Surgical Versus Non-Surgical Treatment of Advanced Oropharyngeal Squamous Cell Carcinoma: A Systematic Review and Meta-Analysis

Abstract

Background:

This review aims to compare survival outcomes of surgical versus non-surgical treatment of advanced-stage OPSCC.

Methods:

CINAHL, Cochrane Library, PubMed, and Scopus were searched. Studies comparing survival of patients with advanced OPSCC treated surgically (+/− adjuvant therapy) or non-surgically were included. Meta-analyses hazard ratio (HR) and proportions (%) were performed for the outcomes of 2-, 3-, and 5-year overall (OS), disease-free (DFS), disease-specific (DSS), and relapse-free (RFS) survival.

Results:

44 studies reporting on (N = 35,681) patients were included, with a mean age of 58.9. For advanced-stage disease, primary treatment with surgery and chemoradiotherapy (CRT) demonstrated the best 5-year OS, at 67.1% (95% CI 55.6–77.6). Survival outcomes favored Surgery + CRT over other primary treatment modalities.

Conclusions:

Surgical treatment with adjuvant CRT may be associated with better OS outcomes compared with non-surgical treatment. Further research is needed to determine superior treatment modalities in OPSCC subgroups, such as HPV+ tumors.

Introduction

Head and neck squamous cell carcinoma (HNSCC) includes a group of cancers involving the oral cavity, pharynx, hypopharynx, larynx, nasal cavity, and salivary glands.¹ HNSCC is a very common type of cancer, accounting for 4.5% of global cancer diagnoses and deaths.¹ Oropharyngeal squamous cell carcinoma, a subtype of HNSCC, affects the tonsillar bed, base of tongue (BOT), soft palate, and uvula, and has a global annual incidence of 98,000.^1,2 Risk factors for OPSCC primarily include alcohol, tobacco, and HPV (human papillomavirus) infection.¹ Given the increasing use of alcohol and tobacco in developing countries, as well as the increase in HPV-related cancers, the incidence of OPSCC is rapidly increasing.¹ Additionally, OPSCC when presenting in advanced stages compounds the morbidity and mortality of this disease.³

Primary treatment modalities of advanced stage OPSCC include surgical resection with adjuvant therapy, chemoradiotherapy, induction chemotherapy or various combinations of the three.⁴ Within the setting of advanced-stage OPSCC, the heterogeneity of treatment modalities is notable and may contribute to the wide range of outcomes seen in these cancers. While most cases of advanced-stage OPSCC are treated non-surgically, a National Cancer Database study demonstrated that advanced oropharyngeal cases which are HPV-negative and T4 have been shown to have relatively greater rates of primary surgical therapy.⁵ Despite these trends, there is no well-known standardized or superior treatment modality of advanced-stage OPSCC, leaving significant gaps in our ability to effectively treat these cancers.

The aim of this systematic review and meta-analysis is to describe survival outcomes by treatment modality for advanced-stage oropharyngeal squamous cell carcinoma, and identify any differences in survival between treatment modality options.

Methods

Stage 1: Literature Search

This study was conducted in accordance with the guidelines provided by the preferred reporting items for systematic reviews and Meta-analyses (PRISMA) and was registered under PROSPERO ID #CRD42024516712.³ To develop the research question, the PICO (population, intervention, comparison, outcome) framework was utilized. A comprehensive literature search of CINAHL (EBSCO), PubMed (National Library of Medicine, National Institutes of Health), and Scopus (Elsevier) was performed from inception through February 29, 2024, for English language articles. The search strategy was designed using Medical Subject Headings (MESH) and keywords including advanced oropharynx, tonsillar, or base of tongue squamous cell carcinoma, surgical treatment, non-surgical treatment, and treatment outcomes. The original search strategy was developed for PubMed and subsequently reformatted for CINAHL, and Scopus. The search strategies are outlined in Supplemental Appendix S1. All collected references were uploaded into the Covidence systematic review management software (Veritas Health Innovation Ltd.).

Stage 2: Study Selection

Two authors (M.K.S. and J.M.O.) independently screened articles based on title and abstract followed by full text to identify all articles that met inclusion criteria. Disagreements were resolved by a third author (S.A.N.). Inclusion criteria were (1) advanced oropharynx, tonsillar, or base of tongue squamous cell carcinoma, (2) human subjects, (3) primary surgical intervention of treatment of disease compared with non-surgical primary treatment, (4) outcomes including local control, local-regional control, severe complications, 5-year overall survival, 5-year disease special survival, and 5-year disease free survival all stratified by clinical stage, (5) primary research articles (6) English language. Advanced OPSCC was defined by AJCC 7^th or 8^th edition stage III-IV, due to the retrospective nature of the included studies. No included studies reported on AJCC 8^th edition staging, thus, 7^th edition staging was considered. Studies reporting on cohorts with all stages of OPSCC were included if they specifically reported specifically on cohorts of AJCC stage III-IV, with advanced-stage interventions and outcomes noted. Exclusion criteria were (1) carcinoma other than oropharynx, tonsillar, or base of tongue, (2) any study not comparing primary surgery versus non-surgical treatment, (3) non-English studies, (4) non-extractable data (5) study designs other than primary research.

Stage 3: Study Appraisal

Level of evidence for all included studies was assessed using the Oxford Center for Evidence-Based Medicine criteria (LOE).⁶ The risk of bias of all included studies was determined as “low,” “high,” or “unclear” using the updated The Risk of Bias in Non-Randomized – Interventional (ROBINS-I) assessment tool (ROB). Both the level of evidence and risk of bias were determined independently by two authors (M.K.S.) and (J.M.O.). Disagreements were settled through discussion with a third author (S.A.N.).

Stage 4: Data Extraction

Two authors (M.K.S. and J.M.O.) independently extracted the data into a standardized Excel template and compared for accuracy. Any discrepancies were resolved by a third author (S.A.N.). Author, publication year, country of origin, and demographics were collected. Outcome data including local control (LC), local-regional control (LRC), 2–5-year overall survival (OS), 2–5-year disease specific survival (DSS), and 2–5-year disease free survival (DFS) were collected and stratified by clinical stage and treatment modality, which included surgery alone, radiation alone, chemotherapy alone, combined CRT, surgery with radiation, and surgery with CRT. “Advanced-stage” was defined as AJCC 7^th Edition stages III and IV, as the 7^th edition staging system was utilized by the included studies.

Stage 5: Statistical Analysis

Meta-analysis of continuous measures (age) and meta-analysis of proportions (patient characteristics, OS, DSS, RFS,..) was performed using MedCalc 23.02 (MedCalc Software, Ostend, Belgium). In addition, a meta-analysis of hazard ratio (Surgery + CRT vs. Surgery + Radiation, Surgery + CRT vs. CRT) for DSS and OS was performed by Comprehensive Meta-Analysis version 4 (Biostat Inc, Engle-wood, NJ, USA). Each measure (hazard ratio (HR)/ mean/proportion (%) and 95% confidence interval [CI]) was weighted according to the number of patients affected.

As some studies reported the outcomes in median (first quartile, third quartile), the quantile estimation method was deployed to calculate the pooled estimates.⁷ Heterogeneity among studies was assessed using I² statistics with fixed effects (I² < 50%) and random effects (I² > 50%). In addition, a comparison of proportions was done to compare surgical versus non-surgical for adverse events, LC, LRC, SC, OS, DSS, DFS, and patient safety. Finally, potential publication bias was evaluated by visual inspection of the funnel plot and Begg’s test, which statistically examines the asymmetry of the funnel plot.^8–10 A P<.05 was considered to indicate a significant difference for all statistical tests.

Results

Study Characteristics

The literature search returned 1337 unique articles. 1236 articles were excluded during title and abstract screening, leaving 101 studies for full-text review. A total of 44 studies were included in this systematic review. The PRISMA diagram, which outlines this search and screening process, is displayed in Figure 1. Articles selected for inclusion were level 3 studies based on the Oxford level of evidence and were published between 1981 and 2022. Descriptive features of these included studies are summarized in Table 1. Critical appraisal of non-randomized studies (Figure 2) indicated an overall acceptably low risk of bias with the most frequent risk of bias being due to missing data or classification of interventions. Finally, a funnel plot with Begg’s test (Kendall’s Tau: 0.074, p = 0.568) demonstrated that all studies were found within the funnel, suggesting low publication bias (Figure 3).

Figure 1: PRISMA Flow Diagram.

Preferred Reporting Items for Systematic Reviews and Meta-analyses flow diagram of study selection.

Table 1:

Descriptive Features of Included Studies

References	Country	Year Published	Patients (n)	Male (n)	Female (n)	Mean Age (years)	Treatment Groups
Aksu et al 14	Turkey	2006	80	41	39	-	Surgery + Radiation, Radiation
Bossi et al 15	Italy	2014	171	141	30	-	Surgery + Radiation, CRT
Chung et al 16	Korea	1997	33	30	3	58.4	Surgery, Radiation
Culie et al 11	France	2021	382	280	102	61.6	Surgery + CRT, CRT
Culie et al 12	France	2021	474	356	118	60.3	Surgery + CRT, CRT
Debenham et al 17	Canada	2016	233	203	30	56	Surgery + CRT, CRT
Fein et al 18	USA	1994	170	117	53	60	Surgery + Radiation, Radiation
Givens et al 19	USA	1981	104	-	-	60	Surgery, Surgery + Radiation, Radiation
Hicks et al 4	USA	1998	76	-	-	-	Radiation
Inagi et al 20	Japan	2002	87	73	14	-	Surgery, Surgery + CRT, Radiation, CRT
Jones et al 21	UK	1998	105	78	27	-	Surgery + Radiation, Radiation
Jones et al 22	UK	2007	123	82	41	-	Surgery + Radiation, Radiation
Kajanti et al 23	Finland	1991	75	54	21	-	Surgery + Radiation, Radiation
Kamran et al 24	USA	2018	22676	19036	3640	-	Surgery + CRT, CRT
Kano et al 25	Japan	2013	186	164	22	-	Surgery, CRT
Kao et al 26	Australia	2018	98	79	19	57.3	Surgery + CRT, CRT
Kim et al 27	Korea	2010	133	104	29	61.5	Surgery + CRT, CRT
Kobayashi et al 28	Japan	2016	62	45	17	63.6	Surgery, Surgery + Radiation, CRT
Kokemueller et al 29	Germany	2011	341	226	115	58.8	Surgery + CRT, CRT
Kravets et al 30	Ukraine	2020	211	176	35	56.1	Surgery + CRT, CRT
Kurita et al 31	Japan	2003	64	38	26	-	Surgery + CRT, CRT
Lanza et al 32	Italy	2000	115	108	7	63	Surgery + Radiation, CRT
Ling et al 33	China	2013	210	113	97	-	Surgery, Surgery + Radiation, CRT
Lybak et al 34	Norway	2017	226	174	52	62	Surgery + CRT, CRT
Mahmoud et al 35	USA	2018	1873	1571	302	-	Surgery + CRT, CRT
McDowell et al 36	Australia	2014	31	21	10	61.6	Surgery + CRT, CRT
Meccariello et al 37	Italy	2020	129	102	27	62.6	Surgery + CRT, CRT
O’Connell et al 38	Canada	2013	344	141	203	58.7	Surgery + Radiation, Surgery + CRT, Radiation, CRT
Park et al 39	Korea	2013	114	103	11	-	Surgery + Radiation, Surgery + CRT, CRT
Park et al 40	Korea	2018	102	85	17	-	Surgery + CRT, CRT
Rana et al 41	Germany	2011	398	224	174	49.5	Surgery + CRT, CRT
Rash et al 42	USA	2018	51	38	13	-	Surgery + CRT, CRT
Seikaly et al 43	Canada	2016	200	158	42	55.4	Surgery + CRT, CRT
Sessions et al 44	USA	2002	332	209	23	-	Surgery, Surgery + Radiation, Radiation
Sher et al 45	USA	2018	884	727	175	61.5	Surgery, Surgery + Radiation, Surgery + CRT, Radiation, CRT
Sher et al 46	USA	2019	2754	1978	776	-	Surgery + CRT, Radiation, CRT
Shirazi et al 47	USA	2006	74	61	13	-	Surgery + CRT, CRT
Smith et al 48	USA	2015	80	61	19	-	Surgery, Surgery + CRT, CRT
Song et al 49	Korea	2017	586	523	63	55	Surgery + CRT, CRT
Thakar et al 50	India	2021	146	135	11	54.6	*
Tsai et al 51	Taiwan	2021	1180	1052	128	-	Surgery, CRT
van de Pol et al 52	Netherlands	2004	88	64	24	53.4	Surgery + Radiation, Radiation
Yen et al 53	Taiwan	2022	60	57	3	-	*
Zubair et al 54	UK	2021	272			-	*

-Data not provided in text.

^*Reported on hazard ratios between treatment groups.

Figure 2: Risk of Bias.

Figure 3: Funnel Plot.

Funnel plot to assess for publication bias.

Patient Demographics

A total of 44 unique studies with (N=36,133) patients with OPSCC were included for analysis. The mean age was 58.9 (range 19–96.5) years. 77.2% (95% CI 74.0%−80.2%, p < 0.0001) were male and 22.0% (95% CI 19.1%−25.1%) were female. 89.6% (95% CI 83.6%−94.3%) of the included patients had advanced-stage disease, which was defined as AJCC stage III-IV by AJCC 7^th edition staging guidelines, due to the timing of the studies’ publication. Sufficient patient information required to re-stage patients according to AJCC 8^th edition guidelines was unavailable in the included studies, thus the 7^th edition staging was included for analysis. The cohort was well-distributed by HPV status: 37.2% (95% CI 17.4%−59.6%) of patients had HPV-positive disease, 36.9% (95% CI 21.7%−53.5%) had HPV-negative disease, and 34.5% (95% CI 15.9%−56.1%) of patients had disease with unknown HPV status. The most cited tumor primary site was the tonsil, followed by the lateral pharyngeal wall and base of tongue. 60.1% (95% CI 46.3%−73.1%) of patients endorsed current or past alcohol use and 71.8% (95% CI 61.2%−81.4%) of patients endorsed current or previous tobacco use. Characteristics of included patients are described in Table 2. Characteristics of each treatment group are described in Table 3, and comparative data between treatment groups is available in Supplemental Table S2.

Table 2:

Characteristics of Included Patients

Characteristic	Proportion of Patients	95% CI
Male	77.2%	74.0%−80.2%
Female	22.0%	19.1%−25.1%
Advanced-Stage (III-IV) Disease	89.6%	83.6%−94.3%
HPV-Positive	37.2%	17.4%−59.6%
HPV-Negative	36.9%	21.7%−53.5%
HPV-Unknown	34.5%	15.9%−56.1%
Tonsil Primary	55.1%	46.0%−64.0%
Lateral Pharyngeal Wall Primary	39.0%	13.2%−68.7%
Base of Tongue Primary	38.5%	30.7%−46.7%
Current/History of Alcohol Use	60.1%	46.3%−73.1%
No Alcohol Use	39.7%	26.8%−53.4%
Current/History of Tobacco Use	71.8%	61.2%−81.4%
No Tobacco Use	27.7%	18.3%−38.2%
Unknown Tobacco Use	6.1%	2.2%−11.8%

Table 3:

Characteristics of Treatment Groups

Characteristic	Surgery + CRT (n = 12,493)	CRT (n = 19,874)	Surgery + Radiation (n = 1,029)	Surgery Alone (n = 900)	Radiation Alone (n = 1,305)
Male	82.8% (95% CI 80.1%−85.3%)	80.5% (95% CI 77.4%−83.5%)	68.5% (95% CI 48.2%−85.7%)	84.7% (95% CI 77.4%−90.7%)	62.0% (95% CI 42.9%−79.4%)
Female	17.2% (95% CI 14.7%−20.0%)	19.5% (95% CI 16.5%−22.6%)	31.5% (95% CI 14.3%−51.8%)	15.3% (95% CI 9.3%−22.6%)	38.0% (95% CI 20.6%−57.1%)
Age (years)	58.0 (95% CI 55.7–60.3)	60.0 (95% CI 58.3–61.6)	58.3 (95% CI 55.0–61.5)	61.0 (95% CI 51.5–70.6)	62.6 (95% CI 56.8–68.4)
T1-T2	67.5% (95% CI 57.8%−76.5%)	41.6% (95% CI 35.8%−47.5%)	51.6% (95% CI 38.5%−64.7%)	61.4% (95% CI 43.1%−78.2%)	43.5% (95% CI 17.4%−71.7%)
T3-T4	31.6% (95% CI 22.8%−41.0%)	56.1% (95% CI 50.3%−61.9%)	56.3% (95% CI 33.5%−77.8%)	38.6% (95% CI21.8%−56.9%)	62.5% (95% CI 24.7%−93.0%)
N0-N1	37.8% (95% CI 30.7%−45.1%)	30.8% (95% CI 25.0%−37.0%)	42.7% (95% CI 25.2%−61.3%)	58.5% (95% CI 31.7%−82.8%)	57.6% (95% CI 51.0%−64.0%)
N2-N3	60.0% (95% CI 49.9%−69.6%)	68.6% (95% CI 63.7%−73.3%)	56.1% (95% CI 37.2%−74.1%)	28.9% (95% CI 0.0%−82.6%)	36.6% (95% CI 24.7%−49.5%)
Stage I/II	10.5% (95% CI 5.1%−17.5%)	4.7% (95% CI 2.1%−8.3%)	13.0% (95% CI 0.6%−37.6%)	24.6% (95% CI 1.0%−82.6%)	21.3% (95% CI3.0%−50.0%)
Stage III/IV	90.5% (95% CI 81.6%−96.7%)	96.6% (95% CI 93.3%−98.8%)	87.0% (95% CI 62.4%−99.4%)	81.8% (95% CI 37.8%−99.9%)	78.7% (95% CI 50.0%−97.0%)
HPV +	41.8% (95% CI 15.9%−70.4%)	35.3% (95% CI 13.4%−60.8%)	-	-	-
HPV -	31.8% (95% CI 16.5%−49.5%)	35.4% (95% CI 19.0%−53.7%)	-	-	-
Current/History of Alcohol Use	56.5% (95% CI 33.2%−78.4%)	73.4% (95% CI 67.6%−78.9%)	-	-	-
No Alcohol Use	42.9% (95% CI 21.3%−66.0%)	25.7% (95% CI 22.5%−29.1%)	-	-	-
Current/History of Tobacco Use	71.8% (95% CI 59.6%−82.7%)	74.2% (95% CI 61.5%−85.2%)	-	81.8% (95% CI 73.5%−88.3%)	-
No Tobacco Use	27.0% (95% CI 17.1%−38.2%)	17.5% (95% CI16.2%−40.6%)	-	18.2% (95% CI 11.7%−26.5%)	-

-Data not provided in text.

Surgery + CRT Group

There were (n = 12,493) patients in this treatment group. This treatment group was comprised of significantly more males (82.8%, 95% CI 80.1%−85.3%) than the CRT (p < 0.0001), Surgery + Radiation (p < 0.0001), and Radiation Alone (p < 0.0001) groups. Additionally, this group had a significantly higher proportion of T1-T2 disease (67.5%, 95% CI 57.8%−76.5%) than all other treatment groups (vs. CRT Alone p < 0.0001, vs. Surgery + Radiation p < 0.0001, vs. Surgery Alone p = 0.0002, versus Radiation Alone p < 0.0001), as well as a significantly higher proportion of HPV-positive disease (41.8%, 95% CI 15.9%−70.4%) than the CRT group (p < 0.0001). Finally, this treatment group had significantly lower proportions of alcohol (56.5%, 95% CI 33.2%−78.4%) and tobacco use (71.8%, 95% CI 59.6%−82.7%) than the CRT group (p < 0.0001). Additional comparative data between treatment groups is available in Supplemental Table S2. For patients with advanced disease, 5-year OS was 67.1% (95% CI 55.6%−77.6%). Survival outcomes consistently favored this treatment cohort over others. Surgery + CRT demonstrated significantly better 2-, 3-, and 5-year OS (2-year OS 82.0%, 95% CI 75.5%−87.5%; 3-year OS 79.2%, 95% CI 59.8%−93.4%; 5-year OS 67.1%, 95% CI 55.6%−77.6%) than CRT alone (p < 0.0001), as well as better 5-year OS (67.1%, 95% CI 55.6%−77.6%) than surgery + radiation (p = 0.0002) for patients with advanced disease. Additionally, this cohort demonstrated significantly better 2-, 3-, and 5-year DSS (2-year DSS 83.5%, 95% CI 66.7%−95.2%; 3-year DSS 89.7%, 95% CI 72.2%−99.0%; 5-year DSS 77.5%, 95% CI 66.2%−87.1%) than CRT alone (p < 0.0001) for patients with advanced disease. Additional outcomes for this treatment group are outlined in Tables 4–7.

Table 4:

2-Year Outcomes

Outcome	Surgery + CRT (n = 12,493)	CRT (n = 19,874)	Comparison of Proportions (%)
OS, Advanced-Stage	82.0% (95% CI 75.5%−87.5%)	53.9% (95% CI 29.7%−77.2%)	28.1 (95% CI 27.1%−29.1%, p < 0.0001)
DSS, Advanced-Stage	83.5% (95% CI 66.7%−95.2%)	45.4% (95% CI 19.2%−73.0%)	38.1 (95% CI 37.1%−39.1%, p < 0.0001)

OS = Overall Survival, DSS = Disease-Specific Survival

Table 7:

Loco- and Locoregional Control Outcomes

	Surgery + CRT (n = 12,493)	CRT (n = 19,874)	Surgery + Radiation (n = 1,029)	Comparison of Proportions (%)
LC, Advanced Stage	81.1% (95% CI 44.0%−99.7%)	80.4% (95% CI 73.0%−86.6%)	71.7% (95% CI 24.6%−99.5%)	0.7 (95% CI −0.2%−1.6%, p = 0.1205)*; 9.4 (95% CI5.2%−13.9%, p < 0.0001)^; 8.7 (95% CI 4.6%−13.2%, p < 0.0001)~
LRC, Advanced Stage	78.2% (95% CI 64.6%−89.3%)	85.7% (95% CI 78.4%−91.3%)	80.9% (95% CI 63.9%−93.4%)	7.5 (95% CI 6.6%−8.4%, p < 0.0001)*; 2.8 (95% CI –1.4%−6.2%, p = 0.1863)^; 4.8 (95% CI 1.3%−8.8%, p = 0.0055)

LC = Local Control, LRC = Locoregional Control

^*Comparing Surg + CRT and CRT

^{^}Comparing Surg + CRT vs Surg + Radiation, ~Comparing CRT and Surg + Radiation

CRT Group

There were (n = 19,874) patients in this treatment group. This treatment group had a significantly higher proportion of T3-T4 disease (56.1%, 95% CI 50.3%−61.9%) than the Surgery + CRT (p < 0.0001) and Surgery Alone (p < 0.0001) groups, as well as a significantly higher proportion of N2-N3 disease (68.6%, 95% CI 63.7%−73.3%) than all other treatment groups (vs. Surgery + CRT p < 0.0001, versus Surgery + Radiation p < 0.0001, vs. Surgery Alone p < 0.0001, vs. Radiation Alone p < 0.0001). Finally, this treatment group had significantly higher proportions of alcohol (73.4%, 95% CI 67.6%−78.9%) and tobacco use (74.2%, 95% CI 61.5%−85.2%) than the Surgery + CRT treatment group (p < 0.0001). Additional comparative data between treatment groups is available in Supplemental Table S2. For patients with advanced disease, 5-year OS was 53.9% (95% CI 37.7%−69.7%). When compared with the surgery + CRT cohort, this group demonstrated significantly worse 2-, 3-, and 5-year OS (2-year OS 53.9%, 95% CI 29.7%−77.2%; 3-year OS 70.2%, 95% CI 57.1%−81.8%; 5-year OS 53.9%, 95% CI 37.7%−69.7%; p < 0.0001; see Tables 4–6) for patients with advanced stage disease. Additionally, this group demonstrated significantly worse locoregional control (85.7%, 95% CI 78.4%−91.3%) than the surgery + CRT cohort in patients with advanced disease (p < 0.0001, see Table 7). Additional outcomes for this treatment group are outlined in Tables 4–7.

Table 6:

5-Year Outcomes

Outcome	Surgery + CRT (n = 12,493)	CRT (n = 19,874)	Surgery + Radiation (n = 1,029)	Comparison of Proportions (%)
OS, Advanced-Stage	67.1% (95% CI 55.6%−77.6%)	53.9% (95% CI 37.7%−69.7%)	63.1% (95% CI 31.9%−89.2%)	13.2 (95% CI 12.1%−14.3%, p < 0.0001)*; 4.0 (95% CI –0.6%−8.8%, p = 0.0863)^; 9.2 (95% CI 4.4%−13.7%, p = 0.0002)~
DSS, Advanced-Stage	77.5% (95% CI 66.2%−87.1%)	63.5% (95% CI 49.1%−76.7%)	55.9% (95% CI 49.1%−62.5%)	14.0 (95% CI 13.0%−15.0%, p < 0.0001)*; 21.6 (95% CI 16.9%−26.4%, p < 0.0001)^; 7.6 (95% CI −16.8%−26.9%, p = 0.5480)~
DFS, Advanced-Stage	58.6% (95% CI 50.8%−66.1%)	39.7% (95% CI 2.1%−87.8%)	-	18.9 (95% CI 17.8%−20.0%, p < 0.0001)*
RFS, Advanced-Stage	70.9% (95% CI 49.1%−88.6%)	69.1% (95% CI 42.0%−90.4%)	71.9% (95% CI 35.2%−96.7%)	1.8% (95% CI 0.7%−2.9%, p = 0.0019)*; 1.0 (95% CI-3.5%−5.2%, p = 0.6568)^; 2.8 (95% CI –1.7%−6.9%, p = 0.2177)~

OS = Overall Survival, DSS = Disease-Specific Survival, DFS = Disease-Free Survival, RFS = Recurrence-Free Survival

-Data not provided in text.

^*Comparing Surg + CRT and CRT,

^{^}Comparing Surg + CRT vs Surg + Radiation, ~Comparing CRT and Surg + Radiation

Surgery + Radiation Group

There were (n = 1,029) patients in this treatment group. This treatment group had a significantly higher proportion of female patients (31.5%, 95% CI 14.3%−51.8%) than all groups (vs. Surgery + CRT p < 0.0001, vs. CRT p < 0.0001, vs. Surgery Alone p < 0.0001) except Radiation Alone. Additional comparative data between treatment groups is available in Supplemental Table S2. The 5-year overall survival (OS) was 63.1% (95% CI 31.9%−89.2%) for patients with advanced disease in this group. When compared with the surgery + CRT cohort, this group demonstrated 21.6% lower 5-year DSS (95% CI 16.9%−26.4%, p < 0.0001), as well as 8.7% lower rates of local control (95% CI 4.6%−13.2%, p < 0.0001). Additional outcomes for this treatment group are outlined in Tables 6 and 7.

Surgery Alone Group

There were (n = 900) patients in this treatment group. This treatment group had the highest proportion of male patients (84.7%, 95% CI 77.4%−90.7%) of all treatment groups, as well as the highest rate of N0-N1 disease (58.5%, 95% CI 31.7%−82.8%) of all treatment groups. This group also had the highest proportion of tobacco use (81.8%, 95% CI 73.5%−88.3%), which was significantly higher than all other treatment groups (vs. Surgery + CRT p < 0.0001, vs. CRT p < 0.0001). Additional comparative data between treatment groups is available in Supplemental Table S2. The 5-year DSS for patients with advanced stage disease in this treatment group was 69.5% (95% CI 24.3%−98.8%) and LRC was 68.6% (95% CI 21.0%−99.1%).

Radiation Alone Group

There were (n = 1,305) patients in this treatment group. This treatment group had the highest proportion of female patients (62.0%, 95% CI 42.9%−79.4%), which was significantly higher than all other treatment groups (vs. Surgery + CRT p < 0.0001, vs. CRT p < 0.0001, vs. Surgery + Radiation p = 0.0011, vs. Surgery Alone p < 0.0001), as well as the highest average age of any treatment group. Additional comparative data between treatment groups is available in Supplemental Table S2. The 5-year DSS was 31.9% (95% CI 18.3%−47.2%) and LRC was 62.2% (95% CI 18.8%−96.0%) for patients with advanced disease in this treatment group. When compared with surgery + CRT, this cohort demonstrated 45.6% lower 5-year DSS (95% CI 41.9%−49.1%, p < 0.0001). Additionally, the rates of locoregional control (LRC) for this cohort were 16.0% lower (95% CI 12.4%−19.8%, p < 0.0001) than for surgery + CRT.

Survival Outcome Comparison

Survival outcomes consistently favored Surgery + CRT over other primary treatment modalities. Surgery + CRT as compared to Surgery + Radiation demonstrated 5-year OS (HR = 1.3729, 95% CI 1.1274–1.6719, p = 0.0016) and 5-year DSS (HR = 2.1513, 95% CI 1.7303–2.6748, p < 0.0001). Surgery + CRT as compared to CRT only demonstrated 5-year OS (HR = 1.4355, 95% CI 1.2569–1.6396, p < 0.0001) and 5-year DSS (HR 1.5548, 95% CI 1.2783–1.891, p < 0.0001). These results are demonstrated in Table 8.

Table 8:

Hazard Ratios (HRs) for 5-Year Survival Outcomes

Treatment	5-Year Overall Survival	5-Year Disease-Specific Survival
Surgery + CRT	Reference	Reference
CRT	1.4355 (95% CI 1.2569–1.6395, p < 0.0001)	1.5548 (95% CI 1.2783, 1.1891, p < 0.0001)
Surgery + Radiation	1.3729 (95% CI 1.1274, 1.6719, p = 0.0016)	2.1513 (95% CI 1.7303, 2.6748, p < 0.0001)

Discussion

In the face of increasing incidence of OPSCC and advanced-stage disease presentations, there is considerable controversy as to whether to pursue radiation-based or surgery-based treatment modality. This systematic review and meta-analysis attempts to add to the literature by comparing survival outcomes of advanced staged oropharyngeal squamous cell carcinoma treated by surgical and non-surgical methods.

This systematic review and meta-analysis demonstrates that primary surgical treatment with combined chemoradiotherapy (surgery + CRT) may be superior to other treatment modalities for advanced stage oropharyngeal cancer, as measured by survival outcomes. Data demonstrated that surgery + CRT showed improved 5-year OS as compared to surgery + radiation (HR = 1.3729, 95% CI 1.1274–1.6719, p = 0.0016) and CRT alone (HR = 1.4355, 95% CI 1.2569–1.6396, p < 0.0001), as well as improved 5-year DSS over surgery + radiation (HR = 2.1513, 95% CI 1.7303–2.6748, p < 0.0001) and CRT alone (HR 1.5548, 95% CI 1.2783–1.891, p < 0.0001) for patients with advanced stage disease. The significance of these findings cannot be understated when considering the morbidity and mortality of advanced stage oropharyngeal squamous cell carcinoma, which often presents in advanced stages, with reports demonstrating up to 84% of patients presenting to cancer referral centers having advanced disease at the time of presentation.³

These results, however, must be interpreted in the context of other factors influencing treatment groups. Due to the retrospective nature of the available data, the treatment groups studied have differences that must be considered when interpreting results. The Surgery + CRT group, which demonstrated significantly improved survival outcomes compared to other treatment groups, was comprised of more T1–2 disease, more HPV-positive disease, and patients with less alcohol and tobacco use. Each of these modifiers is correlated with better disease outcomes than more advanced T-stage, HPV-negative disease, and increased alcohol and tobacco use. Additionally, to be considered for surgery, these patients had to have disease anatomically favorable or accessible for surgical intervention. Thus, the Surgery + CRT group likely was selected at time of treatment for their surgical and adjuvant course based on their lower burden of disease. Despite the interent differences in treatment groups, however, this data demonstrates that patients undergoing Surgery + CRT for their advanced-stage OPSCC had favorable survival outcomes compared to patients undergoing primary non-surgical treatment courses, suggesting that when patients with advanced-stage OPSCC are good surgical candidates, primary surgical treatment of their disease is favorable for survival outcomes.

This information is, to our knowledge, the first of its kind. Multiple recent studies, including Culie et al, have demonstrated that upfront surgical treatment of OPSCC, including advanced-stage disease, may yield increased overall survival, RFS, and DFS, as compared to upfront nonsurgical treatment.^11,12 However, this is the first systematic review and meta-analysis to our knowledge to demonstrate that these survival benefits may be powered to influence clinical practice. For this reason, we believe that this data could prove to be integral for clinical practice. Though contraindications to surgical treatment are not uncommon, this data demonstrates that, when possible, patients may significantly benefit from surgical treatment with CRT in the setting of advanced stage OPSCC, as evidenced by increased overall survival.

Limitations

There are several limitations that should be considered for this systematic review and meta-analysis. Most notably, this meta-analysis sought to capture survival data about advanced-stage OPSCC disease and to stratify results by HPV status. Unfortunately, many studies reported data about advanced-stage disease outcomes, regardless of HPV status, or about HPV-dependent disease outcomes, regardless of disease stage. For these reasons, we were unable to collect data and perform meta-analysis of HPV-positive and HPV-negative advanced stage disease outcomes. Given our knowledge of the different behaviors of HPV-positive and HPV-negative disease, it is both a limitation of this study and a call for further research that we were unable to collect and analyze data about advanced stage disease, stratified by HPV status.

A second limitation is again influenced by HPV status. The AJCC 8^th edition head and neck cancer staging guidelines took effect in January 2018. These guidelines significantly changed the way head and neck cancers are staged, including OPSCC, as staging is now affected by HPV status of the disease.¹³ Given the range of publication dates, as well as the retrospective nature of all studies included in this systematic review and meta-analysis, the 7^th edition guidelines were used to stage the cases we included in this study. There will be a lag in survival outcomes data for tumors staged with the new, 8^th edition guidelines, and this should prompt further research on the topic.

Additionally, though we sought to collect data regarding severe complications from treatment and morbidity following treatment protocols, few studies reported on these outcomes, and we were ultimately unable to collect this data. This data is integral to extrapolating the clinical significance of our findings, as no invasive treatment is without risk and survival must be interpreted within the context of patient morbidity. For this reason, further research is needed about the morbidity and complications of surgery + CRT in the treatment of advanced stage oropharyngeal squamous cell carcinoma in order to fully contextualize these results.

Finally, our treatment groups included a heterogeneous variety of treatment protocols, including surgical approaches, chemotherapeutic protocols, and radiation strategies. This limitation is inherent in studies comparing treatment groups between institutions and internationally but is nevertheless present in this study and must be considered when interpreting its results. Most importantly, there is inherent selection bias in the included studies which guides the patients who are selected for surgical or nonsurgical therapy. Patients with poorer functional status or worse comorbidities may be selected for nonsurgical therapy.

Conclusions

Our systematic review and meta-analysis suggests that, for patient with advanced-stage oropharyngeal squamous cell carcinoma, primary treatment with surgery and combined chemoradiotherapy may be superior to non-surgical or surgical-only treatment modalities, as measured by survival outcomes. Our findings strengthen the evidence to support surgical management of head and neck cancer when possible and serve as a call for additional research about the superior treatment modalities for HPV-positive and HPV-negative advanced-stage disease.

Supplementary Material

Supinfo

Supplemental Appendix S1: Search Strategies

Search strategies utilized for article selection.

Supplemental Table S2: Comparative Data Between Treatment Groups

Comparative data between treatment groups with p-values.

Table 5:

3-Year Outcomes

Outcome	Surgery + CRT (n = 12,493)	CRT (n = 19,874)	Comparison of Proportions (%)
OS, Advanced-Stage	79.2% (95% CI 59.8%−93.4%)	70.2% (95% CI 57.1%−81.8%)	9.0 (95% CI 8.0%−10.0%, p < 0.0001)
OS, HPV-Positive	93.4% (95% CI 90.6%−95.7%)	88.5% (95% CI 83.4%−92.8%)	4.9 (95% CI 3.5%−6.3%, p < 0.0001)
OS, HPV-Negative	80.5% (95% CI 77.5%−83.3%)	65.0% (95% CI 62.6%−67.4%)	15.5 (95% CI 11.7%−19.1%, p < 0.0001)
DSS, Advanced-Stage	89.7% (95% CI 72.2%−99.0%)	81.1% (95% CI 58.7%−96.0%)	8.6 (95% CI 7.8%−9.4%, p < 0.0001)
RFS, Advanced-Stage	78.6% (95% CI 53.8%−95.5%)	61.2% (95% CI 33.9%−85.2%)	17.4 (95% CI 16.4%−18.4%, p < 0.0001)

OS = Overall Survival, DSS = Disease-Specific Survival, RFS = Recurrence-Free Survival