PET/CT follow up for Head and Neck Squamous Cell Carcinoma The following is a presentation of a prospective protocol, named PET/CT follow up for Head and Neck Squamous Cell Carcinoma (PET Follow), including patients who have completed radiotherapy treatment for squamous cell carcinoma of the head and neck (HNSCC).

The purpose of this study is to investigate the diagnostic performance of 18F-fluorodeoxy-D-glucose (FDG) Positron Emission Tomography/ Computed Tomography (PET/CT) in patients with HNSCC after curative intended treatment.

PET/CT follow up for Head and Neck Squamous Cell Carcinoma The following is a presentation of a prospective protocol, named PET/CT follow up for Head and Neck Squamous Cell Carcinoma (PET Follow), including patients who have completed radiotherapy treatment for squamous cell carcinoma of the head and neck (HNSCC).

Purpose The purpose of this study is to investigate the diagnostic performance of 18F-fluorodeoxy-D-glucose (FDG) Positron Emission Tomography/ Computed Tomography (PET/CT) in patients with HNSCC after curative intended treatment.

Background Head and neck cancer HNSCC is the most frequent malignancy in the head and neck area and arises from the mucosal lining in the oral cavity, naso-, oro-, hypo-pharynx and larynx. HNSCC is a heterogeneous disease in regard to involved anatomical subsite and prognosis. Patients with human papilloma virus (HPV) related tumors in oropharynx have a good prognosis, whereas patients with tobacco and alcohol related tumors have a relatively poor prognosis. Treatment involves surgery and or radiotherapy (RT) with or without concomitant chemotherapy depending on disease stage and anatomical sublocation. Even with optimal treatment approximately 40-50% of the patients experience recurrence after treatment and the majority of recurrences occurs within 2 years after treatment. Early detection of possible recurrence is important to enable possible salvage treatment early. In Denmark patients are followed up, according to DAHANCA guidelines (www.dahanca.dk), with clinical examination 2, 6, 12, 18, 24 months after ended treatment and then yearly until five years after end of treatment. Although clinical examination of HNSCC patients can be difficult, routine follow up with imaging is not recommended in Denmark but may be used on clinical suspicion of relapse. Especially the neck is difficult to assess clinically, and some centers have implemented routine imaging 2-3 months after ended treatment. However, there are no consensus guidelines regarding imaging modality, timing nor how to interpretate the results. This lack of consensus challenges the management of the patients following imaging procedures with uncertain and varying specificity and sensitivity. Figure 1 below depicts the current follow up management in Denmark.

Imaging Both surgical and radiotherapeutic treatment alter the anatomy in HNSCC patients which challenge both clinical examination and conventional anatomical imaging with CT and Magnetic Resonance Imaging (MRI). Imaging with FDG-PET is based on assessment of cellular metabolism and thus less influenced by changes in anatomical structures and could be of value to distinguish between treatment related changes and recurrence. On the other hand, increased FDG can be seen in residual/recurrent tumour as well as in inflammation caused by radiotherapy or recent surgery. Previously, planned neck dissection after RT has been routine in some countries. But in a recent clinical randomized trial (PET-NECK), surveillance with PET/CT proved non-inferior compared to routine planned neck dissection after RT treatment. In the PET-NECK study patients were classified in three categories based on the PET/CT assessment as follows: 1) incomplete response (ICR); 2.) equivocal response (EQR) and 3.) complete response (CR). Both patients with ICR and EQR underwent neck dissection and only patients with CR continued with conventional follow up in the PET-NECK study. Based on the PET-NECK study, several institutions have changed the management of patients from planned neck dissection to surveillance with PET/CT and preserved neck dissection only for patients with ICR or EQR. However, the follow up management has in some institutes changed beyond the PET-NECK study using further surveillance for patients with ICR or EQR, despite lack of direct evidence to validate the safety of this approach. These institutes have gradually adopted a more conservative, watch and wait, approach using PET/CT surveillance for patients with EQR and in some cases patients with ICR on the evaluation PET/CT after treatment. These patients have a repeated PET/CT scan 1-6 months later instead of immediate neck dissection. This change in day to day practice is based on institutional experience and a few retrospective studies reporting, that some patients may be spared from unnecessary neck dissection. The watch and wait approach are driven mainly by two reasons. 1) The high negative predictive value of the evaluation PET/CT scan can identify patients who can be spared of neck dissection; And 2) The lower positive predictive value reported in meta-analyses raise a concern of unnecessary neck dissection in patients whom on the repeated PET/CT are converted from EQR or ICR to CR. Especially in patients with HPV related disease who have a fairly good prognosis and tumors that seems to take longer to respond to RT. Planned neck dissection after RT has never been the standard of care in Denmark and post treatment evaluation with imaging has been preserved for patients with clinical suspicion of recurrence. As such, the PET-NECK study has in principal not changed the routine follow up guidelines for HNSCC patients in Denmark. However, also in Denmark a growing concern for overtreating patients has led to a more conservative approach using PET/CT scans for surveillance and gradually adopting a watch and wait approach without clear consensus recommendations nor evidence supporting this approach.

There are two main concerns with the watch and wait strategy both with several potential issues and pitfalls. The first concern relates to the lack of clinical guidelines on whom, how and when to scan. There is no agreement nor consensus on the timing of the subsequently PET/CT scans, and no well-established standardization on how to interpretate the PET/CT scans. And there is no direct evidence recommending a watch and wait approach, which has led to some ambiguity in different recommendations for follow up management for HNSCC patients. The second concern relates to the uncertainties in the diagnostic performance in PET/CT. There is a high physiological FDG uptake in head and neck area challenging the interpretation, especially after surgery or radiotherapy, and the potential correlation between FDG uptake and vital tumor cells is low and not well understood.

In this study both concerns will be addressed by investigating the diagnostic performance of PET/CT scans in the follow up management of patients treated for HNSCC.

Aim

To achieve the purpose of the study the following specific aim (SA) will be addressed:

-SA1: Investigate the diagnostic performance of PET/CT scan in a post treatment setting and calculate positive and negative predictive values.

Methods This is a prospective two stage single arm study where patients undergoing surveillance after treatment for HNSCC will be offered inclusion in the study to quantify the sensitivity, specificity, positive and negative predictive values of PET/CT 3 months after radiotherapy treatment. Patients who have completed curative intended radiotherapy for HNSCC in the oral cavity, pharynx or larynx and meet the inclusion criteria will be asked to participate in the study.

Study plan Approximately 600 hundred patients were treated for HNSCC in eastern Denmark in 2020 (DAHANCA annual report; available in danish only www.DAHANCA.dk). We therefore expect 300-400 patients would be relevant candidates for this study and 200-300 to be accrued yearly after initiation of the protocol. The enrollment in the study will not delay any treatment in case of recurrence. Figure 1 depicts the current conventional follow up and figure 2 illustrates the follow up for the patients included in the study. Changes from conventional follow up are marked in red.

All included patients will have a PET/CT performed three months after ended treatment. Patients with EQR or ICR will undergo neck dissection.

PET/CT The PET/CT scans will be performed in accordance with EANM guidelines. The interpretation of the PET/CT scans regarding N-site will be performed according to the Deauville scoring system. Deauville relies on visual inspection of the relative difference in tumor metabolism compared to surrounding normal tissue and/or background uptake in the mediastinal blood pool and liver. The Deauville criteria use 5-point scales, where scores 1 and 2 effectively represent a complete metabolic response. The Deauville scoring system is already known and used in lymphoma patients and proved to minimize indeterminate scan results in the study by Zhong et al.

A central review board will be established between the three participating departments of Clinical Physiology and Nuclear Medicine at Rigshospitalet, Herlev and Køge. The PET/CT scans will be assessed by consultants at two of the involved departments and in case of lack of consensus, the third department will be involved to ensure consensus by at least two departments. Beside above the PET/CT scans will be assessed and described as done in clinical routine.

Surgery The neck dissection will be performed according to institutional guideline and the preoperative workup is done as in normal routine as are the postoperative follow up. Suspicious lymph nodes will be marked separately according to anatomical orientation per-operatively. The remaining specimen will be assessed and marked according to specified involved neck levels.

Pathology During the histologic processing all identified lymph nodes will be removed and examined with HE-staining, complemented if needed with other immunohistochemical stains as done in routine diagnostic setting. The per-operatively marked lymph nodes will be sectioned continuously and all slices embedded in paraffin. A pathological positive lymph node is defined by presence of vital tumor cells.

Statistical considerations In the PET NECK study 47 of the 270 (17%) patients in the surveillance group had CR at primary site but ICR or EQR at N-site, and 19 patients (7%) had ICR in both T and N site. In the retrospective studies in literature the numbers of patients with EQR or ICR are higher and reported from 37-40% in patients with HPV related HNSCC to 55% in patients with HPV unrelated HNSCC and 51% in a cohort including patients with both HPV related and unrelated HNSCC. In other words, the numbers of patients with CR, EQR and ICR varies considerably between the different studies.

Stage one Based on the ambiguous results presented above, the first step in this two-stage study is 1) to investigate how many patients will achieve CR, EQR and ICR on PET/CT performed three months after treatment and 2) to identify a group of EQR patients on imaging with low risk of pathologically viable tumor (<5%).

In terms of statistical power assessment, a sample size of 160 achieves 80% power to detect a difference (P1-P0) of 0.0550 using a one-sided exact test with a target significance level of < 0.0500 (The actual significance level achieved by this test is 0.0305). These results assume that the population proportion under the null hypothesis (P0) is 0.0500.

In other words, a sample of 160 patients is required to determine that the patients in that group has less than 10.5% risk of viable tumor cells assuming the underlying true risk is 5%. Assuming 20% attrition or dropout, we should design for 200 patients undergoing NECK dissection. With 200-300 patients included yearly and 20-50% with EQR or ICR based on the literature, we expect 40-100 patients will have a NECK dissection performed yearly. Based on this assumption we expect to complete the first stage within two to three years.

This is a realistic level for training a risk model in the first stage of the study. However, given that the model will be data generated in this first stage, there is a need for a confirmatory stage where the treatment is unchanged, but the risk model is fixed for validation.

The included variables will include patient information such as smoking history and age and clinical examination including ultrasound and clinical assessment of remission at T- and N- site (complete remission or incomplete remission), p16 status, UICC stage, subsite and PET results (CR, EQR and IR). Please see below subheading "Respect for the subjects' physical and mental integrity and privacy".

Annual interim analyses During both stage one and stage two an annual interim analysis will be conducted and presented for the steering committee.

Stage two In the confirmatory stage we will proceed with a corresponding sample and sample size as in stage one for model building and validate the model for predicting a low risk (<5%) of pathological viable tumor cells of stage one. The combined dataset of stage one and stage two will be used to determine the diagnostic properties of imaging (positive predictive value and negative predictive value) using best expert judgement.

Further stages Further stages will be considered adapted to the findings from stage one and two and the most important clinical questions remaining and will be conducted as shared decision making with the patients based on the predictive values established from stage one and two. The future stages will aid in avoiding unnecessary operations in patients with low risk and avoid delaying salvage treatment with extra control scans in patients with high risk. Separate protocols for further stages will be sent to the scientific ethics committee for approval and will not be initiated before approval.

Steering committee A steering committee will be established to assess the annual interim analyses. Beside the already involved investigators, we propose one person from each of the participating institutions and departments.

Inclusion criteria for the project patients

• Informed consent
• Age above 18
• Completed curative treatment for HNSCC of the oral cavity, nasopharynx, oropharynx, hypopharynx or larynx Exclusion criteria
• Patient refusal
• Patients with clinical N0 neck
• Patients who had neck dissections prior to radiotherapy
• Patients who are clinical inoperable for any reason

Investigators Capitol Region Department of Otorhinolaryngology, Head & Neck Surgery and Audiology, Rigshospitalet Signe B Gram, Jacob H Rasmussen, Irene Wessel and Christian von Buchwald. Department of Oncology, Section of Radiotherapy, Rigshospitalet Ivan R Vogelius and Jeppe Friborg Department of Oncology, Herlev Hospital Elo Verner Andersen Department of Clinical Physiology and Nuclear Medicine, Rigshospitalet Barbara M Fischer and Danijela Dejanovic Department of Clinical Physiology and Nuclear Medicine, Herlev Charlotte Birk Christensen Department of Pathology, Rigshospitalet Giedrius Lelkaitis Region Zealand Department of Otorhinolaryngology, Head & Neck Surgery, Køge Hospital Gitte Hvilsom Department of Oncology, Næstved Hospital Mohammad Farhadi Department of Clinical Physiology and Nuclear Medicine, Køge Hospital Oriol Puig Calvo Dissemination of results The results whether positive, negative or inconclusive will be published in international peer-reviewed journals and submitted to national and international conferences.

Inclusion Criteria:

• Informed consent
• Age above 18
• Completed curative treatment for HNSCC of the oral cavity, nasopharynx, oropharynx, hypopharynx or larynx

Exclusion Criteria:

• Patient refusal
• Patients with clinical N0 neck
• Patients who had neck dissections prior to radiotherapy
• Patients who are clinical inoperable for any reason