A Nordic survey on the management of head and neck CUP

Abstract

Conclusion: The management of Head and Neck Cancer of Unknown Primary (HNCUP) patients varies both between centres within and also between the Nordic countries. This study contributes to a continuing discussion of how to improve the accuracy of diagnosis and quality of treatment of HNCUP patients.

Objectives: The initiative for this study was based on the lack of common guidelines for diagnostic procedures and for treatment of HNCUP patients in the Nordic countries constituting a region having a rather homogeneous population.

Method: A structured questionnaire was sent to all university hospitals in the five Nordic countries.

Results: Four of the five Nordic countries use either national guidelines or specific protocols when handling HNCUP. The main diagnostic tools are PET-CT, fine needle aspiration, endoscopic evaluation with biopsies, and most often bilateral tonsillectomy. At 21 of 22 university hospitals the treatment decision is made at a multidisciplinary conference. Three of seven Swedish centres use only radiotherapy or chemoradiotherapy to treat N+ HNCUP patients. Robotic surgery for biopsy of the tongue base is beginning to become an alternative to targeted biopsies in Sweden and Finland. Narrow Band Imaging is used only in Finland.

Keywords:

• Cancer of unknown origin
• follow up
• multidisciplinary tumour board meeting
• PET-CT

Introduction

The definition of Head and Neck Cancer of Unknown Primary (HNCUP) is the presence of a cervical lymph node metastasis, for which therapy will be initiated even though there is no specific evidence of a primary tumour. Extensive clinical and radiological examinations are typically undertaken before initiating treatment of HNCUP.

The incidence of HNCUP varies between 3–7% of all neck lumps [1]. Many patients with an occult neck lump initially diagnosed as a HNCUP are later identified with a primary tumour. Waltonen et al. [2] found that nearly half of all primary tumours could be located using PET-CT and endoscopy. A more recent report suggested that males constitute 72% of all cases, the median age at onset being ∼55 years (range = 42–87 years) [3]. Lymph node metastases located in the supraclavicular fossa are frequently found to be of infra-clavicular origin. Fine needle aspiration cytology and modern molecular diagnostics have improved the assessment of cancer of unknown primary. Accurate cytological diagnosis of HNCUP is essential when choosing treatment modality, and may become increasingly important in the HPV era [4,5].

It has been recognized that the diagnostic and treatment modalities for HNCUP cases differ, and the present study is the first attempt to systematically assess these differences in the Nordic countries (Iceland, Norway, Denmark, Sweden, and Finland). The recommendation is that the diagnostic work-up should include both PET-CT and panendoscopy with biopsies, including bilateral tonsillectomy [2,6]. Such an action may offer the greatest likelihood of successfully identifying an occult primary tumour. However, the availability of PET-CT varies, as do the guidelines to use it.

Therapeutic approaches vary between centres, and treatment modalities include surgery (lymph node excision or neck dissection) with or without post-operative radiotherapy; radiotherapy, alone or in combination with chemotherapy; and radiotherapy, followed by surgery. In the early stage (N1), neck dissection and radiotherapy seem to have similar efficacy, whereas more advanced cases (N2, N3) necessitate a combination of modalities [6]. One retrospective study concluded that radiotherapy after neck dissection influenced the overall survival [7], while another study suggested that IMRT was the most beneficial radiotherapeutic approach [8]. Mistry et al. [9] reported that the overall 5-year survival was better for HNCUP patients than for those with a known primary of comparable nodal stage. The overall 5-year survival rate for the HNCUP group was found to be 55%. In line with this, another study reported the 5-year survival rate to be poorer among HNCUP patients where a primary tumour was later detected, when compared with patients having a persistent unknown primary (22% vs 52%). Significant prognostic factors in HNCUP were M-class, smoking, alcohol consumption, and tonsillectomy [6].

The aim of this study was to compare diagnostic procedures for and treatment of HNCUP at the head and neck cancer centres of university hospitals in the five Nordic countries.

Materials and methods

A multiple-choice questionnaire was designed in English by the three authors, based on clinical experience and on appropriate current literature. The questionnaire was designed to identify differences in diagnostic and treatment protocols between the Nordic university hospitals. Questions regarding outcome data were not included in this survey. The questionnaire was first tested at the Finnish university hospitals. Consequently, physicians in Finland were used as a pilot group in order to fine-tune the questions. After collecting the comments from the Finnish university hospitals, the questionnaire was revised accordingly by the authors. It now comprised four parts (General information (2 questions), Diagnostic procedures for HNCUP (11 questions), Treatment of HNCUP (15 questions), Follow-up of HNCUP (3 questions)). Early in 2015 the revised questionnaire was sent to all university hospitals in the Nordic countries (Figure 1), representing a total population of ∼25 million people.

Figure 1. Map showing all the university hospitals managing HNCUP included in this study; (1) Reykjavik, (2) Oslo, (3) Bergen, (4) Trondheim, (5) Tromsoe, (6) Aalborg, (7) Aarhus, (8) Odense, (9) Herlev, (10) Copenhagen, (11) Lund, (12) Gothenburg, (13) Linköping, (14) Örebro, (15) Stockholm, (16) Uppsala, (17) Umeå, (18) Helsinki, (19) Turku, (20) Tampere, (21) Kuopio, and (22) Oulu.

All university hospitals in Iceland (1), Sweden (7), Norway (4), and Finland (5) kindly responded to the questionnaire. Denmark referred to their national guidelines for HNCUP (www.DAHANCA.oncology.dk/Brows_Web_Guidelines 2013), and a senior oncologist representing DAHANCA completed the questionnaire accordingly.

Results

The management of HNCUP in Iceland and Sweden followed a specific protocol based on local routines regarding diagnostic procedure and treatment. Finland had national guidelines for the treatment of head and neck cancer, but no specific protocol for the diagnostics of HNCUP. In Denmark, all centres applied the same national guidelines, while, in Norway, two university hospitals used local protocols, but no specific protocol was followed at the remaining two centres. Patients were discussed at weekly Multidisciplinary Tumour Board Meetings (MTBM) at all sites except in Iceland.

In Sweden, two out of seven centres performed PET-CT (Neck-Chest) as a primary imaging modality, and the remaining centres performed conventional CT (Neck-Chest) first. PET-CT was applied in all cases of HNCUP in Denmark and Finland. In Iceland, PET-CT was unavailable. In Norway two centres performed PET-CT in selected cases, whereas the others used CT or MRI as their first choice of imaging (Table 1).

Table 1. Diagnostic procedures.

	n	%
Imaging modality
PET/CT	14	64
CT	7	32
MRI	1	5
Cytological method (needle aspiration)
Fine	22	100
Middle	0	0
Thick	0	0
Sites included for scopy
B, E, H, L	11	50
H, L	11	50
Biopsies
N, TB, T	17	77
N, TB	2	9
TB, T	2	9
T	1	5
Tonsillectomy
Unilateral	2	9
Bilateral	17	77
Bilateral in selected cases	3	14
Robotic surgery
Yes	5	23
No	17	77
Assessment under anaesthesia
Palpation	22	100
Narrow band imaging	5	23
Toluidine blue	0	0

B: bronchoscopy; E: esophagoscopy; H: hypopharyngoscopy; L: laryngoscopy; N: nasopharynx; TB: tongue base; T: tonsil.

Fine needle aspiration of the cervical mass was reportedly carried out at all the participating centres in the Nordic countries. The relation to HPV positivity was evaluated at all centres (except in Iceland), with either p16 or PCR technique. EBV analysis was not a routine procedure in any of the Nordic countries.

Panendoscopy, including bronchoscopy, esophagoscopy, hypopharyngoscopy, and laryngoscopy, was performed at all centres in Finland. In Iceland and Denmark, selective endoscopies were performed, including hypopharyngoscopy and laryngoscopy. In Sweden and Norway, local routines were followed. Twenty-one of 22 centres performed endoscopic procedures after the primary imaging.

Biopsy samples to detect a primary tumour were taken from the tongue base, and epipharynx at all centres in Denmark and Sweden. In Finland, Norway, and Iceland, biopsies were taken according to their local routines. In addition, tonsillectomy was performed either ipsi- or bilaterally at all centres. Most centres in Denmark (5/5), Sweden (6/7), Norway (2/4), and Finland (4/5) performed bilateral tonsillectomy as a standard procedure. The remaining five centres performed bilateral tonsillectomy only in selected cases.

Robotic surgery for biopsy of the tongue base was beginning to become an alternative to targeted biopsies in Sweden and Finland; the other countries had not yet introduced this procedure into their protocol. In Finland, Narrow Band Imaging was used during endoscopy to improve the information where a biopsy should be taken.

In Sweden, three centres did not perform neck dissections in N+ HNCUP patients; instead, they used solely radiotherapy or chemoradiotherapy. All other university hospitals used radiotherapy or chemoradiotherapy combined with neck dissection (Table 2).

Table 2. Treatment.

	n	%
N1
Treatment
Surgery	5	23
RT	2	9
CRT	7	32
Surgery + RT	1	5
Surgery + CRT	7	32
Radiotherapy dose (Gy)
50–70	16	72
0ther	1	5
No RT	5	23
Radiotherapy of the neck*
Ipsilateral	5	24
Bilateral	11	52
No RT	5	24
N2
Treatment
Surgery	—	—
RT	2	9
CRT	2	9
Surgery + RT	7	32
Surgery + CRT	11	50
Radiotherapy dose (Gy)
50–70	21	95
0ther	1	5
Radiotherapy of the neck*
Ipsilateral	4	19
Bilateral	17	81
N3
Treatment
Surgery	—	—
RT	1	5
CRT	2	9
Surgery + RT	7	32
Surgery + CRT	12	54
Radiotherapy dose (Gy)
50–70	22	100
Radiotherapy of the neck*
Ipsilateral	2	10
Bilateral	19	90

* Data missing from one centre.

Concerning follow-up, there was a disparity between Finland and the other Nordic countries. In Finland, patients treated for HNCUP were always monitored by the ENT surgeon, whereas the other Nordic countries most often used a combined approach with visits alternatingly to ENT physicians and oncologists. All Nordic patients were followed up for 5 years after treatment. Most patients were initially seen by the primary treating centres, although a combination follow-up was often organized, where every second visit was performed by the referring hospital.

Discussion

The initiative to undertake this study was the lack of internationally accepted guidelines for diagnostic procedures and treatment specifications for HNCUP patients. The study was carried out at a tertiary hospital level in northern Europe. The five Nordic countries combined have a rather homogeneous population (∼ 25 million) regarding socioeconomic and health-related factors. The countries’ healthcare systems are similar, but there are still no unified treatment protocols between them to tackle HNCUP. It, therefore, seemed justified to evaluate the differences in existing guidelines in this field. We chose to undertake the present investigation, specifically the management of HNCUP, as this would merely demonstrate the differences in guidelines between the centres and not, for example, their availability to use the various advanced reconstructive or oncological techniques.

It is noteworthy that this survey covered all 22 university hospitals in the five Nordic countries treating these patients. This collaborative effort was made possible by the existing membership network of the Scandinavian Society for Head and Neck Oncology. The Society strives to promote collaboration between head and neck surgeons and oncologists in the Nordic countries, to stimulate multi-centre studies, and to assist any advances toward unified treatment guidelines.

In the present study 14 of the 22 centres (64%) used PET-CT as a primary imaging modality, although even PET-CT is limited in its ability to detect small tumours. In a meta-analysis comprising 13 studies, the sensitivity and specificity values of PET-CT were 0.84 and 0.96, respectively, whereas those of conventional imaging were 0.63 and 0.96, indicating that PET-CT may be more sensitive than conventional imaging [10]. However, PET-CT may have a problem with high false-positivity rates, due to relatively low specificity, which is why multiple biopsies from suspicious sites are recommended even after PET-CT [11]. Furthermore, it is important that treatment initiation is not delayed due to lacking availability of PET-CT [12].

Five centres in Sweden and Finland used robotic surgery to improve the quality of biopsies from the tongue base. In a study by Karni et al. [13] on the use of transoral laser microsurgery, a 94% primary tumour detection rate was demonstrated, suggesting better specificity than with targeted biopsies. To date, only Finnish university hospitals use Narrow Band Imaging on a regular basis in order to optimize the quality of the biopsy. In a recent review, Koivunen et al. [14] reported an increased detection rate of the primary tumour from 80% to 90%, following the introduction of Narrow Band Imaging technique, combined with PET-CT and trans-oral diagnostic procedures. Together, these studies add valuable information to a growing body of evidence that new diagnostic tools can improve diagnostic accuracy for primary tumours. MTBM is regarded as the gold standard for treatment decision and optimization of patient care. Nguyen et al. [15] conclude that, when all team members of the MTBM take part in the treatment decision-making, it will be based on performance status rather than inherent bias due to age or perceived comorbidity. It should be noted that all the Nordic university hospitals except one used MTBM as standard routine.

All countries except one either followed a specific protocol or national guidelines for the management of HNCUP. Only three centres treated patients with single modality radiotherapy/chemoradiotherapy for N+ necks, whereas the majority combined neck dissection with radiotherapy/chemoradiotherapy (86%).

The great majority of previous studies are based on retrospective data. Shoushtari et al. [16] found that pre-operative IMRT (50–56 Gy) followed by neck dissection gave excellent overall and disease-free survival in N1–N2a disease (100%). Furthermore, they found that patients with N2b–N3 disease, with a significantly lower survival rate (66.7%, p = 0.017), benefitted from concurrent chemotherapy, targeted therapeutic agents or accelerated radiotherapy regimens, in addition to surgery. Cuaron et al. [17], however, found that conventional radiotherapy produced excellent locoregional control of HNCUP with acceptably low levels of late toxicity. Chen et al. [18] reported that concurrent chemoradiotherapy was associated with significantly increased toxicity without definite benefit in overall survival or locoregional control in the treatment of HNCUP. It has been suggested that post-operative radiotherapy does not influence the rate of neck relapse (p = 0.72), although when Issing et al. [7] compared radiotherapy alone with neck dissection plus post-operative radiotherapy, they found a significantly improved survival rate in patients who underwent a diagnostic bilateral tonsillectomy in addition to combined modality treatment. We speculate that the reason why some centres do not perform neck dissection could be that the occult primary tumour often arises in the oropharyngeal region, which is included in the radiotherapy field. Furthermore, the probability of an HPV-positive tumour in this region is high [5] and the prognostic implications of HPV-positive nodes in HNCUP are similar to those in oropharyngeal primary cancers [4]. When radiotherapy or chemo-radiotherapy is given, a small occult primary tumour will be cured, and most likely never even be detected during the follow-up. For this reason, some centres maintain a conservative approach to surgery, sparing the patient from surgical side-effects, even though a selective neck dissection is associated with little post-operative morbidity.

In Denmark, all centres have consented to follow their national guidelines. During this study, Sweden published its first national guidelines (Huvud och halscancer – Nationellt vårdprogram 2015 Aug), although there is some scope for local variation in the diagnostic work-up and treatment recommendations. In Finland all university hospitals follow national guidelines for the treatment of head and neck cancer, but there are obvious variations regarding certain details, as could be observed for the diagnosis of HNCUP in the current study. The other Nordic countries adapt their management according to local traditions.

This survey contributes to an on-going discussion intended to improve the quality of diagnosis and treatment of HNCUP patients. In conclusion, the main diagnostic tools currently used are PET-CT, fine needle aspiration, and endoscopic evaluation with biopsies from the nasopharynx, tongue base, and bilateral tonsillectomy. Treatment modalities include radiotherapy/chemoradiotherapy or neck dissection combined with radiotherapy/chemoradiotherapy. The addition of chemotherapy is generally given when the patient is fit to receive it and below 70 years of age.

The current study has established a suitable network and platform in northern Europe based on a population exceeding 25 million. Such a multi-centre network can be used to compare survival data according to the given treatment, and thereby facilitate future evaluation of survival differences between the two major treatment groups: radiotherapy/chemoradiotherapy, or surgery combined with radiotherapy/chemoradiotherapy. Future multi-centre studies using this starting point may hopefully improve our understanding of the diagnostic tools, and how to manage HNCUP. Moreover, this Nordic network could be used to establish common guidelines for follow-up of patients with HNCUP, regarding recurrence, late detection of primary tumours, and treatment squealae.

Acknowledgements

We gratefully acknowledge all colleagues in the Nordic countries who have contributed to the data collection: Iceland: Arnar Gudjonson; Norway: Jan Folkvard Evensen, Odd Monge, Åsa Karlsdottir, Marianne Brydoy, Hanne Tondel, John Sigurd Lybeck, and Oddveig Rikardsen; Denmark: Kenneth Jensen; Sweden: Johan Wennerberg, Eva Brun, Magnus Niklasson, Hedda Haugen, Maciej Tytor, Måns Agrup, Jan Rzepecki, Anders Westerborn, Johan Reizenstein, Anders Högmo, Helena Sjödin, Tomas Ekberg, Lena Cederblad, Magnus Wahlgren, and Karin Söderström; and Finland: Leif Bäck, Kauko Saarilahti, Jaakko Pulkkinen, Ilpo Kinnunen, Paula Lindholm, Reidar Grénman, Heikki Minn, Jussi Laranne, Tuija Wigren, Matti Pukkila, Leena Voutilainen, Petri Koivunen, and Merja Korpela. We would also like to acknowledge the Scandinavian Society for Head and Neck Oncology.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.