Introduction
Unilateral radiotherapy for lateralised oropharyngeal squamous cell carcinoma (SCC) can reduce morbidity from treatment compared with bilateral cervical irradiation. Unilateral radiotherapy reduces xerostomia, dysphagia, dependence on gastrostomy feeding, and other functional impairments.Reference Jellema, Slotman, Doornaert, Leemans and Langendijk1, Reference Al-Mamgani, van Rooij, Fransen and Levendag2 For selected patients with T1–T2, N0–N1 (early) tonsillar SCCs, unilateral irradiation achieves comparable oncological outcomes and is an established de-escalation strategy.Reference Al-Mamgani, van Werkhoven, Navran, Karakullukcu, Hamming-Vrieze and Machiels3–Reference Taku, Chronowski, Brandon Gunn, Morrison, Gross and Moreno6
A barrier to unilateral treatment is uncertainty about the presence of synchronous contralateral primary oropharyngeal malignancies. The prevalence of synchronous contralateral tonsillar malignancies is reported at between 1 and 8 per cent, with human papillomavirus (HPV)-positive tumours demonstrating lower rates (approximately 3 per cent) than HPV-negative disease (up to 8 per cent).Reference Charlton, Mughal, Sharin, Sahota, Mansuri and Mair7 This uncertainty underpins the practice of routine contralateral tonsillectomy used at some centres, providing histological confirmation of a clear contralateral tonsil prior to radiotherapy planning.
However, contralateral tonsillectomy carries procedural morbidity, including pain for two weeks and post-tonsillectomy haemorrhage, and may delay definitive treatment, which in head and neck cancers is associated with inferior outcomes, with most cancers progressing in a period of four weeks.Reference Jensen, Nellemann and Overgaard8
The present study was conducted to determine the rate of synchronous contralateral tonsillar malignancies among patients with oropharyngeal SCC, in order to evaluate whether the diagnostic yield of routine contralateral tonsillectomy justifies its associated morbidity and the potential for delay in definitive treatment.
Materials and methods
A retrospective review was performed of all patients treated for oropharyngeal SCC at a tertiary centre between 1 January 2022 and 31 December 2024, where radiotherapy or transoral robotic partial pharyngectomy are offered for early oropharyngeal malignancies. Patients were identified through departmental operative records and the Somerset Cancer Registry. Each patient was assigned a unique identifier, and clinical data were collected from the electronic patient records on a standardised proforma before transfer to an anonymised database for analysis. The study was registered with the clinical audit department at Imperial College Healthcare NHS Trust (reference ENT_072).
Inclusion criteria were patients with histologically confirmed oropharyngeal SCC who underwent definitive radiotherapy or surgery with curative intent. Exclusion criteria were treatment with palliative intent, death before treatment and treatment at other institutions.
All patients underwent comprehensive staging by clinical examination, magnetic resonance imaging (MRI) and/or computed tomography (CT), and positron emission tomography (PET)-CT, according to multidisciplinary team consensus. The contralateral tonsil was considered normal on cross-sectional imaging when there was no asymmetric enlargement, abnormal enhancement or restricted diffusion, and PET-CT demonstrated no abnormal avidity to fluorodeoxyglucose.
Decision-making between unilateral and bilateral cervical radiotherapy was undertaken by the radiotherapy department. Patients with early lateralised tonsillar tumours (T1–T2) and some with T3 disease, and early ipsilateral nodal disease (N0–N1), were considered potential candidates for unilateral radiotherapy, based on the consensus criteria of the Royal College of Radiologists, UK.9 For these patients, contralateral tonsillectomy was performed to look for a synchronous contralateral malignancy prior to finalising the radiotherapy plan. Patients were scheduled for bilateral radiotherapy without contralateral tonsillectomy when they had tumours of the midline base of tongue, extensive primary tumours crossing the midline, or bilateral or contralateral regional lymph node involvement.
Contralateral tonsillectomy specimens were processed using standard formalin fixation and paraffin embedding with serial sectioning at 3–4 mm intervals. Sections were examined by specialist head and neck histopathologists. Immunohistochemistry for p16 was performed using the CINtec p16 (E6H4) antibody (Ventana Medical Systems (now Roche), Oro Valley, Arizona, USA) with the Bond Refine Detection Kit (Leica Biosystems, Newcastle upon Tyne, UK). Diffuse, strong nuclear and cytoplasmic staining in 70 per cent or more of tumour cells was considered p16-positive. Human papillomavirus status was initially determined by in situ hybridisation using a high-risk HPV probe cocktail (Ventana Medical Systems (Roche)), being replaced by HPV genotyping at Health Services Laboratories, London, UK, from November 2023.
The primary outcome was the rate of histologically confirmed SCC in the contralateral tonsillectomy specimen. Secondary outcomes included complication rates from contralateral tonsillectomy and any treatment delays associated with staged diagnostic procedures. Procedures were classified as same-sitting when the diagnostic biopsy and contralateral tonsillectomy were performed under a single general anaesthetic, and as staged when they were carried out on different dates.
Descriptive statistics were used to summarise patient demographics and tumour characteristics. Continuous variables are reported by their median and range. 95 per cent confidence intervals were calculated for proportions. The Mann–Whitney U test was used to compare time intervals between groups. Statistical analysis was performed using R (R Foundation for Statistical Computing, Vienna, Austria). A two-sided p-value of less than 0.05 was considered statistically significant.
Results and analysis
Eighty-one patients with oropharyngeal SCCs met the inclusion criteria during the period of study. Demographic and tumour characteristics are summarised in Table 1. Most patients were male (59 out of 81, 73 per cent) with a median age of 64 years (range, 43–82 years). Primary tumour sites were the palatine tonsil (n = 42, 52 per cent), base of tongue (n = 30, 37 per cent) and multiple oropharyngeal subsites (n = 9, 11 per cent). Immunostaining for p16 was positive in 63 of 78 patients (81 per cent) with test data. Human papillomavirus in situ hybridisation and/or HPV genotyping were positive in 44 of 54 patients tested (81 per cent), Overall, 65 of 80 patients (81 per cent) with at least one test performed were HPV-associated.
Patient and tumour characteristics
Table 1 Long description
Patient demographics and tumor characteristics are summarized for a cohort of 81 patients. Most patients were male, with 59 men making up 73 percent of the group. The median age was 64 years, with ages ranging from 43 to 82 years. Primary tumor site was most often the tonsil at 42 patients, or 52 percent, followed by base of tongue at 30 patients, or 37 percent; multiple subsites accounted for 9 patients, or 11 percent. HPV association was reported for 80 patients, and 65 of those 80 were HPV-associated, corresponding to 81 percent. Because HPV status was not available for one patient, the HPV percentage reflects those with reported results rather than the full cohort.
HPV = human papillomavirus
There was suspicion of a separate and contralateral tonsillar lesion in two patients on contrast-enhanced MRI, and in four on PET-CT; one patient was positive on both imaging modalities. Of these five patients with abnormal imaging, four underwent contralateral tonsillectomy, and none had a malignancy on histological examination. In total, 42 patients underwent contralateral tonsillectomy. The remaining 39 were scheduled for bilateral radiotherapy based on tumour extent or nodal distribution, as determined by the multidisciplinary team (Table 2, Figure 1).
Comparison of patient and tumour characteristics by contralateral tonsillectomy status
Table 2 Long description
Patient demographics and tumour features are compared between those who had contralateral tonsillectomy (42 patients) and those who did not (39 patients). Median age was 60 years with tonsillectomy versus 64 years without, and both groups were mostly male (74 percent and 72 percent). Primary site differed: tonsil was most common with tonsillectomy (74 percent), while base of tongue was most common without tonsillectomy (56 percent). Tumour stage was generally earlier in the tonsillectomy group (T1 29 percent, T2 55 percent) and more advanced in the no-tonsillectomy group, including T4 in 38 percent versus none with tonsillectomy. Nodal stage also differed, with N1 more frequent with tonsillectomy (55 percent vs 23 percent) and higher nodal categories appearing more often without tonsillectomy (N2c 15 percent and N3 10 percent vs none). HPV-associated disease was common in both groups but higher with tonsillectomy (86 percent vs 74 percent), and smoking status was broadly similar, though current smoking was higher without tonsillectomy (38 percent vs 24 percent). Tumours approaching or crossing the midline were much more frequent without tonsillectomy (59 percent vs 12 percent). Percentages may not sum to 100 due to rounding and small counts in some categories.
HPV = human papillomavirus
Patient flow diagram. SCC = squamous cell carcinoma.
Synchronous contralateral tonsillar SCC was identified in only one patient (2.4 per cent; 95 per cent confidence interval = 0.1-12.6 per cent), and this case was occult to both MRI and PET-CT. At the multidisciplinary team meeting where contralateral tonsillectomy was planned, the right tonsillar primary was staged clinically as T1N1,Reference Brierley, Gospodarowicz and Wittekind10 with a single ipsilateral lymph node and no radiological evidence of extranodal extension, and the patient met the criteria for unilateral radiotherapy at the time. Histopathological examination of the contralateral tonsillectomy and neck dissection tissue subsequently demonstrated a synchronous HPV-positive left tonsillar SCC (pathologically confirmed T1) and extranodal extension within the ipsilateral nodal disease. The extranodal extension confirmed on histopathological examination necessitated bilateral oropharyngeal and cervical chemoradiation, in line with departmental protocol. The patient was treated by transoral robotic partial pharyngectomy of the right tonsillar tumour and right neck dissection, followed by 65.1 Gy in 30 fractions and carboplatin to the left synchronous tonsillar primary and neck bilaterally, and remained cancer-free at 13 months. Because the extranodal extension alone necessitated bilateral chemoradiation, the synchronous contralateral primary did not alter the treatment this patient received.
Median follow-up duration for the cohort was 16 months (range, 2–37 months), and no metachronous contralateral tonsillar malignancies developed during the period of study.
Post-tonsillectomy haemorrhage from the contralateral tonsillectomy site requiring hospital re-admission occurred in 1 of 42 patients (2.4 per cent). A return to the operating theatre was not necessary. No other surgical complications were recorded from contralateral tonsillectomy.
Of 42 patients, 31 (74 per cent) underwent contralateral tonsillectomy at the same sitting as panendoscopy and biopsy of the primary tumour. The remaining 11 patients (26 per cent) had staged procedures, with a median interval of 21 days (range, 6–42 days) between tumour biopsy and tonsillectomy: 4 had their biopsies performed at the out-patient clinic followed by contralateral tonsillectomy under general anaesthesia, and 7 required 2 general anaesthetics. The median time from the initial biopsy to definitive treatment was 41 days in the same-sitting group and 34 days in the staged group (Mann–Whitney U test, p = 0.77). The staged group required an additional procedure (and a second general anaesthetic in 7 of the 11 patients), introducing a median 21-day interval during which the treatment plan remained unconfirmed.
Discussion
Synchronous contralateral tonsillar SCC occurred in only 1 of 42 patients (2.4 per cent) undergoing contralateral tonsillectomy for apparently unilateral oropharyngeal SCC. Its detection did not change the laterality of treatment, as extranodal extension identified in the neck dissection on histopathological examination necessitated bilateral oropharyngeal and cervical chemoradiation regardless. The results of contralateral tonsillectomy did not alter management in the cohort.
This synchronous contralateral malignancy rate is consistent with the lower end of that reported in published literature (1–8 per cent) and with a recent meta-analysis demonstrating reduced rates in HPV-positive populations.Reference Charlton, Mughal, Sharin, Sahota, Mansuri and Mair7 Eighty-one per cent of our cohort was HPV-associated, reflecting contemporary oropharyngeal cancer epidemiology, in which European prevalence of HPV in oropharyngeal SCC has risen from about 35 per cent before 2000 to over 70 per cent in more recent series.Reference Mehanna, Beech, Nicholson, El-Hariry, McConkey and Paleri11 A systematic review reported that synchronous contralateral malignancies were often missed at clinical and imaging evaluations preoperatively.Reference Charlton, Mughal, Sharin, Sahota, Mansuri and Mair7 Our single positive case fits this pattern, being occult on both MRI and PET-CT. However, given an absolute prevalence of approximately 3 per cent in HPV-positive populations, few patients stand to benefit from routine contralateral tonsillectomy. The same review recommended a tailored approach, with contralateral tonsillectomy generally warranted for unknown primary disease;Reference Charlton, Mughal, Sharin, Sahota, Mansuri and Mair7 the present cohort comprised patients with known, lateralised primary cancers.
Unilateral intensity-modulated radiotherapy reduces xerostomia and dysphagia compared with bilateral treatment by sparing the contralateral parotid gland and pharyngeal constrictor muscles.Reference Jellema, Slotman, Doornaert, Leemans and Langendijk1, Reference Al-Mamgani, van Rooij, Fransen and Levendag2 Pooled contralateral regional failure rates following unilateral irradiation are only 2–3 per cent in series from the HPV era,Reference Al-Mamgani, van Werkhoven, Navran, Karakullukcu, Hamming-Vrieze and Machiels3, Reference Razavian, D’Agostino, Steber, Helis and Hughes4 with five-year overall survival greater than 90 per cent.Reference Rackley, Namelo, Palaniappan, Cole, Owens and Evans5, Reference Taku, Chronowski, Brandon Gunn, Morrison, Gross and Moreno6 The consensus statements of the Royal College of Radiologists and American Radium Society support unilateral radiotherapy for well-lateralised T1–T2, N0–N1 tonsillar SCCs located more than 10 mm from the midline.9, Reference Tsai, Galloway, Margalit, Bakst, Beadle and Beitler12 Contralateral tonsillectomy may be performed in this setting to provide histological confirmation before unilateral radiotherapy volumes are finalised.
Routine contralateral tonsillectomy carries morbidity and may delay definitive treatment. The 2.4 per cent re-admission rate for haemorrhage in this series is lower than the 16 per cent reported in ‘NATTINA’, the largest contemporary UK randomised trial of adult tonsillectomy.Reference Wilson, O’Hara, Fouweather, Homer, Stocken and Vale13 Twenty-six per cent of patients who underwent tumour biopsy and contralateral tonsillectomy had these as staged procedures, with a median 21-day interval (range, 6–42 days) between biopsy and contralateral tonsillectomy. Head and neck cancers progress while awaiting radiotherapy,Reference Jensen, Nellemann and Overgaard8 and each four-week delay to radical radiotherapy is associated with about a 9 per cent increase in mortality rate (hazard ratio = 1.09, 95 per cent confidence interval = 1.05–1.14), and 6 per cent for surgery., Reference Hanna, King, Thibodeau, Jalink, Paulin and Harvey-Jones14 The median time to definitive treatment did not differ significantly between same-sitting and staged groups in this cohort (41 vs 34 days, p = 0.77), which may reflect selection bias.
No synchronous contralateral malignancies were identified in the 11 patients who underwent a staged biopsy and contralateral tonsillectomy. The single positive case in the series underwent tonsillectomy at the same sitting as panendoscopy and biopsy. No patients were upstaged between their biopsy and contralateral tonsillectomy, and the time from biopsy to definitive treatment was not longer in the staged group than in the same-sitting group. The interval between staged procedures, therefore, did not delay definitive treatment or yield a diagnosis that altered management in the present cohort. Seven of the 11 staged patients required a second general anaesthetic before commencing definitive treatment.
This study has limitations inherent to retrospective design. Contralateral tonsillectomy was performed preferentially in smaller, more lateralised tumours (Table 2), so the two groups are not directly comparable. With 42 procedures, the upper 95 per cent confidence bound for the synchronous malignancy rate was 12.6 per cent, so a rate as high as that reported in HPV-negative disease cannot be excluded. All candidates for unilateral radiotherapy underwent contralateral tonsillectomy, so the present study cannot assess the outcomes of unilateral radiotherapy without contralateral tonsillectomy. However, among 295 patients with p16-positive tonsillar squamous cell carcinoma treated by unilateral transoral robotic surgery, in whom the contralateral tonsil was excluded from adjuvant radiation and not excised, only one (0.3 per cent) developed a contralateral tonsillar primary over a median follow-up of 48 months, which was treated with re-irradiation.Reference Parhar, Shimunov, Brody, Cannady, Newman, O’Malley, Chalian, Rassekh, Weinstein and Rajasekaran15 The median follow-up duration of 16 months in the present study does not capture long-term oncological outcome such as contralateral regional relapse. Toxicity and functional data were not collected prospectively.
• Unilateral radiotherapy for lateralised oropharyngeal squamous cell carcinoma reduces toxicity compared with bilateral treatment
• Contralateral tonsillectomy is sometimes used to exclude synchronous malignancies and permit unilateral irradiation of the primary cancer and ipsilateral neck
• Reported prevalence of synchronous contralateral tonsillar malignancy is 1–8 per cent, with lower rates in human papillomavirus (HPV)-positive disease
• In a predominantly HPV-positive cohort, synchronous contralateral tonsillar malignancy occurred in only 2.4 per cent of patients, without altering management
• Routine contralateral tonsillectomy carries morbidity and may delay treatment if performed after biopsy
• Findings support selective omission of routine contralateral tonsillectomy in appropriately staged patients with early nodal disease
Conclusion
Routine contralateral tonsillectomy resulted in no findings that altered management. The single occult tonsillar SCC required bilateral chemoradiation regardless, owing to extranodal extension identified on histopathological examination. While contralateral tonsillectomy provides some information that may contribute to radiotherapy planning, the low diagnostic yield, procedural morbidity and potential to delay treatment support omitting contralateral tonsillectomy as part of broader de-escalation strategies for HPV-associated oropharyngeal SCC.
Competing interests
None declared.
Ethical standards
The authors confirm the study did not require ethics committee approval and was conducted in accordance with the Helsinki Declaration of 1975, as revised in 2024.
Open access
