 (RTOG 0920) (NCT00956007) [41]	Phase III Open-label USA (multicenter)	Surgery + RT + cetuximab	290	Age: 57.5 (20–79) Male: 69.0 White: 85.9 Current/former smoker: 61.4 ECOG 0/1: 100	OC: 63.4 OP: 23.4 HP: 0 L: 13.1 HPV: 19.0	All intermediate-risk patientsd PNI: 39.7 VE: 27.2	86.4c	0.81 (0.60–1.08)	DFS 0.75 (0.57–0.98) Any recurrence (local, regional, or distant) or death
		Surgery + RT	287	Age: 57 (27–80) Male: 71.1 White: 86.8 Current/former smoker: 67.6 ECOG 0/1: 100	OC: 63.8 OP: 20.9 HP: 0 L: 15.3 HPV: 17.4	All intermediate-risk patients PNI: 39.0 VE: 25.8	86.4c	Reference	Reference
Tobias et al., 2010 (UKHAN1) (NCT00002476) [42]	Phase III Multinational	Surgery + RT ± methotrexate ± vincristine ± bleomycin ± 5-FU	118	Age: 60 (36–81) Male: 74.6 White: – Current/former smoker: – ECOG 0/1: –	OC: 36.4 OP: 21.2 HP: 9.3 L: 28.8 HPV: –	PSM, advanced stage at presentation: 100	120c	0.94 (0.70–1.27)e	EFS 1.03 (0.78–1.37)e Recurrence, new tumor, or death in patients who were disease-free 6 months after randomization
		Surgery + RT	135	Age: 56 (32–76) Male: 77.0 White: – Current/former smoker: – ECOG 0/1: –	OC: 34.1 OP: 25.2 HP: 13.3 L: 20.7 HPV: –	PSM, advanced stage at presentation: 100	120c	Reference	Reference
Wozniak et al., 2020[43]	Poland (multicenter)	Surgery + RT + cisplatin	54	Age ≥ 57 (percentage): 46.3 Male: 68.5 White: – Current/former smoker: – ECOG 0/1: 100	OC: 46.3 OP: 53.7 HP: 0 L: 0 HPV: –	Risk score 7–14: 100f	64.8c	1.03 (0.57–1.87)	–
		Surgery + RT	57	Age ≥ 57 (percentage): 57.9 Male: 68.4 White: – Current/former smoker: – ECOG 0/1: 100	OC: 50.9 OP: 49.1 HP: 0 L: 0 HPV: –	Risk score 7–14: 100f	64.8c	Reference	–

Interventions appearing before and after “surgery” were administered in the neoadjuvant and adjuvant settings, respectively. En dashes (–) indicate values that were not reported. Additions to the standard of care are shown in italic font in the treatment arms column

5-FU fluorouracil, CI confidence interval, DFS disease-free survival, ECOG Eastern Cooperative Oncology Group, EFS event-free survival, ENE extranodal extension, HP hypopharynx, HPV human papillomavirus, HR hazard ratio, KPS Karnofsky performance status, L larynx, LVI lymphovascular invasion, N sample size, OC oral cavity, OP oropharynx, PFS progression-free survival, PNI perineural invasion, PSM positive surgical margins, RT radiotherapy, VE vascular tumor embolism, wk week

^aMedian (range) is reported for age, % is reported for male patients, white patients, current/former smokers, performance score 0 or 1, tumor sites, and HPV status unless otherwise specified

^b% of patients who tested positive for HPV is reported relative to the entire population of the respective treatment arm

^cReported for the entire population

^dDefined as negative final resection margins and no evidence of ENE, but with one or more of resection margins < 5 mm, pathologic T3, T4a, or N2 tumor, LVI and/or PNI, or a T2 oral cavity cancer with > 5 mm depth of invasion

^eThe 95% confidence interval was calculated from the reported 99% confidence interval

^fBased on the scoring system proposed in Peters et al. [52]

Table 2.

Pathological response with neoadjuvant treatments as reported in the included trials

Trial	Trial characteristics	Treatment arms	N	Population characteristicsa	Tumor location and HPV status (%)a,b	High-risk features, %	pCR definition, %	mPR definition, %
Ferris et al., 2023 (NCT04080804) [44]	Phase II Open-label USA (single center)	Nivolumab + ipilimumab + surgery	10	Age: 63 (32–81)c Male: 60.6c White: – Current/former smoker: – ECOG 0/1: 100	OC: 75.8c OP: 15.2c HP: 0 L: 9.1c HPV: 9.1c	–	100% pathological response (no further description provided) 0	> 90% pathological response (no further description provided) 10.0
		Nivolumab + relatlimab + surgery	13	Age: 63 (32–81)c Male: 60.6c White: – Current/former smoker: – ECOG 0/1: 100	OC: 75.8c OP: 15.2c HP: 0 L: 9.1c HPV: 9.1c	–	100% pathological response (no further description provided) 7.7	> 90% pathological response (no further description provided) 15.4
		Nivolumab + surgery	10	Age: 63 (32–81)c Male: 60.6c White: – Current/former smoker: – ECOG 0/1: 100	OC: 75.8c OP: 15.2c HP: 0 L: 9.1c HPV: 9.1c	–	100% pathological response (no further description provided) 0	> 90% pathological response (no further description provided) 0
He et al., 2025 (CAMORAL) (ChiCTR2000037980) [26]	Phase II Open-label China (multicenter)	Camrelizumab + nab-paclitaxel + carboplatin + surgery + RT + camrelizumab ± cisplatin	63	Age: 54.2 (mean) (SD 10.6) Male: 73 White: – Current/former smoker: – ECOG 0/1:100	OC: 79.4 OP: 19 HP/L: 1.6 HPV: 4.8	–	Absence of viable tumor in the entire tumor bed and all sampled lymph nodes 25.4	≤ 10% viable tumor in the entire tumor bed and all sampled lymph nodes 47.6
Kim et al., 2024 (NCT03737968) [45]	Phase II Open-label South Korea (single center)	Durvalumab + tremelimumab + surgery + RT ± chemotherapy + durvalumab	24	Age: 61 (30–85)c Male: 100 White: – Current/former smoker: 75.0 ECOG 0/1: 100	OC: 12.5 OP: 54.2 HP: 20.8 L: 12.5 HPV: 45.8	–	Absence of any viable tumor after neoadjuvant treatment 4.2	≤ 10% tumor viability in the surgical specimen after neoadjuvant treatment 4.2
		Durvalumab + surgery + RT ± chemotherapy + durvalumab	21	Age: 61 (30–85)c Male: 81.0 White: – Current/former smoker: 76.2 ECOG 0/1: 100	OC: 23.8 OP: 47.6 HP: 19.0 L: 9.5 HPV: 42.9	–	Absence of any viable tumor after neoadjuvant treatment 4.8	≤ 10% tumor viability in the surgical specimen after neoadjuvant treatment 9.5
Le et al., 2022 (NCT01927744) [46]	Phase II Double-blinded USA (single center)	Docetaxel + cisplatin (carboplatin) + erlotinib + surgery	24	Age: 60 (29–74) Male: 58.3 White: 79.2 Current/former smoker: 58.3 ECOG 0/1: 95.8	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	–	–	Complete pathologic response or partial pathologic response with minimal residual disease ≤ 10% viable tumor cells. Lymph node status was not included in the definition 30.4
		Docetaxel + cisplatin (carboplatin) + surgery	28	Age: 57 (26–71) Male: 60.7 White: 78.6 Current/former smoker: 39.3 ECOG 0/1: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	–	–	Complete pathologic response or partial pathologic response with minimal residual disease ≤ 10% viable tumor cells. Lymph node status was not included in the definition 41.7
Liu et al., 2025 (NEOPBOSCC) (NCT04649476) [47]	Phase II Open-label China (single center)	Camrelizumab + docetaxel + cisplatin + 5-FU + surgery + RT ± chemotherapy	34	Age: 49.5 (mean) (33–68) Male: 91.2 White: – Current/former smoker: 64.7 ECOG 0/1: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	–	Absence of viable tumor in the entire tumor bed and all sampled lymph nodes 29.4	≤ 10% viable tumor in the entire tumor bed and all sampled lymph nodes 76.4
		Camrelizumab + surgery + RT ± chemotherapy	34	Age: 52 (mean) (32–67) Male: 82.4 White: – Current/former smoker: 70.6 ECOG 0/1: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	–	Absence of viable tumor in the entire tumor bed and all sampled lymph nodes 0	≤ 10% viable tumor in the entire tumor bed and all sampled lymph nodes 14.7
Mastrolonardo et al., 2025 (NCT03854032) [48]	Phase II Open-label USA (multicenter)	Nivolumab + linrodostat + surgery	31	Age: 62.5 (45–84) Male: 74.2 White: – Current/former smoker: 61.3 ECOG 0/1: –	OC: 35.5 OP: 58.1 HP/L: 6.5 HPV: 48.4	–	100% pathologic response at primary site 12.9	≥ 90% pathologic response at primary site 19.4
		Nivolumab + surgery	30	Age: 60.5 (44–86) Male: 90.0 White: – Current/former smoker: 66.7 ECOG 0/1: –	OC: 43.3 OP: 50.0 HP/L: 6.7 HPV: 50	–	100% pathologic response at primary site 10	≥ 90% pathologic response at primary site 10
Uppaluri et al., 2025 (KEYNOTE-689) (NCT03765918) [22]d	Phase III Open-label Multinational	Pembrolizumab + surgery + pembrolizumab + RT ± cisplatin	363	Age: 60 (29–82) Male: 78.8 White: 78.2 Current/former smoker: 80.7 ECOG 0/1:100	OC: 60.3 OP: 9.6 HP: 7.7 L: 22.3 HPV: 3.3	ENE, PSM < 1 mm: 32.5	No residual invasive SCC within the resected primary tumor specimen and all sampled regional lymph nodes 3.0	≤ 10% residual viable invasive SCC within the resected primary tumor specimen and all sampled regional lymph nodes 9.4
Wang et al., 2025 (NCT05522985) [49, 50]	Phase II Open-label China (single center)	Toripalimab + cisplatin + nab-paclitaxel + surgery	61	Age: 60 Male: 75.4 White: – Current/former smoker: 52.5 ECOG 0/1: 100	OC: 37.7 OP: 16.4 HP: 34.4 L: 6.6 HPV: –	–	Complete disappearance of the tumor within the resected tumor bed 42.6	Residual tumor tissue < 10% in all sections within the resected tumor bed 60.7
		Cisplatin + nab-paclitaxel + surgery	61	Age: 59 Male: 85.2 White: – Current/former smoker: 49.2 ECOG 0/1: 100	OC: 29.5 OP: 19.7 HP: 34.4 L: 14.8 HPV: –	–	Complete disappearance of the tumor within the resected tumor bed 18.0	Residual tumor tissue < 10% in all sections within the resected tumor bed 37.7
Zhong et al., 2013 (NCT01542931)d [33]	Phase III Open-label China (single center)	Docetaxel + cisplatin + 5-FU + surgery + RT	128	Age: 55 (29–74) Male: 71.1 White: – Current/former smoker: 53.9 KPS > 60: 100	OC: 100 OP: 0 HP: 0 L: 0 HPV: –	ENE: 13.2	Absence of any tumor cells in the resected specimen 13.4	Minimal residual disease with < 10% viable tumor cells 27.7
Zhong, 2022 (EAGLE) (NCT01434394)d [34]	Phase III Open-label China (multicenter)	Cetuximab + docetaxel + cisplatin + surgery + RT	134	Age: – Male: – White: – Current/former smoker: – ECOG 0/1: –	OC/OP: 100 HP: 0 L: 0 HPV: –	–	– 10.5	–

Interventions appearing before and after “surgery” were administered in the neoadjuvant and adjuvant settings, respectively. En dashes (–) indicate values that were not reported. Neoadjuvant treatments are shown in italic font in the treatment arms column

5-FU fluorouracil, ENE extranodal extension, ECOG Eastern Cooperative Oncology Group, HP hypopharynx, HPV human papillomavirus, L larynx, KPS Karnofsky performance status, mPR major pathological response, N sample size, OC oral cavity, OP oropharynx, pCR pathological complete response, PSM positive surgical margins, RT radiotherapy, SCC squamous cell carcinoma

^aMedian (range) is reported for age, % is reported for male patients, white patients, current/former smokers, performance score 0 or 1, tumor sites, and HPV status unless otherwise specified

^b% of patients who tested positive for HPV is reported relative to the entire population of the respective treatment arm

^cReported for the entire population

^dOnly trial arms evaluating neoadjuvant treatments are presented in the table

Survival Outcomes

A total of 19 trials reported the survival outcomes of interest in the target population. Two, four, and 13 trials evaluated systemic treatments in the perioperative (pembrolizumab, camrelizumab + nab-paclitaxel + carboplatin), neoadjuvant (docetaxel + cisplatin + fluorouracil [5-FU], leukocyte interleukin [with or without cyclophosphamide + indomethacin + zinc], cetuximab + docetaxel + cisplatin) and adjuvant (cisplatin, atezolizumab, nivolumab, pembrolizumab, nimotuzumab, lapatinib, docetaxel, cetuximab, and methotrexate ± vincristine ± bleomycin ± 5-FU) settings, respectively (Table 1). Trials typically evaluated the addition of these systemic treatments to surgery + RT or surgery + RT + cisplatin, except Jiang et al. [38] and Ma et al. [40] (both replaced adjuvant cisplatin with adjuvant docetaxel, following surgery + RT). Most trials were phase III studies and had an open-label design, often with sample sizes larger than 200 patients. Around half of the trials were conducted in multiple countries; four were conducted only in China, two only in India, one only in Poland, one only in Germany, and two only in the USA. All except four trials were conducted in multiple centers. Median follow-up durations mainly ranged between 30 and 65 months, with three trials reporting median follow-up durations of 100 months or longer. Trials were generally of high quality with low risk of bias across the evaluated domains, with only five trials with some concerns for bias arising from the randomization process or deviations from intended interventions (Supplementary Appendix D).

Most populations comprised entirely patients with an Eastern Cooperative Oncology Group (ECOG) performance score of 0 or 1 or a Karnofsky performance score of 70 or higher. Five trials were exclusively conducted in patients with oral cavity cancers (three evaluating neoadjuvant treatments and two adjuvant treatments), one (evaluating adjuvant treatments) exclusively in patients with oropharyngeal cancers, and one (evaluating neoadjuvant treatments) in a mix of patients with oral cavity or oropharyngeal cancers. In the remaining studies (evaluating perioperative or adjuvant treatments), tumor locations were more evenly distributed. Except for one trial, only a small percentage of patients had HPV-positive oropharyngeal cancer across the seven trials that reported this characteristic. Populations were similar in terms of sex (typically 70–90% male patients), race (often white), and smoking status (predominantly current or former smokers), with mean or median age generally ranging between 45 and 60 years across the treatment arms.

There was heterogeneity in terms of the definition and distribution of tumors that were considered at high risk of recurrence. High-risk features were reported with different underlying definitions in terms of qualifying characteristics, such as presence of ENE, PSM, tumor invasion to ≥ 2 lymph nodes, pathological tumor stage, involvement of lymph nodes at level IV or V (oral cavity or oropharyngeal tumors), perineural invasion, and vascular tumor embolism. The definition of PSM also varied, with Bourhis et al. [21] and Uppaluri et al. [22] defining it as margins < 1 mm from the invasive front of tumor, Bernier et al. [11] and Harrington et al. [36] as ≤ 5 mm from the tumor, and Cooper et al. [12] and Dietz et al. [28] as microscopically involved mucosal margins of resection (microscopic PSM). Other trials did not provide clear definitions for PSM. Regardless of the exact definitions, populations in Bernier et al. [11], Bourhis et al. [21], Cooper et al. [12], Harrington et al. [36], Jiang et al. [38], and Wozniak et al. [43] entirely comprised patients whose tumors were considered at high risk of recurrence (based on postoperative assessments of tumor specimens and surgical margins), whereas around 30–60% of patients in most other trials had tumors with high-risk features. Lastly, Machtay et al. [41] identified its entire population as having intermediate-risk tumors, defined as negative final resection margins and no evidence of ENE, but with one or more of resection margins < 5 mm, pathologic T3, T4a, or N2 tumor, lymphovascular and/or perineural invasion, or a T2 oral cavity cancer with > 5 mm depth of invasion.

Event-Free Survival

Across the 18 trials reporting EFS, outcome definitions were provided in all trials except Zhong et al., 2022 [34] (only reported on in a conference poster). The definition of EFS (or analogous outcomes) often included as events any progression (or recurrence) at local, regional, and distant sites as well as death due to any cause, except in Cooper et al. [12], Dietz et al. [28], Haddad et al. [35], Talor et al. [32], and Tobias et al. [42], where additional clinical outcomes such as “second primary” tumor, “new second primary HNSCC lesion”, and “new tumor” were also considered events. EFS was consistently reported in the overall population in all trials except in Tobias et al. [42], where it was reported in patients who were disease-free 6 months after randomization. Details of outcome definitions and their trial-specific labels (e.g., event-free survival, disease-free survival) are provided in Table 1.

Uppaluri et al. [22] demonstrated significant improvement with the addition of perioperative pembrolizumab to surgery + RT ± cisplatin in the overall population (hazard ratio 0.73; 95% confidence interval 0.58–0.92) as well as within subgroup of patients with combined positive score (CPS) ≥ 10 (0.66; 0.49–0.88) and CPS ≥ 1 (0.70; 0.55–0.89). Similarly, in He et al. [26], addition of perioperative camrelizumab with neoadjuvant nab-paclitaxel + carboplatin to surgery + RT ± cisplatin resulted in significant improvement in the overall population (0.34; 0.17–0.68). No significant improvement in EFS was observed with the addition of neoadjuvant docetaxel + cisplatin + 5-FU [29] or neoadjuvant leukocyte interleukin (with or without cyclophosphamide + indomethacin + zinc) [32] to surgery + RT ± cisplatin. No significant improvement was observed also with the addition of neoadjuvant docetaxel + cisplatin + 5-FU [33] or neoadjuvant cetuximab + docetaxel + cisplatin [34] to surgery + RT.

Among the trials evaluating the addition of adjuvant treatments, surgery + RT + cisplatin improved EFS versus surgery + RT in Bernier et al. [11] (0.75; 0.56–0.99) and Cooper et al. [12] (0.78; 0.61–0.99 after a median follow-up of 45.9 months), whereas it was comparable to surgery + RT in Laskar et al. [39]. Of note, long-term (10-year follow-up) results from Cooper et al. [27] also showed comparable EFS between these two treatments. Bourhis et al. [21] showed significant improvement in EFS (specified as disease-free survival) with the addition of adjuvant nivolumab to surgery + RT + cisplatin in the overall population (0.76; 0.60–0.98). No improvement was observed with the addition of adjuvant pembrolizumab (Dietz et al. [28]) or adjuvant nimotuzumab (Iyer et al. [37]) to surgery + RT + cisplatin. The addition of adjuvant cetuximab to surgery + RT improved EFS versus surgery + RT in Machtay et al. [41] (0.75; 0.57–0.98). Replacing adjuvant cisplatin with docetaxel was associated with comparable EFS in Jiang et al. [38] and with significantly worse EFS in Ma et al. [40].

Overall Survival

Across the 18 trials reporting OS, no significant improvement in OS was observed with the addition of perioperative pembrolizumab, perioperative camrelizumab with neoadjuvant nab-paclitaxel + carboplatin, neoadjuvant docetaxel + cisplatin + 5-FU or neoadjuvant leukocyte interleukin (with or without cyclophosphamide + indomethacin + zinc) to surgery + RT ± cisplatin. It is noteworthy to mention that since the protocol-specified criterion for statistical significance was not met in Uppaluri et al. [22] for OS within the subgroup of patients with CPS ≥ 10, subsequent OS hypotheses were not formally tested at the first interim analysis and are planned to be re-evaluated in subsequent follow-up analyses. No significant improvement was observed with the addition of neoadjuvant docetaxel + cisplatin + 5-FU or cetuximab + docetaxel + cisplatin to surgery + RT.

Among the trials evaluating the addition of adjuvant treatments to SoC, significant improvement in OS was only observed with the addition of adjuvant cisplatin to surgery + RT. Specifically, surgery + RT + cisplatin improved OS versus surgery + RT in Bernier et al. [11] (0.7; 0.52–0.95), whereas it was comparable to surgery + RT in Cooper et al. [12], Wozniak et al. [43], and Laskar et al. [39]. Replacing adjuvant cisplatin with docetaxel (Jiang et al. [38]; Ma et al. [40]) following surgery + RT did not improve OS. It should be noted that the p value for the reported OS hazard ratio in He et al. [26] was < 0.05. However, this p value was only nominal, i.e., with no multiplicity adjustment, and therefore did not indicate a statistically significant difference based on prespecified hypothesis testing.

Pathological Response

Ten trials reported on pathological response following neoadjuvant systemic treatments in the target population. In six trials, neoadjuvant treatments were evaluated in all intervention arms, whereas they were only evaluated in the experimental arms (and not the control arms) of He et al. [26], Uppaluri et al. [22], Zhong et al., 2013 [33], and Zhong et al., 2022 [34]. As pathological response is only relevant for intervention arms with neoadjuvant treatment, only these arms are further described in this section. The evaluated neoadjuvant treatments included pembrolizumab, docetaxel + cisplatin + 5-FU, cetuximab + docetaxel + cisplatin, nivolumab (alone or with ipilimumab, relatlimab, or linrodostat) durvalumab (alone or with tremelimumab), docetaxel + cisplatin (with or without erlotinib), cisplatin + nab-paclitaxel (alone or with toripalimab), and camrelizumab (alone, with docetaxel + cisplatin + 5-FU, or with nab-paclitaxel + carboplatin) (Table 2).

The 10 trials in this category included three large phase III studies (with > 100 patients in individual treatment arms) and seven small phase II trials (often with < 35 patients in individual treatment arms). Only one trial was multinational; three were conducted in the USA only, five in China only, and one in South Korea. Most trials were conducted in single centers. Trials were generally of high quality with low risk of bias (Supplementary Appendix D).

Most populations comprised entirely patients with an ECOG performance score of 0 or 1 or a Karnofsky performance score of 60 or higher. Three trials were exclusively conducted in patients with oral cavity cancers, and one was in a mix of patients with oral cavity or oropharyngeal cancers. Patients in the remaining studies often had cancers in the oral cavity or oropharynx. Across the trial arms evaluating neoadjuvant treatments, the percentage of patients with HPV-positive oropharyngeal cancer was < 10% in Uppaluri et al. [22], He et al. [26], and Ferris et al. [44], whereas around half of the populations had such lesions in Kim et al. [45] and Mastrolonardo et al. [48]. Furthermore, while all trial populations comprised patients with stage III–IV disease, around half of the patients had stage II disease in Kim et al. [45]. Populations were similar in terms of sex (predominantly male patients) and smoking status (often current or former smokers), with mean or median age generally ranging between 55 and 65 years across trial arms evaluating neoadjuvant treatments.

While definitions of pCR and mPR were similar in terms of the minimum required percentage of residual viable tumor (0% for pCR and ≤ 10% for mPR), there was heterogeneity in terms of the scope of lesions that was evaluated. Among the nine trials that reported pCR, Uppaluri et al. [22], He et al. [26], and Liu et al. [47] evaluated the resected primary tumor specimen and all sampled regional lymph nodes, while Mastrolonardo et al. [48] and Wang et al. [49] focused on the primary tumor site only. The other trials did not clearly specify the scope of their evaluations. Across the treatment arms that included at least one neoadjuvant chemotherapy agent, pCR rates were 10.5%, 13.4%, 18.0%, 25.4%, 29.4%, and 42.6% with cetuximab + docetaxel + cisplatin, docetaxel + cisplatin + 5-FU, cisplatin + nab-paclitaxel, camrelizumab + nab-paclitaxel + carboplatin, camrelizumab + docetaxel + cisplatin + 5-FU, and toripalimab + cisplatin + nab-paclitaxel, respectively. In the treatment arms without any neoadjuvant chemotherapy agent, pCR rates ranged from 0 to 12.9%.

Among the nine trials that reported mPR, Uppaluri et al. [22], He et al. [26], and Liu et al. [47] evaluated the resected primary tumor specimen and all sampled regional lymph nodes, while Le et al. [46], Mastrolonardo et al. [48], and Wang et al. [49] focused on the primary tumor site only. The other trials did not clearly specify the scope of their evaluations. Rates of mPR ranged from 27.7% to 76.4% and from 0 to 19.4% across the treatment arms with and without chemotherapy regimens, respectively. Of note, postoperative assessment of high-risk features of resected specimens were only reported in Uppaluri et al. [22] and Zhong et al., 2013 [33] (with slightly different definitions), where 32.5% (pembrolizumab) and 13.2% (docetaxel + cisplatin + 5-FU) of patients had high-risk tumors, respectively.

Four trials explored the correlation between pathological response and survival-related outcomes. In Zhong et al., 2013 [33], a favorable (mPR) pathological response was associated with improved EFS (hazard ratio 0.31; 95% confidence interval 0.12–0.81) and OS (0.35; 0.13–0.94) on multivariate Cox model analysis versus unfavorable response. Similarly, Zhong et al., 2022 [34] found that patients who achieved pCR had superior EFS (hazard ratio 0.31; p = 0.01) and OS (hazard ratio 0.38; p = 0.03) compared to those who did not. In Liu et al. [47], patients with pathological response (mPR) had improved EFS rates compared to non-responders (p < 0.0001). Lastly, Le et al. [46] found that patients achieving pathological response (mPR) had a trend towards improved PFS; however, results were not significant.

Discussion

Since the early 2000s, a risk-adapted approach has been adopted as SoC for the management of patients with resectable LA-HNSCC, with surgery + RT as the backbone and additional adjuvant cisplatin chemotherapy reserved for tumors that are at higher risk of recurrence. While findings from the pivotal Bernier et al. and Cooper et al. trials showed improvements in certain clinical outcomes with the addition of concomitant adjuvant cisplatin chemotherapy to RT, these improvements are suboptimal [11, 12]. To address the unmet need for a novel intervention in this population, alternative treatments have been evaluated in clinical trials over the past two decades. The current systematic review aimed to provide an up-to-date understanding of the efficacy of these treatments when added to SoC.

Trials evaluating EFS and OS demonstrated substantial heterogeneity. The timing of systemic therapy relative to surgery varied, with interventions delivered in perioperative (e.g., pembrolizumab), neoadjuvant (e.g., docetaxel-based triplet chemotherapies), and adjuvant (e.g., cisplatin, cetuximab) settings. Some trials enrolled exclusively patients with high-risk tumors or intermediate-risk tumors. While some studies restricted eligible tumor locations to oral cavity, others included patients with lesions from broader head and neck sites. Certain trials had more strict definitions for PSM, and others included additional tumor characteristics (e.g., perineural invasion) besides ENE and PSM in their definition of high-risk tumors. Only one trial (Zhong et al., 2022 [34], reported in a conference poster) did not provide a definition for EFS. While all other trials included recurrence and death in their EFS definitions, there were a few that additionally counted second primary tumors as events.

Perioperative immunotherapy with pembrolizumab and perioperative immunotherapy with camrelizumab with neoadjuvant nab-paclitaxel + carboplatin significantly improved EFS versus risk-adapted SoC, i.e., surgery + RT ± cisplatin (chemotherapy for patients with high-risk tumors) in Uppaluri et al. [22] and He et al. [26], respectively. With perioperative pembrolizumab, similar improvements in EFS were observed within subgroups of patients with CPS ≥ 10 and CPS ≥ 1 of Uppaluri et al. [22], both of which were prespecified co-primary endpoints of the trial. Although safety outcomes were not the focus of the current study, it is noteworthy to mention that rates of any-grade and grade 3–4 treatment-related adverse events with perioperative pembrolizumab in combination with surgery + RT ± cisplatin were similar to surgery + RT ± cisplatin in Uppaluri et al. [22] (Supplementary Appendix E). Treatment-related adverse events were not presented in He et al. [26] (conference slide deck).

The addition of adjuvant immunotherapy with nivolumab to surgery + RT + cisplatin in patients with high risk of recurrence (defined by presence of ENE and/or PSM, ≥ 4 nodal involvements, multiple perineural invasions) improved EFS (specified as disease-free survival) in Bourhis et al. [21]. Rates of any-grade and grade 3–4 treatment-related adverse events with adjuvant nivolumab arm were similar to the control arm, particularly in the first 100 days after the last treatment of lead-in and concomitant phases of the trial (Supplementary Appendix E). The addition of adjuvant cisplatin to surgery + RT improved EFS in Bernier et al. [11] in patients with tumors broadly exhibiting any one of pT3/pT4, T1/T2 with N2/N3, T1/T2 with N0/N1 and any unfavorable pathological findings (ENE, PSM, perineural invasion, or vascular tumor embolism), as well as oral cavity or oropharyngeal tumors with involved lymph nodes at level IV or V [11]. Although initial findings from Cooper et al. [12] showed similar improvements in EFS with addition of adjuvant cisplatin to surgery + RT in patients with ENE, microscopic PSM, and/or invasion of ≥ 2 regional lymph nodes, the estimated treatment effect lost its statistical significance in the 10-year follow-up analysis of the trial [27]. Despite that, EFS (as well as locoregional control) remained significantly improved with the addition of adjuvant cisplatin to surgery + RT within the subgroup of patients with ENE and/or microscopic PSM (unplanned subgroup analysis) [27]. Similar to the long-term results from Cooper et al. [27], no significant improvement in EFS was observed with the addition of adjuvant cisplatin to surgery + RT in Laskar et al. [39] in patients with oral cavity cancers. Lastly, the addition of adjuvant cetuximab to surgery + RT improved EFS in Machtay et al. [41] in patients with intermediate-risk tumors. No improvement was observed with the addition of any of the other evaluated interventions. Replacing adjuvant cisplatin with adjuvant docetaxel (following surgery + RT) resulted in comparable EFS in Jiang et al. [38] but worse EFS in Ma et al. [40].

Although significant improvement in OS was observed with surgery + RT + cisplatin compared to surgery + RT in Bernier et al. in patients with high-risk tumors [11], comparable results were observed between the two interventions in other trials that had similar populations or were conducted more recently [12, 39, 43]. Findings from our review suggest that immunotherapies, targeted therapies, and cytotoxic regimens, as evaluated in various neoadjuvant and adjuvant settings, were generally comparable to SoC in terms of OS. Of note, improvement in OS with the addition of perioperative pembrolizumab to surgery + RT ± cisplatin was not significant versus SoC in patients with CPS ≥ 10 in the first interim analysis results of Uppaluri et al. [22]. As such, testing of subsequent OS hypotheses was not formally conducted at this time and is planned for future analyses of that trial [22]. Similarly, the statistical analysis of OS with the addition of adjuvant nivolumab to surgery + RT + cisplatin in Bourhis et al. [21] requires more mature data at this time and is planned to be performed when the prespecified number of deaths is reached.

Although the current review focused on randomized controlled trials (i.e., studies with the highest quality) and on overall trial populations (rather than subgroups) to maximize internal and external validity of findings reported herein, considerable heterogeneity in interventions, population characteristics, and outcome definitions makes it challenging to compare said findings across studies without access to individual patient-level data and adjusting for between-trial differences. Nevertheless, our findings are consistent with the recent US Food and Drug Administration approval of pembrolizumab as the first immune checkpoint inhibitor in this population. Results from Uppaluri et al. [22], He et al. [26], and Bourhis et al. [21] generally support the use of immune checkpoint inhibitors in addition to the current SoC. It should be noted that, while EFS improvements are encouraging, it is not a guaranteed surrogate for OS. Several included trials report immature or heterogeneous OS, while having heterogenous populations in terms of tumor location and HPV status, which limits generalizability of results. As such, our conclusions regarding the role of immunotherapy and other perioperative strategies are intentionally conservative and bounded pending longer follow-up and subgroup-specific evidence from ongoing and future trials.

The subset of trials reporting on pathological response following neoadjuvant therapy was also heterogeneous. These recent trials often evaluated immunotherapies with inhibitors of programmed cell death protein 1/death ligand 1 as single agents (e.g., nivolumab, pembrolizumab), with chemotherapy-heavy regimens (e.g., camrelizumab + docetaxel + cisplatin + 5-FU), or as part of novel combinations with other immunotherapies (e.g., nivolumab + ipilimumab, nivolumab + relatlimab). Population characteristics varied, with some trials exclusively conducted in patients with oral cavity or oropharyngeal cancers or in populations where around half of patients had HPV-positive oropharyngeal cancers, all known to have more favorable prognosis. Variations in the distribution of prognostic factors such as tumor location and HPV-positivity status for oropharyngeal cancers, with potential effect-modifying properties, complicate indirect comparisons across trials without appropriate adjustments through rigorous statistical analyses.

Only two trials reported on high-risk patients, although their definitions of high-risk features were different. While trials used thresholds of 0% viable tumor for pCR and ≤ 10% for mPR, the scope of tissue evaluated (primary tumor only versus primary tumor plus regional lymph nodes) differed, with only three and four trials providing a clear scope within the subsets of studies reporting on pCR and mPR, respectively. Most treatment arms had small (< 30 patients) sample sizes, with some having as few as 10 patients. These differences and limitations add considerable uncertainty to potential cross-trial comparisons. Despite that, a trend was observed across the treatment arms in terms of numerically higher pCR and mPR rates with immunochemotherapy combinations versus purely immunotherapy-based regimens.

Future trials comparing immunochemotherapy with immunotherapy alone would help better understand their relative benefits and risks as only one trial (Liu et al. [47]) directly compared them to each other in the current evidence base. Of note, a similar trend was observed in a study in 258 patients with resectable non-small cell lung cancer receiving neoadjuvant therapy, where immunochemotherapy resulted in a higher mPR rate compared to immunotherapy alone (53.9% versus 8.7%, respectively; p < 0.001) [51]. Lastly, four trials explored the correlations between pathological response and survival-related outcomes, of which three [33, 34, 47] found positive associations, albeit often as part of exploratory or post hoc analyses. In Le et al., patients achieving mPR had a trend towards improved progression-free survival although results were not significant (likely due to limited sample size) [46]. These findings suggest that achieving pathological response with neoadjuvant therapies may be associated with more favorable survival outcomes in patients with resectable LA-HNSCC. Trial-level and patient-level correlation analyses of future trials can provide a more precise estimate of this association.

This review focused on survival- and response-related outcomes in the target population. Review of safety outcomes revealed inconsistent reporting of data across the trials (Supplementary Appendix E). For example, over half of the trials comparing survival outcomes did not report any-grade or grade 3–4 treatment-related adverse events, while most others only reported one of the two outcomes. Although these data were more frequently presented in trials reporting pathological response, information on preoperative treatment-related adverse events (i.e., with neoadjuvant treatments) was not provided separately for treatment arms that additionally included adjuvant RT (± cisplatin), making it challenging to isolate the toxicity rates attributable to the neoadjuvant treatment, thereby making a qualitative comparison across the trials more challenging. The variability and limited availability of long-term safety reporting observed across the included trials highlight a clear need for future studies to standardize definitions, time horizons, and reporting practices for treatment-related adverse events.

This review leveraged highly sensitive searches in peer-reviewed literature guided by pre-specified study eligibility criteria. It involved two researchers to minimize screening and data extraction errors. The scope was restricted to recent randomized controlled trials reporting clinically relevant outcomes to maximize the quality of evidence and relevance to the emerging treatment landscape. Comparative measures of survival were selected as the primary outcomes to facilitate decision-making for clinicians and other stakeholders. Despite the above strengths, there are limitations to this study as is the case with all systematic reviews. Trials published around or after the search date may not have been systematically captured. As such, a manual search was performed to ensure relevant publications were included. There is also a risk of publication bias as some clinical trials fail to publish results while others are published only in the form of conference abstract or poster, which present limited information. An extensive search of recent proceedings of relevant oncology conferences and clinical trial registries was performed to mitigate the impact of this potential bias on the results of the current review. We acknowledge patients were not directly involved in the current analysis, and only aggregate-level data that were already published were incorporated from the literature into the current study. Furthermore, no data have been presented by sex and/or gender. Lastly, search and selection were restricted to trials published in English and therefore there is a risk that non-English publications were not identified.

Conclusion

This systematic review highlights the evolving treatment landscape for patients with resectable LA-HNSCC and underscores the limitations of current risk-adapted SoC in achieving optimal long-term outcomes. Despite some differences in the definition of EFS across trials, immunotherapy with inhibitors of programmed cell death protein 1 shows promise as an addition to surgery with adjuvant RT ± cisplatin, primarily through EFS improvements observed in randomized controlled trials, while most other evaluated interventions have not demonstrated consistent EFS benefits. Benefits in OS have not been consistently demonstrated to date as data may still be immature in key trials. Evidence on pathological response suggests potential prognostic value for long-term survival outcomes, but heterogeneity in definitions and small sample sizes constrain definitive conclusions. Overall, these findings emphasize the need for larger studies with longer follow-up data and more consistent endpoint definitions to compare patient-relevant clinical outcomes and inform future therapeutic strategies. Ongoing and future randomized controlled trials will be critical to refining the role of immunotherapy and combination regimens in this population.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contribution

All authors (Ali Mojebi, Yuexin Tang, Sam Keeping, Sanjay Merchant, Behzad Bidadi, and Dandan Zheng) contributed to the study conception and design. Material preparation, data collection, and quality control of data were performed by Ali Mojebi and Dandan Zheng. The first draft of the manuscript was written by Ali Mojebi, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. All named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take responsibility for the integrity of the work, and have given their approval for the manuscript to be published.

Funding

This research and the journal’s Rapid Service Fee and Open Access Fee were funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.

Data Availability

The data summarized in this systematic literature review were derived from the individual published studies that were included in the evidence base. All 25 included studies have been cited with complete references and can be accessed through their respective journals. All data are available within the paper (Tables 1 and 2) or its Supplementary Appendix.

Declarations

Conflict of Interest

Dandan Zheng, Yuexin Tang, and Sanjay Merchant are employees of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA and are shareholders of Merck & Co., Inc., Rahway, NJ, USA. Behzad Bidadi was an employee of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and a shareholder of Merck & Co., Inc., Rahway, NJ, USA, at the time of study execution. Ali Mojebi and Sam Keeping are employees of Precision AQ LLC, which received funding from Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.

Ethical Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.