Effectiveness of bilateral tongue base mucosectomy by transoral robotic surgery or transoral laser microsurgery in combination with tonsillectomy in identifying head and neck primary cancer of unknown primary: a randomized phase 2 protocol (RoboCUP trial)

Abstract

Background

In cases of prevalent lymphadenopathy in the head and neck cancer region, identifying the primary tumor site allows for more precise radiotherapy targeting. This improves treatment by reducing the volumes of mucosal irradiation and potentially lowering morbidity. An extensive diagnostic workup, including FDG PET-CT imaging and bilateral tonsillectomy, has been shown to identify the primary cancer in 60% of cases. Mucosectomy of the tongue base holds promise for detecting additional primary sites. When performed using minimally invasive endoscopic techniques such as transoral robotic surgery (TORS) or transoral laser microsurgery (TLM), mucosectomy results in minimal morbidity. However, the effectiveness of tongue base mucosectomy in detecting primary tumors has yet to be evaluated in a randomized trial involving patients with lymphadenopathy of unknown primary.

Methods

The RoboCUP trial is a multicentre, open-label, randomized, non-comparative phase 2 trial aiming to evaluate the benefit of bilateral TORS or TLM-assisted tongue base mucosectomy in association to tonsillectomy in the assessment of prevalent cervical lymphadenopathy with a negative exhaustive diagnostic workup. The main endpoint is the proportion of patients with detection of a primary cancer. Surgery will consist in tongue base mucosectomy plus tonsillectomy in the experimental arm, and the standard of care, i.e. tonsillectomy alone in the control arm. Patients will then be treated by intensity modulated radiotherapy, possibly with chemotherapy as radiosensitizer, per current guidelines. Using a single-stage Fleming design, 36 patients will be enrolled in the experimental arm, and 36 patients in the control arm.

Discussion

This trial aims to improve the diagnostic performances, i.e. detection of primary tumor, in patients with head and neck carcinoma of unknown primary. It is expected that the subsequent therapeutic changes could enhance radiotherapy accuracy, and could improve the prognosis, toxicity profiles and quality-of-life of patients.

Trial registration

NCT04767048, registered February, 19, 2021.

Background

Head and neck carcinoma of unknown primary (CUP) is defined as cervical lymph node metastasis without a primary tumor. It represents 3 to 5% of head and neck cancers [1]. Therapeutic management includes cervical node dissection followed by radiotherapy of mucosal areas at risk. Historically, radiotherapy was delivered to large mucosal volumes from the nasopharynx to the larynx mucosal and cervical region. The benefit in terms of local control of comprehensive mucosal irradiation is however controversial owing to the significant sequelae, including severe dysphagia and xerostomia [2]. Due to excess morbidity with such large irradiated volumes, radiotherapy practice has evolved toward attempts to reduce volumes based on areas of lymphatic involvement. However, due to lack of evidence for node-level guided mucosal irradiation, strategies vary widely across teams. Better identification of the primary tumor site is one of the priorities in this field as it would allow for irradiation limited to the mucosa of the primary site. This is expected to reduce the morbidity of the multimodal treatment. Initial diagnostic procedures therefore should play a pivotal role. Efforts remain necessary to achieve a clear consensus [3].

Fluorodeoxyglucose positron emission tomography (FDG-PET)/CT-scan imaging has a detection rate of primary cancer of 29 to 40% [4]. Ipsi or bilateral tonsillectomy for diagnostic purposes yields primary detection rates in up to 60% of cases [5]. The tongue base is another likely primary site but bilateral mucosectomy (to account for bilateral lymphatic drainage) of the tongue base could not be performed in a minimally invasive way before the generalization of endoscopic surgical techniques (laser and transoral robotic surgery (TORS)) [6]. Recent retrospective or non-randomized prospective American studies, of small sample size, suggest increased primary tumor detectability with systematic transoral tongue base mucosectomy in addition to ipsilateral tonsillectomy and FDG PET/CT, yielding primary tumor detection in 80% of cases [7–8]. One possible drawback is the occurrence of postoperative hemorrhage in 5% of cases, requiring a usually-efficient hemostatic procedure in the operating room in 3% of cases. Finally, there seems to be room for transoral tongue base mucosectomy (TORS or laser assisted) in addition to the tonsillectomy and PET in the diagnostic workup of cervical CUP but randomized studies are warranted to get sufficient level of evidence.

We thus propose a randomized phase II study to assess the detection rate of bilateral TORS or laser-assisted tongue base mucosectomy in combination with tonsillectomy in patients with cervical lymphadenopathies of unknown primary after clinical and FDG PET/CT assessment.

Methods/design

Trial schedule and objectives

The RoboCUP study is a multicenter, open-label, randomized phase II trial aiming to evaluate the interest of performing tongue base mucosectomy in addition to a tonsillectomy for the detection of a primary cancer.

Randomization 1:1 will be stratified by study centre through the minimization method, to allocate patients into one of the following arms:

• Experimental arm: TORS or laser-assisted tongue base mucosectomy in addition to unilateral or bilateral tonsillectomy.

• Control arm: unilateral or bilateral tonsillectomy alone.

The flow-chart of the study is displayed in Fig. 1.

The RoboCUP protocol and this manuscript have been written following standard protocol items, namely recommendations for interventional trials (SPIRIT).

Fig. 1

Flow chart of the randomized phase 2 RoboCUP study

Primary objective

The main objective is to determine the proportion of primary cancer detected through histological analysis of tissue samples from bilateral TORS or laser-assisted tongue base mucosectomy in addition to ipsilateral tonsillectomy in patients with cervical CUP after an extensive imaging workup showing no primary cancer.

Secondary objectives

Secondary objectives are:

• To evaluate the failure rate, defined by the absence of detection of a primary cancer, in patients undergoing TORS or laser-assisted tongue base mucosectomy plus tonsillectomy (experimental arm); the standard arm being without mucosectomy.

• To evaluate, in both arms:

• Per and post-surgical morbidity: hemorrhages, pain, dysphagia.

• Multi modal treatment morbidity at 6 months: hemorrhages, pain, dysphagia.

• Hospitalization duration.

• Time until the initiation of radiotherapy (surgical complications, healing), dose and volume of radiotherapy.

• Type of adjuvant chemotherapy, if recommended.

• Quality of life: EORTC QLQ-C30 and QLQ H&N43.

• Progression-free survival and overall survival.

• The diagnostic value of FDG PET / CT imaging.

• The prevalence of HPV-positive lesion, determined by strong staining of p16 protein in immunohistochemistry, and confirmed by the presence of HPV virus within the tumor.

• Quality of radiotherapy and differences in radiotherapy volumes in the two arms.

Eligibility criteria

Patients have to fulfill the following main inclusion criteria

• Patient with at least one metastatic cervical lymphadenopathy of squamous cell carcinoma either by fine-needle aspiration or adenectomy, a FDG PET-CT, a contrast-enhanced cervico-thoracic CT scan and an ENT work-up (nasofibroscopy) not finding any primary tumor.

• Performance Status < 2.

• Patient aged 18 or over.

• Patient affiliated with the French social security system.

• Informed consent signed.

Main criteria to exclude subjects from the trial are

• History of head and neck or skin face squamous cell carcinoma.

• History of cervico-facial radiotherapy.

• Primary tumor discovered during panendoscopy.

• Difficult airway management with little accessibility to the tongue base, hindering mucosectomy.

• Hemostatic disorders contra-indicating tonsillectomy and mucosectomy.

• Distant metastases.

• Inoperable / unresectable cervical node.

• Patient with a contraindication to radiotherapy.

• Pregnant or breastfeeding woman.

• Women of childbearing potential without effective contraception.

• Patient under guardianship or unable to give informed consent.

• Patient unable to undergo the trial follow-up for geographic, social or psychological reasons.

Surgical procedure

Study participation will be proposed during the surgery consultation to patients who meet the eligibility criteria. Surgical procedure will be performed within 4 weeks after randomization.

Tonsillectomy

It will be performed for patients of each arm (experimental and control), by doing an extra-capsular ablation of the lymphoid tissue of the palatal tonsil after incision of the anterior pillar. The main side effects or potential risks are: hemorrhage (1%), secondary post-tonsillectomy bleeding, postoperative pain, swallowing disorders, infection and healing disorders, which are expected to be transient.

Tongue base mucosectomy

This procedure will be performed only in patients assigned to the experimental arm. Tongue base mucosectomy involves the removal of lymphoid tissue from the base of the tongue while preserving the underlying musculature. Hemostasis is achieved gradually to minimize the risk of hemorrhage. The area is left uncovered, with expected healing of the excisional site within approximately ten days. Performing this procedure via transoral robotic surgery provides enhanced exposure of the tongue base and surrounding valleys, allowing for improved surgical and oncological control. Alternatively, this procedure can be performed using transoral laser surgery. No significant differences have been demonstrated between these two techniques, and the choice primarily depends on the accessibility of the equipment. The main contraindications include inadequate exposure and uncontrolled hemostasis disorders. Potential complications or risks include hemorrhage (5%), secondary post-tonsillectomy bleeding, postoperative pain, swallowing difficulties, infection, wound healing issues, and reduced lingual mobility [8].

Additional treatments

Intensity Modulated RadioTherapy (IMRT) treatment will be initiated within 6 weeks after surgery. The radiotherapy modalities (dose, duration, etc.) will be as the discretion of the investigators depending on the results of the surgery. Radiotherapy reports and Digital Imaging and Communications in Medicine (DICOM) Radiotherapy (plan, images, structure and dose) will be provided to allow independent analysis of volumes (lymph node areas, head and neck mucous membranes) and radiation doses by two radiation oncologists blind to the treatment arm and to each other before collegial supervision. Chemotherapy will be indicated to potentiate radiotherapy, according to usual standards.

Histological analysis

The analysis of the excised tissues will be performed during definitive anatomo-pathology, without frozen section analysis, according to precise specifications including at least the histological mapping of each of clearly identified samples (tonsillectomy, right and left tongue base). If a primary cancer is detected, the report will specify histological type, tumor size, keratinizing morphology, tumor differentiation, excision’s margins, presence of vascular emboli or perineural invasion and HPV status according to immunohistochemistry staining of p16 protein and confirmed by the search for HPV viral DNA (centralized analysis at the end of the study).

The patient will be excluded if a primary lesion is found outside the tonsil region or tongue base during surgery.

Pathological analysis

The first step of the macroscopic examination (before fixation) will consist in measuring each fragment in its three dimensions (anteroposterior, width and thickness) and describing the appearance of the mucosa. After formalin fixation, the second stage of the macroscopic examination will include:

• Identification of excision margins.

• Inclusion of each hemi-mucosectomy (0.5 cm thick section) mapped along the longest axis, according to serial and numbered sections,

• A formalin-fixed paraffin-embedded (FFPE) section on each paraffin block.

• If there is no identifiable lesion, pathological examination will be completed with 3 staggered cuts on each block (40 to 80 microns apart).

The histological report will describe:

• Presence / absence of a neoplastic lesion.

• If any lesion is identified:

  • Histological type ± in situ component.

  • p16 status.

  • EBV status (only if completed, not mandatory).

  • Dimensions.

  • Infiltration depth (in mm).

  • Uncertainties based on specimen quality such as presence of artifacts, assessable margins between section and lesion, potentially making the sample unassessable.

  • Quality of excision: R0 / R1.

  • pTNM stage.

  • Presence of perineural invasion or vascular emboli.

If required, an immunohistochemical analysis will complete the morphological analysis (markers according to the histological type of the tumor, search for micro-infiltration of the chorion by cytokeratin AE1/3, etc.).

In the presence of invasive squamous cell carcinoma, the immunohistochemical analysis will systematically include p16 protein immunostaining. At the end of the study, a tumor block will be sent to the promoter’s centre to perform centralized analyzes (search for HPV viral DNA) if a malignant lesion is detected.

Statistical methods

Sample size

• Experimental arm: TORS tongue base mucosectomy in addition to a tonsillectomy for the identification of a primary tumor site.

Primary tumor detectability after ipsi or bilateral tonsillectomy is set at 60% [5, 7]. Performing a mucosectomy of the tongue base in addition to tonsillectomy will be considered as useless if the detection rate of primary cancer is lower than 60% (H0), or relevant if it allows to detect primary cancer in at least 80% of cases (H1). According to a single-stage Fleming design, alpha risk level of 5% and power of 80%, 33 assessable patients are required in the experimental arm, including at least 25 patients in whom a primary cancer will be detected to assess the interest of performing mucosectomy before tonsillectomy. Inclusion of 36 patients is planned to anticipate inclusion of non-assessable patients (around 10%).

• Control arm: tonsillectomy alone.

To avoid selection bias and to assess primary cancer detection rate in a population of patients operated for a tonsillectomy, 36 patients will be included in the control arm (expecting at least 33 assessable patients).

Analysis

Analyses will be conducted on an intent-to-treat basis. The proportion of detected primary tumors will be estimated separately in both arms, with a 95% confidence interval. Excepted for clinical characteristics, no direct comparison between arms is expected due to small sample size and lack of power (as in randomized phases 2 in general). Data analysis will be based on the use of non-parametric tests that will identify variables associated with detection of primary tumor. Survival endpoints will be estimated by Kaplan Meier method, with patients lost to follow-up censored on the date of their latest news.

Study assessments

A total of five assessment periods, depicted in Table 1, are planned during this study:

• Baseline assessment, within the 4 weeks before randomization.

• Early post-surgical assessment, between 1 and 3 days after surgery.

• Intermediate post-surgical assessment, 1 month after surgery.

• Radiotherapy period assessment, at the first and the last days of treatment.

• Post-radiotherapy assessment, at 3 months and 6 months after the end of treatment.

Table 1 Overview of study assessments of the robocup trial

Clinical assessment (ECOG performance status, BMI,…), pain assessment (visual analogue scale), dysphagia’s effects on the quality of life (MD Anderson Dysphagia Inventory Head & Neck questionnaire, MDADI-HN) will be performed at each time point. Clinical assessment will be completed by the collection of post-surgical complications at early and intermediate post-surgical assessment periods, and by the collection of toxicities at radiotherapy and post-radiotherapy assessment periods. Use of enteral nutrition (nasogastric tube) will be assessed at each assessment. Tumor evaluation, through contrast enhanced CT and FDG PET-CT, will be performed at baseline and three months after the end of radiotherapy. Nasofibroscopy will be also performed at baseline and after radiotherapy. Quality-of-life, through EORTC QLQ-C30 and the specific HN43 module for head and neck cancer patients, will be evaluated at baseline, early after surgery, during radiotherapy and post-radiotherapy.

Patients will be followed up to 6 months after the end of radiotherapy, even if progression is observed after surgery, unless the patient prematurely withdraws from the study. Patient’s participation will be discontinued at any time under the following circumstances:

• Intraoperatively discovery of a primary cancer outside the base of the tongue or of the oropharynx.

• Disease progression in the lymph node making the patient inoperable.

• Patient’s or investigator’s decision.

• Intercurrent illness or other reason that requires discontinuation of study treatment.

• Patient lost to follow-up.

Discussion

Improving the detection of a primary tumor in the oropharynx, as expected from a tongue base mucosectomy as proposed in this randomized study, is meant to avoid probabilistic mucosal irradiation, as primary agnostic irradiation would result in substantial mucosal volumes. Detection of an oropharyngeal primary would help to limit primary target volumes to the oropharyngeal mucosa instead of the nasopharyngeal, oral cavity, oropharyngeal, laryngeal and hypopharyngeal mucosae. As these mucosae are involved in many critical functions, as a result of reduced mucosal irradiation, it may be expected to preserve the patient from severe cutaneous toxicity, dysphagia (including a risk of prolonged parenteral nutrition), xerostomia in the early and long terms [9]. All of these elements can impact patient’s quality of life, as well as some societal and professional aspects of their life.

Systematic transoral tongue base mucosectomy in addition to ipsilateral tonsillectomy and PET-CT scan seems to improve primary cancer detection up 80%. However, the literature is mostly limited to retrospective studies in small groups of patients. Moreover, these studies are performed on north American & Scandinavian countries which a large incidence of HPV-positive oropharyngeal cancer (up to 90% of oropharyngeal squamous cell carcinoma) [8]. If transoral tongue base mucosectomy in these population with high HPV prevalence seems to be interesting, it has to be confirmed on larger prospective studies and on different populations with less HPV-induced oropharyngeal cancer [10, 11].

To include this diagnostic procedure in the global management of head and neck unknown primary itsmorbidity should be limited. We have thus planned to investigate the impact of tongue base mucosectomy and tonsillectomy comparatively to tonsillectomy alone in terms of pain, swallowing, bleeding complications and procedure-related side effects.

Finally, the post-operative treatment, especially radiotherapy dose and volume, remains debated [3]. We will prospectively collect treatment information between the participating centers to analyze the differences in management and definitions in radiation volume in order to correlate them with toxicities, thereby measuring the clinical impact of tongue base mucosectomy on personalized multidisciplinary treatment.