Oncological Outcomes of Salvage Oropharyngectomy in Irradiated Neck for Recurrent or Metachronous Squamous Cell Carcinoma

ABSTRACT

Objective

To evaluate oncological outcomes and their predictors following salvage oropharyngectomy.

Methods

This single‐center retrospective study included patients who underwent a salvage oropharyngectomy in an irradiated neck for recurrent or metachronous oropharyngeal squamous cell carcinoma (OPSCC) between 2014 and 2023.

Results

Ninety‐four patients were included. Five‐year overall survival (OS), disease‐specific survival (DSS), and local recurrence‐free survival (LRFS) were 31.7%, 58.5%, and 55.2%, respectively. In multivariable analysis, pN+ status (HR 2.32; p = 0.031), pT status ≥ 3 (HR 2.03; p = 0.020), and age (HR 1.05; p = 0.014) were associated with OS. DSS was associated with pT status ≥ 3 (HR 5.24; p < 0.001), metachronous (vs. recurrent) OPSCC (HR 0.32; p = 0.006), and cN+ status (HR 2.31; p = 0.053). LRFS was associated with metachronous OPSCC (HR 0.30; p = 0.002) and pT status ≥ 3 (HR 3.60; p = 0.001).

Conclusions

This last‐resort procedure is associated with poor survival outcomes. In this series, patients with p16‐positive OPSCC did not fare better than their p16‐negative counterparts.

1. Introduction

The incidence of oropharyngeal squamous cell carcinoma (OPSCC), in contrast with other head and neck subsites, has risen over the past decades. It is mainly due to the emergence of human papillomavirus (HPV)‐related OPSCC, especially in developed countries. Whereas preventive measures targeting tobacco and alcohol consumption have led to a decline in the incidence of HPV‐negative OPSCC, the prevalence of HPV‐positive OPSCC continues to rise, representing over 80% of OPSCC in certain regions across the United States [1].

Radiotherapy (RT) or chemoradiotherapy (CRT) is often considered first‐line treatment, as it is often associated with better functional outcomes compared to upfront surgery and acceptable survival outcomes, especially for HPV‐positive OPSCC [2]. The increased use of RT has consequently led to a rise in surgical procedures performed in previously irradiated areas.

Indeed, with a 5‐year local recurrence rate estimated at 4%–7% [3, 4], and a rate of metachronous head and neck cancer of about 13 per 1000‐person‐year [5], salvage oropharyngectomy will be considered in a number of cases. Given the poor survival and functional outcomes, careful patient selection and informed consent are paramount to perform this last‐resort surgery [6].

The literature lacks robust data regarding prognostic data as well as patient's and tumor's operability criteria in the specific context of salvage oropharyngectomy, especially in the new era of HPV‐related OPSCC, intensity‐modulated radiation therapy (IMRT), and avoidance of mandibulotomy. Existing series are often dated and/or sometimes include non‐oropharyngeal tumor sites [6, 7, 8, 9, 10]. While this is a common clinical situation, most studies have relatively small sample sizes [11, 12, 13]. Furthermore, the influence of HPV status in the oncological outcome after such surgery remains to be elucidated. Given the evolving epidemiology, the role of HPV or p16 status warrants further investigation in contemporary cohorts.

The aim of this study is to evaluate oncological outcomes following salvage oropharyngectomy in the irradiated neck for recurrent or metachronous OPSCC, in the modern era of high HPV prevalence, IMRT, and mandibulotomy avoidance. We also aimed to identify prognostic factors for survival, with particular emphasis on p16 status.

2. Materials and Methods

This single center retrospective study was conducted in a French tertiary referral center between January 2014 and July 2023. We included all patients who underwent a salvage oropharyngectomy, that is, surgery performed in an irradiated neck for recurrent or metachronous OPSCC. We excluded patients who received a RT dose < 50 Gy to the head and neck region, tumors from other subsites which secondarily invaded the oropharynx, non‐SCC histologies, patients with distant metastasis, or patients with a synchronous invasive head and neck SCC.

All cases were discussed in multidisciplinary team meetings, and indications for surgery followed international guidelines [14]. All patients received preoperative anesthetic assessment, and a comprehensive geriatric assessment was performed in elderly patients.

Data were collected in the patients' electronic medical record. They included preoperative data (demographics, patient history, tumor characteristics), perioperative data (type of surgery performed, pathology results), and survival status at last follow‐up. Tumors were staged with both the seventh and eighth edition of the American Joint Cancer Committee (AJCC) TNM classification to compare their accuracy of prognostication. The main endpoints were survival outcomes, that is, overall survival (OS), disease‐specific survival (DSS), local recurrence‐free survival (LRFS). The postoperative and functional outcomes of that same cohort have also been studied and are the subject of another article.

This study was conducted within a General Data Protection Regulation‐compliant and secure system, in accordance with the French legal framework of MR‐004 set up by the National Commission on Informatics and Liberty. This study received approval by the local research committee and was recorded in the Health Data Hub (reference: 19270385).

Data were summarized by median and range (min; max) for continuous variables, and by frequency and percentage for qualitative variables. The number of missing data was reported for each variable. OS was defined as the time between the date of surgery and the date of death from any cause; patients alive were censored at last follow‐up news. DSS was defined as the time between the date of surgery and the date of cancer‐related death. Patients alive were censored at the date of last contact, and patients who died from other causes were censored at their date of death. LRFS was defined as the time from the date of surgery to the date of first local recurrence. Patients without local recurrence were censored at the date of last contact or date of death. Survival rates were estimated using the Kaplan–Meier method with their 95% confidence interval (95% CI). Univariable and multivariable analyses were performed using the Logrank test and Cox proportional hazards model; hazard ratios (HRs) were estimated with 95% CIs. Variable selection for multivariable analyses was based on clinical relevance and results of univariable analyses. All analyses were performed with STATA 19 software; all tests used were two‐sided, and a p < 0.05 was considered statistically significant.

3. Results

A total of 94 patients (72 males and 22 females, M/F gender ratio = 3.27/1) met the eligibility criteria. The main patient's characteristics are shown in Table 1.

TABLE 1.

Patient characteristics.

	Total (N = 94)
Age at the time of the surgery (year) (n = 94)
Median (range)	62 (42.0; 83.0)
BMI (n = 93)
Median (range)	22.8 (13,6; 33,9)
Underweight	17 (18.3%)
Healthy weight	50 (53.8%)
Overweight	20 (21.5%)
Obese	6 (6.5%)
Missing	1
ASA score (n = 93)
1	4 (4.3%)
2	61 (65.6%)
3	27 (29.0%)
4	1 (1.1%)
Missing	1
History of other head and neck cancer (n = 93)
No	32 (34.4%)
Yes	61 (65.6%)
Missing	1
Smoker (n = 93)
Never smoker	3 (3.2%)
Current smoker	29 (31.2%)
Former smoker	61 (65.6%)
Missing	1
Tobacco use (n = 92)
No or less than 20 pack‐years	9 (9.8%)
Yes and more than 20 pack‐years	83 (90.2%)
Missing	2
Alcohol abuse (> 20 or 30 g/day according to gender) (n = 88)
No	36 (40.9%)
Yes, current	16 (18.2%)
Yes, former	36 (40.9%)
Missing	6

Abbreviations: ASA = American Society of Anesthesiologists, BMI = body mass index.

Patient's primary treatment always included RT, but also surgery and chemotherapy (CT) in 47.3% and 65.6% of cases, respectively. The indication for the primary treatment was predominantly OPSCC, accounting for 54.8% of cases. Other head and neck cancer sites were represented in similar proportions, namely the oral cavity, larynx, and hypopharynx (14%, 14%, and 16.1%, respectively). This treatment occurred a median time of 45.2 months before the secondary treatment, that is, the salvage oropharyngectomy.

Tumor characteristics are displayed in Tables 2 and 3. Among patients who presented with nodal metastasis on pathological examination, the Levels IA, IB, IIA, IIB, III, IV, and V were involved in 7.7%, 23.1%, 76.9%, 15.4%, 7.7%, 15.4%, and 7.7%, respectively. All patients with a nodal metastasis in Level IIB (n = 2) had a metastasis in Level IIA. The only patient with a nodal metastasis in Level V had a history of Level I–IV neck dissection. Due to the relatively low number of patients with nodal metastasis and the marginal differences in N‐status between the seventh and eighth AJCC staging system, we could not highlight a prognostic difference between the two classifications. Finally, Table 4 shows the characteristics of the surgical and postoperative treatment of the OPSCC.

TABLE 2.

Tumor clinical data.

	Total (N = 94)
Tumor origin (n = 94)
Tonsil	22 (23.4%)
Base of tongue	22 (23.4%)
Glossotonsillar sulcus	26 (27.7%)
Soft palate	14 (14.9%)
Posterior pharyngeal wall	6 (6.4%)
Vallecula	4 (4.3%)
Tumor extension (n = 94)
Tonsil	32 (34.0%)
Base of tongue	42 (44.7%)
Glossotonsillar sulcus	48 (51.1%)
Soft palate	34 (36.2%)
Vallecula	10 (10.6%)
Larynx	2 (2.1%)
Hypopharynx	6 (6.4%)
Oral cavity	27 (28.7%)
Crossing the midline	31 (33.0%)
Context (n = 94)
Metachronous OPSCC	53 (56.4%)
Recurrent OPSCC (persistent or recurrent disease)	41 (43.6%)
If recurrent OPSCC (n = 41)
Persistent disease (< 6 months)	11 (26.8%)
Recurrent disease (> 6 months)	30 (73.2%)

Abbreviation: OPSCC = oropharyngeal squamous cell carcinoma.

TABLE 3.

Tumor pathological data.

	Total (N = 94)
p16 status (n = 84)
Negative	68 (81.0%)
Positive	16 (19.0%)
Missing	10
Differentiation (n = 88)
Poorly differentiated	12 (13.6%)
Moderately differentiated	40 (45.5%)
Well differentiated	36 (40.9%)
Missing	6
Size (in mm) (n = 90)
Median (range)	24.5 (2.0; 80.0)
Missing	4
Depth of invasion (in mm) (n = 78)
Median (range)	7.0 (0.5; 35.0)
Missing or in situ	16
Mucosal margin (n = 93)
Positive	12 (12.9%)
Clear	81 (87.1%)
Missing	1
If clear mucosal margin, minimum distance (in mm) (n = 76)
0–2 mm	36 (47.4%)
≥ 3 mm	40 (52.6%)
Missing	5
Deep margin (n = 93)
Positive	7 (7.5%)
Clear	86 (92.5%)
Missing	1
If clear deep margin, minimum distance (in mm) (n = 68)
0–2 mm	28 (41.2%)
≥ 3 mm	40 (58.8%)
Missing	18
Excision margin (n = 93)
Positive mucosal or deep margin	15 (16.1%)
Clear mucosal and deep margins	78 (83.9%)
Missing	1
Perineural invasion (n = 94)
No	42 (50.0%)
Yes	42 (50.0%)
Missing	10
Lymphovascular embolism (n = 91)
No	67 (73.6%)
Yes	24 (26.4%)
Missing	3
Extranodal extension (n = 94)
No	85 (90.4%)
Yes	9 (9.6%)
pT (AJCC 7th edition) (n = 93)
1	37 (39.8%)
2	34 (36.6%)
3	13 (14.0%)
4a	4 (4.3%)
Tis	5 (5.4%)
Missing	1
pN (AJCC 7th edition) (n = 94)
x (no neck dissection)	38 (40.4%)
0	43 (45.7%)
1	5 (5.3%)
2b	3 (3.2%)
2c	5 (5.3%)
pN (AJCC 8th edition) (n = 94)
x (no neck dissection)	38 (40.4%)
0	43 (45.7%)
1	8 (8.5%)
2c	1 (1.1%)
3b	4 (4.3%)
cN (n = 94)
cN0	77 (81.9%)
cN+	17 (18.1%)
If neck dissection performed, occult nodal metastasis (cN0pN+) (n = 56)
No	53 (94.6%)
Yes	3 (5.4%)

Abbreviation: AJCC = American Joint Committee of Cancer.

TABLE 4.

Treatment characteristics.

	Total (N = 94)
Surgery
Surgical approach (n = 94)
Transoral	28 (29.8%)
Transcervical‐transoral	50 (53.2%)
Transcervical	9 (9.6%)
TORS	7 (7.4%)
Neck dissection (n = 94)
No	38 (40.4%)
Unilateral	37 (39.4%)
Bilateral	19 (20.2%)
Mandibulectomy (n = 94)
No	79 (84.0%)
Marginal	8 (8.5%)
Segmental	5 (5.3%)
Mandibulotomy	2 (2.1%)
Reconstruction (n = 94)
Free flap	57 (60.6%)
Pedicled flap	16 (17.0%)
Primary closure or local flap	10 (10.6%)
Secondary healing	11 (11.7%)
Tracheostomy (n = 94)
No	32 (34.0%)
Yes	62 (66.0%)
Adjuvant treatment
None	53 (56.4%)
Yes	41 (43.6%)
Reirradiation (n = 41)
No	16 (39.0%)
Yes	25 (61.0%)
Re‐resection for positive margins (n = 41)
No	30 (73.2%)
Yes	11 (26.8%)
Chemotherapy (n = 41)
No	36 (87.8%)
Yes	5 (12.2%)
Immunotherapy (n = 41)
No	36 (87.8%)
Yes	5 (12.2%)

Abbreviation: TORS = transoral robotic surgery.

After a median follow‐up of 74.3 months [56.0; 95.2], 45.7% of patients experienced at least one recurrence. The main pattern of recurrence was local. The first recurrence occurred in the primary site (79.1%) and/or neck (30.2%) and/or in the form of distant metastasis (14.3%).

Five‐year OS, DSS, and LRFS were 31.7% [95% CI = 21.7–42.1], 58.5% [95% CI = 44.3–70.3], and 55.2% [95% CI = 41.7–66.8], respectively. The median OS was 30.8 months [95% CI = 23.8–40.9] while the median LRFS was 71.3 months [95% CI = 28.3—Not reached]. Among the 66 patients (70.2%) who died during follow‐up, the cause was the OPSCC (45.3%), treatment toxicity (9.4%), or other causes (45.3%).

In univariable analysis, factors significantly associated with worse OS were cN+ status (HR 2.11 [95% CI = 1.19–3.73]; p = 0.009), pT status ≥ 3 (HR 2.27 [95% CI 1.27–4.07]; p = 0.005), as well as a statistical trend for pN+ status (vs. pN0/pNx, HR 1.73 [95% CI 0.90–3.32]; p = 0.096) and older age at surgery (HR 1.03 [95% CI 1.00–1.06]; p = 0.095). Importantly, p16‐positive status (HR 0.80 [95% CI 0.40–1.57]; p = 0.509) was not associated with OS. Neither were all the other studied variables, for example, American Society of Anesthesiologists (ASA) score, body mass index (BMI), smoking status, alcohol consumption, cancer status (recurrence vs. residual disease vs. metachronous OPSCC), tumor origin, midline involvement, pathologic factors.

In multivariable analysis, p16 status was not associated with OS in a model adjusted on independent prognostic factors for OS: pN+ status (HR 2.32 [95% CI 1.08–4.98]; p = 0.031), pT status ≥ 3 (HR 2.03 [95% CI 1.12–3.68]; p = 0.020), and age at surgery (HR 1.05 [95% CI 1.01–1.08]; p = 0.014) (Table 5). Figure 1 shows the OS according to p16‐status and smoking status; however, the number of events did not allow us to perform multivariable analysis with stratification by p16‐status and smoking status at the same time.

TABLE 5.

Results of multivariable analyses for overall survival, disease‐specific survival, and local recurrence‐free survival.

	HR [95% CI]	p
Overall survival
Age at surgery (years)	1.05 [1.01; 1.08]	0.014
p16‐status
Negative	1.00
Positive	0.55 [0.25; 1.20]	0.132
pT‐status
1–2	1.00
3–4	2.03 [1.12; 3.68]	0.020
pN‐status
Nx/N0	1.00
N+	2.32 [1.08; 4.98]	0.031
Disease‐specific survival
Context
Recurrent OPSCC	1.00
Metachronous OPSCC	0.32 [0.14; 0.72]	0.006
pT‐status
1–2	1.00
3–4	5.24 [2.41; 11.39]	< 0.001
cN‐status
N0	1.00
N+	2.31 [0.99; 5.41]	0.053
Local recurrence‐free survival
p16‐status
Negative	1.00
Positive	1.32 [0.56; 3.13]	0.525
Context
Recurrent OPSCC	1.00
Metachronous OPSCC	0.30 [0.14; 0.65]	0.002
pT‐status
1–2	1.00
3–4	3.60 [1.65; 7.82]	0.001

Note: Bold indicates statistically significant p‐values; italic indicates statistical trends.

Abbreviation: OPSCC = oropharyngeal squamous cell carcinoma.

FIGURE 1.

Kaplan–Meier overall survival curves for p16‐negative patients and p16‐positive patients with and without history of smoking > 20 pack‐years.

In univariable analysis, factors associated with DSS were cN+ status (HR 2.84 [95% CI 1.28–6.29]; p = 0.007), pT status ≥ 3 (HR 5.38 [95% CI 2.52–11.47]; p < 0.001), greater tumor size (HR 1.04 [95% CI 1.02–1.06]; p < 0.001), perineural invasion (HR 3.58 [95% CI 1.51–8.48]; p = 0.002), positive versus clear mucosal and deep margins (HR 0.42 [95% CI 0.19–0.96]; p = 0.033), margin status in three groups (HR 0.23 [95% CI 0.07–0.76] if mucosal and deep margins ≥ 3 mm, HR 0.52 [95% CI 0.23–1.22] if mucosal and deep margins clear < 3 mm; p = 0.037), and metachronous OPSCC (HR 0.34 [95% CI 0.16–0.73]; p = 0.004). There was also a statistical trend for pN+ status (HR 2.18 [95% CI 0.88–5.39]; p = 0.082) and lymphovascular embolism (HR 2.09 [95% CI 0.95–4.59]; p = 0.061). Again, p16‐positive status (HR 1.58 [95% CI 0.69–3.61]; p = 0.276) was not associated with DSS.

In multivariable analysis, only the following variables remained significant: pT status ≥ 3 (HR 5.24 [95% CI 2.41–11.39]; p < 0.001) and metachronous OPSCC (HR 0.32 [95% CI 0.14–0.72]; p = 0.006), with cN+ status at the limit of significance (HR 2.31 [95% CI 0.99–5.41]; p = 0.053) (Table 5). In another model with p16 status adjusted on pT and cN, results were similar.

In univariable analysis, factors associated with better LRFS were metachronous OPSCC (HR 0.32 [95% CI 0.16–0.63]; p < 0.001), first cancer located outside the oropharynx (HR 0.40 [95% CI 0.19–0.85]; p = 0.014). On the other hand, pT status ≥ 3 (HR 3.14 [95% CI 1.52–6.50]; p = 0.001), greater tumor size (HR 1.03 [95% CI 1.02–1.05]; p < 0.001), and perineural invasion (HR 3.24 [95% CI 1.53–6.88]; p = 0.001) were associated with worse LRFS. Of note, the following variables were not associated with LRFS: clear margins (HR 0.80 [95% CI 0.33–1.93]; p = 0.616) and p16‐status (HR 1.81 [95% CI 0.83–3.93]; p = 0.127).

In a multivariable analysis model with p16 status adjusted on independent prognostic factors for LRFS, metachronous OPSCC (HR 0.30 [95% CI 0.14–0.65]; p = 0.002) and pT status ≥ 3 (HR 3.60 [95% CI 1.65–7.82]; p = 0.001) were still associated with LRFS (Table 5). Overall, p16 status was never associated with any survival in multivariable analysis.

4. Discussion

Salvage oropharyngectomy in irradiated neck for recurrent or metachronous OPSCC remains a last‐resort procedure associated with relatively poor survival, that is, 5‐year OS, DSS, and LRFS of 31.7%, 58.5%, and 55.2% in our study. The main pattern of recurrence was local (79.1%) or regional (30.2%) rather than in the form of distant metastasis (14.3%). The main independent prognostic factors for survival were an advanced T‐status, a positive N‐status, and a recurrent (vs. metachronous) OPSCC.

Relatively few OPSCC‐related prognostic factors for OS were identified in our series. This can be explained by the fact that most patients from our cohort were current or former smokers with comorbidities, and sometimes had several cancers. This resulted in a high number of deaths by other causes (45.3%). When analyzing DSS, the number of identified prognostic factors was logically higher.

Table 6 summarizes all published series of salvage oropharyngectomies in irradiated neck, for recurrent and/or metachronous OPSCC, alongside the risk factors for survival highlighted. Our series is the first to find that T‐status was associated with worse OS or LRFS in multivariable analysis. Only one study with a more heterogeneous cohort, including patients without previous RT, made a similar observation. T‐status was an independent prognostic factor for worse locoregional recurrence‐free survival after TORS resection of recurrent OPSCC [20]. Similarly, N‐status was associated with OS and DSS in multivariable analysis in our cohort. Only two studies made similar observations, either in a very specific subset of patients (patients with a previous OPSCC and a recurrent or metachronous SCC also located in the oropharynx), or via the surrogate of extranodal extension [18, 19]. The factors influencing DSS in our study in univariable analysis correspond to the intrinsic pathological characteristics of the tumor, namely T‐status, N‐status, perineural invasion, and to a lesser extent lymphovascular embolism. The impact of surgical margins, however, was limited to LRFS and in univariable analysis. Its prognostic significance remains comparatively modest relative to the aforementioned pathological factors. Only two studies reported negative excision margins to be associated with survival (OS for both) in multivariable analysis [10, 19]. While achieving negative margins remains paramount, one should keep in mind that intrinsic tumor characteristics and patient‐related factors often outweigh the impact of margin status on survival.

TABLE 6.

Literature review of studies which included only patients who underwent salvage oropharyngectomies in irradiated neck for recurrent or metachronous OPSCC.

Article	Country	Number of patients	OS	DFS	Risk factors for any survival
Zafereo 2009 [7]	USA	39	3Y: 42% 5Y: 28%	3Y: 26% 5Y: 22%	Older age, persistent disease, N‐status, positive surgical margins (UV)
Nichols 2011 [10]	Canada and USA	29	2Y: 64.5% 5Y: 43.4%	—	Alcohol abuse, positive margins (MV)
Righini 2012 [6]	France	105	—	1Y: 52% 3Y: 31% 5Y: 21%	Tumor size, time since RT (MV) T‐status, N‐status, tumor site, excision margins, persistent disease (UV)
White 2013 [9]	USA	64 (TORS) vs. 64 (conventional)	2Y: 74 vs. 43%	2Y: 74 vs. 43%	Conventional approach (UV)
Culié 2015 [15]	France	34	3Y: 48% 5Y: 28%	3Y: 42% 5Y: 31%	pT‐status, tumor stage, ASA score (MV)
Patel 2016 [16]	USA and Canada	34	2Y: 62% 3Y: 41% 5Y: 25%	3Y: 28% 5Y: 19%	N‐status, margin status, LVE (UV)
Sweeny 2016 [17]	USA	69	2Y: 47% 5Y: 23%	—	T‐status, N‐status, p‐16 status (UV)
Philouze 2017 [11]	France	52	3Y: 43% 5Y: 31%	3Y: 19% 5Y: 19%	Older age, recurrent vs. metachronous disease (MV) T‐status, time since RT (UV)
Heft Neal 2020 [18]	USA—Canada	120	5Y: 31%	5Y: 25%	Time since RT, smoking status (MV) N‐status and reirradiation after complete excision (MV, in subgroup analysis)
D'Andréa 2022 [19]	France	53 (TORS)	2Y: 59%	2Y: 46.1%	ENE, positive margins (MV)
Balaguru 2024 [12]	USA	26	1Y: 82.9% 2Y: 61.6% 3Y: 49.8% 4Y: 49.8% 5Y: 24.9%	1Y: 66.9% 2Y: 56.2% 3Y: 49.9% 4Y: 49.9% 5Y: 25.0%	—
Our study	France	94	1Y: 79.8% 2Y: 59.3% 3Y: 45.9% 4Y: 34.9% 5Y: 31.7%	DSS 2Y: 77.8% 5Y: 58.5%	T‐status, N‐status, recurrent vs. metachronous disease, older age (MV) Excision margins, tumor size, PNI (UV)

Abbreviations: ASA = American Society of Anesthesiologists, ENE = extranodal extension, LRRFS = locoregional recurrence‐free survival, LVE = lymphovascular embolism, MV = multivariable analysis, PNI = perineural invasion, RT = radiotherapy, UV = univariable analysis, Y = year.

In multivariable analysis, we found that metachronous OPSCC had about three times fewer chances of local recurrence or disease‐related death than recurrent OPSCC. Another study found a similar association [11]. Indeed, recurrent cancers are usually considered more aggressive than primary ones. Additionally, initial RT may have been delivered to a different anatomical site. This may have allowed not only for a less complex surgical resection but, more importantly, for a reirradiation of the oropharynx.

The prognostic influence of HPV‐status and smoking in primary OPSCC is well established [21]. Similarly, HPV‐positive and p16‐positive OPSCC exhibit a longer survival after treatment failure or in case of distant metastasis [22, 23]. However, a prognostic difference according to HPV or p16‐status in the setting of salvage oropharyngectomy yet has to be demonstrated [24]. In our series, patients with p16‐positive OPSCC did not fare better than their p16‐negative counterparts. This is consistent with other studies [15, 16, 18, 25] but remains to be further validated as the number of p16‐positive patients was low in our cohort (n = 16) and about half of them had a history of smoking > 20 pack‐years, which may bring their prognosis closer to that of p16‐negative OPSCC. To date, only the study by Sweeny et al. highlighted an increased risk of recurrence, at the limit of significance (p = 0.05), in p16‐negative patients [17]. Surprisingly, a study by Ansarin et al., which included only patients with small recurrent OPSCC, some of them without history of RT, observed a worse locoregional recurrence‐free survival among HPV/p16‐positive patients, in multivariable analysis (HR 2.63). They attributed it to the low number of HPV‐positive patients (n = 9), and their advanced nodal stage [20]. Of note, two studies reported a better prognosis in salvaged HPV‐positive patients, but it was largely biased by the inclusion of other tumor subsites or the comparison to a historical cohort [26, 27]. Overall, this suggests that the prognosis of HPV‐positive salvaged OPSCC is likely not better (or marginally better) than that of their HPV‐negative counterparts.

Lip‐splitting mandibulotomy (also known as mandibular swing) for exposure is a very common approach to OPSCC, especially in the salvage setting as previous RT and/or surgery may have impaired the extent of mouth opening. The rate of such approach in published series can go up to 50% [12]. A recent meta‐analysis suggested that lip‐mandible preservation technique may be a safe and efficient alternative to lip‐splitting mandibulotomy for treating cancers of the oropharynx and oral cavity. It was associated with decreased rates of mandibular osteomyelitis/osteoradionecrosis, fistula, flap infection, alongside a better aesthetic outcome [28]. Our series is in line with this statement, even among the salvage population, having the highest risk of access difficulty. Indeed, we reported fewer than 3% of mandibulotomies, and our oncological outcomes were comparable to that of the literature. However, as most recurrences were local (79.1%) and as positive excision margins have been reported to be independently associated with a worse survival, it is paramount not to underestimate the importance of exposure [10, 18].

Of note, our rate of occult nodal metastasis (5.4%) matched that of a series by Heft Neal (6%) [29]. Performing a prophylactic neck dissection during this last‐resort procedure, which often requires a trans‐cervical approach for free flap reconstruction, seems reasonable irrespective of the low rate of occult nodal disease. The location of nodal metastases could be relatively atypical (including Levels IA, IB, and V), which reminds one of the modifications of lymphatic drainage in the previously treated neck.

Based on retrospective series, salvage oropharyngectomy, whenever feasible, is believed to provide better survival outcomes than reirradiation [7]. However, this needs to be weighed against the posttreatment functional outcomes and quality of life [19, 30]. Advanced reirradiation techniques, such as IMRT, have allowed for more precise targeting of the tumor sites while minimizing radiation to surrounding healthy tissues, thus improving both the toxicity profile and the patient's quality of life [31, 32]. A subgroup analysis from the study by Heft Neal et al. suggested that patients undergoing salvage surgery with negative margins followed by reirradiation exhibit improved 5‐year OS and DSS [18]. Nevertheless, proving the benefit of postoperative reirradiation remains difficult in such heterogeneous retrospective cohorts. Although reirradiation provides significant benefits in a number of cases, it should be used with caution due to the heightened risk of long‐term toxicity, particularly in patients who have already received a high cumulative radiation dose. Therefore, it should be discussed thoroughly with each patient, balancing potential benefits and associated risks.

Among the limitations of this study are its relatively limited number of patients and events. Furthermore, the cohort was inherently heterogeneous, as it mixed patients with various histories of head and neck cancer. On the other hand, the homogeneity of patient management in this single center, the high number of collected variables, and the very limited amount of missing data are strengths of this study, especially as other published series are smaller and/or dated and/or present gaps.

In the modern era of high HPV prevalence, IMRT and mandibulotomy avoidance, the oncological outcomes of salvage oropharyngectomy in irradiated neck for recurrent or metachronous OPSCC remain poor. This last‐resort procedure is associated with high rates of local recurrences and death from various causes, in this relatively morbid population. The main determinants for survival were an advanced T‐status, a positive N‐status, and a recurrent (vs. metachronous) OPSCC. However, patients with p16‐positive OPSCC did not fare better than their p16‐negative counterparts, which is in keeping with the current literature. A robust understanding of these prognostic factors will help to appropriately select patients for this surgery.

Funding

The authors have nothing to report.

Ethics Statement

This study was conducted within a General Data Protection Regulation‐compliant and secure system, in accordance with the French legal framework of MR‐004 set up by the National Commission on Informatics and Liberty. This study received approval by the local research committee and was recorded in the Health Data Hub (reference: 19270385).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.