Introduction
Multimodal treatment options for head and neck squamous cell carcinoma (SCC) include surgery, radiotherapy and cisplatin-based chemotherapy. These treatment modalities have allowed for greater control of locoregional disease. Unfortunately, this increased local control has not translated to a significant overall survival advantage for patients with head and neck SCC. The dismal increase in overall survival is partially because these treatment modalities have no influence over the rate of development of distant metastases.Reference Hauswald, Simon, Hecht, Debus and Lindel1–Reference Garavellow, Ciardo, Spreafico and Gaini4
Head and neck SCC is currently classified as human papillomavirus (HPV)-positive and HPV-negative based on cancer aetiology. The main traditional risk factors for head and neck SCC are smoking and alcohol use; however, over the past three decades, HPV has emerged as an independent risk factor for head and neck SCC. Defining these groups has implications for treatment options and prognosis. The HPV-positive oropharyngeal head and neck SCC patients seemingly present with more extensive lymph node involvement (greater regional disease burden), yet there is greater locoregional control of disease in these patients and a survival advantage with treatment options, when compared with HPV-negative patients.Reference Joo, Cho, Park, Nam, Kim and Kim5–Reference Näsman, Attner, Hammarstedt, Du, Eriksson and Giraud7
Head and neck SCC incidence is increasing globally, with HPV-positive cases linked to this rise.Reference Jones8 With the increase in locoregional control of disease in head and neck SCC patients as a result of multimodal treatment options, especially in the HPV-positive group, there has been an associated increase in distant metastases formation.Reference Huang, Perez-Ordonez, Weinreb, Hope, Massey and Waldron9 Future developments in head and neck cancer treatment must try to influence the survivorship from distant metastases.
The present methods of head and neck SCC tumour staging include primary tumour site, nodal status, histological grade and the presence of extracapsular spread.Reference Garavellow, Ciardo, Spreafico and Gaini4 These current parameters, which use clinical and pathological data as surrogate markers of cancer biology, are employed to identify those head and neck SCC patients at risk of treatment failure and/or predict overall survival. A greater understanding of this tumour biology and, in particular, how distant metastases occur, is clinically necessary.
Circulating tumour cells are viable cells that have been shed by the primary tumour and circulate within the blood stream of cancer patients. Along with circulating tumour cells, disseminated tumour cells have been identified; these are viable cancer cells detectable within the bone marrow of cancer patients. Circulating tumour cells and disseminated tumour cells have been studied in many cancers, including cervical, colorectal, prostate and breast cancer, and head and neck SCC.Reference Fehm, Banys, Rack, Jäger, Hartkopf and Taran10–Reference Hristozova, Konschak, Stromberger, Fusi, Liu and Weichert14 The bone marrow is seemingly a preferred destination for these cancer cells, in keeping with the ‘seed and soil’ theory originally purported by Stephen Paget.Reference Langley and Fidler15 This theory proposes that the cancer cells (the seeds) will metastasise to an organ (the soil) with favourable conditions for their growth and development.
Circulating and disseminated cancer cell detection
One millilitre of human blood carries approximately 1 billion red blood cells, 7 million white blood cells and 295 million platelets.Reference Möckelmann, Laban, Pantel and Knecht16 It has been estimated that a patient with metastatic cancer will have one circulating cancer cell to one billion of these normal cells, making these circulating tumour cells difficult to detect.Reference Möckelmann, Laban, Pantel and Knecht16
Haematopoietic stem cells reside within the bone marrow, and give rise to lymphocytic and myelocytic cell lineages. On maturation, these haematopoietic cells leave the bone marrow and circulate within the blood. Thus, in healthy individuals, the peripheral blood stream and bone marrow harbour cells of haematopoietic origin (cells originating from the embryological mesodermal germ cell layer).
Two complementary strategies are commonly employed to detect epithelial cells within mixtures of other cell types: detecting a cell marker (e.g. a cell surface protein such as E48, or an epithelial cytoskeleton protein such as a member of the family of cytokeratins) or detecting messenger RNA (mRNA) of a gene that encodes a protein preferentially transcribed in epithelial cells.Reference Berois, Varangot, Aizen, Estrugo, Zarantonelli and Fernández17, Reference Smith, Slade, English, Graham, Lüchtenborg and Sinnett18
The detection of the epithelial cell cytoskeletal proteins, cytokeratins or cell surface proteins is typically performed by immunohistochemistry, using an antibody that has specificity for the protein under investigation.Reference Fehm, Braun, Muller, Janni, Gebauer and Marth19 However, cells undergoing proliferation or apoptosis may release cytokeratins into the blood stream; hence, there may be a low detectable level of cytokeratins in a peripheral blood sample of healthy individuals.Reference Klein20, Reference Schardt, Meyer, Hartmann, Schubert, Schmidt-Kittler and Fuhrmann21 An additional frailty of this approach is the potential loss of cytokeratin expression as cells de-differentiate during the process of carcinogenesis.Reference Alix-Panabières, Riethdorf and Pantel22
Detection of specific mRNA transcripts usually involves polymerase chain reaction.Reference Czerkinsky, Nilsson, Ouchterlony, Tarkowski and Nygren23 For circulating and disseminated tumour cell detection, reverse transcription polymerase chain reaction is performed. Virtually all cells within the human body will contain a full genomic complement. Identifying epithelial cells requires the detection of genes that are expressed specifically within epithelial cells. To do this, mRNA of expressed proteins is converted (reverse transcribed) into copy DNA, which is then detected. Unfortunately, background expression of a gene not normally associated with that cell lineage may also occur. An example of this is normal white blood cells that express carcinoembryonic antigen.Reference Hampton, Walker, Marshall and Juhl24 Therefore, in order to use this approach, a threshold setting has to be determined for the transcript under investigation.
A third possible method of detecting circulating tumour cells and disseminated tumour cells is to culture cells from a peripheral blood sample in order to identify proteins secreted exclusively from epithelial cells. An advantage of this approach is that cells would have to be viable to secrete proteins for this technique to work.
Background work
Much of the work on circulating tumour cell detection has been conducted in breast cancer patients. Cristofanilli et al. conducted a double-blinded, multicentre, prospective study to establish whether circulating tumour cells could be a marker of survival in metastatic breast cancer.Reference Cristofanilli, Budd, Ellis, Stopeck, Matera and Miller25 The authors recruited 177 women with breast cancer and 145 women in a control group. The control group contained 72 premenopausal women and 73 post-menopausal women who had been diagnosed with either benign breast disease or other benign conditions, with no past or current history of an epithelial-derived tumour. The circulating epithelial cell level was found to be 0.1 ± 0.2 cells in 7.5 ml (standard peripheral blood collection tube volume) of whole blood in the healthy group and 0.1 ± 0.9 cells per 7.5 ml in patients with benign breast disease. Sixty-one per cent of the breast cancer cohort had two or more circulating tumour cells per 7.5 ml of whole blood. In the breast cancer group, measurements of circulating tumour cells were taken prior to treatment and then at the first follow-up visit (usually 3 to 4 weeks after treatment), and again at regular intervals of 9 to 12 weeks. Ten cancer patients died prior to the first visit; these patients all had very high levels of circulating tumour cells (counts from 9 to 618 cells per 7.5 ml of blood). Patients were categorised into two groups: those with fewer than five circulating tumour cells per 7.5 ml of whole blood and those with five or greater circulating tumour cells per 7.5 ml of blood. Patients in the latter cohort had shorter median progression-free survival times (2.7 months vs 7.0 months, p < 0.001) and shorter overall survival (10.1 months vs more than 18 months, p < 0.001).
In a much larger (pooled analysis) study, Braun and colleagues examined the significance of disseminated tumour cells in the bone marrow of breast cancer patients.Reference Braun, Vogl, Naume, Janni, Osborne and Coombes26 Nine studies taken together had a total of 4073 stage I, II and III breast cancer patients, allowing analysis of outcomes over a 10-year period. Approximately one-third of patients had detectable disseminated tumour cells within the bone marrow at the time of presentation. These patients were found to have a greater tumour burden, with larger primary tumours, lymph node involvement, a higher histological grade and hormone receptor negativity. Similar to the Cristofanilli et al. study on circulating tumour cells,Reference Cristofanilli, Budd, Ellis, Stopeck, Matera and Miller25 disseminated tumour cells were associated with poorer disease-free survival and an overall poorer outcome. Approximately 50 per cent of the patients with disseminated tumour cells had relapse of the cancer, suggesting that in the remaining 50 per cent of patients the disseminated tumour cells might not have been viable.
Head and neck cancer
Colnot and co-workers studied the clinical significance of disseminated tumour cells within the bone marrow of head and neck cancer patients.Reference Colnot, Nieuwenhuis, Kuik, Leemans, Dijkstra and Snow27 Iliac crest bone marrow aspirations were taken from 139 surgically treated primary head and neck SCC patients. The primary statistical endpoints were disease-free survival and distant metastases free survival. The bone marrow samples of 30 non-cancer patients were used as a control group. Disseminated tumour cells were not found in any of the control group samples, but were found in 40 per cent of the cancer group samples. The results did not demonstrate a statistically significant decrease in survival of those harbouring disseminated tumour cells (although this group fared worse overall). However, head and neck SCC patients with detectable disseminated tumour cells, who had two or more involved lymph nodes, had a statistically significant decrease in survival compared to their counterparts with two or more involved lymph nodes and no disseminated tumour cells.
Hristozova and co-workers assessed the frequency of circulating tumour cells and their association with clinicopathological markers during chemoradiotherapy in head and neck SCC patients.Reference Hristozova, Konschak, Stromberger, Fusi, Liu and Weichert14 Peripheral blood samples were taken prior to commencing treatment, and then at a defined early time point in treatment (defined as less than four weeks). A further peripheral blood sample was taken after four weeks of treatment. The authors discovered that 18 of 42 patients with locally advanced, inoperable head and neck SCC harboured circulating tumour cells prior to the commencement of treatment. No patients within this study had clinical or radiological evidence of distant metastases. There was no association found between tumour volume or HPV status (40 per cent of the patients within this study had oropharyngeal tumours). However, a nodal stage of N2b or greater was associated with the presence of circulating tumour cells. Chemoradiotherapy reduced circulating tumour cell detection, although 20 per cent of individuals had evidence of residual circulating tumour cells after completion of treatment. The study follow-up time was not long enough to discover whether these 20 per cent went on to suffer either locoregional or distant recurrence (i.e. treatment failure). Of note, the authors stated that circulating tumour cell numbers decreased after the radiotherapy treatment sessions had begun, prior to the commencement of chemotherapy; minimal further reduction in circulating tumour cell numbers was observed with the addition of chemotherapy. This finding does not fit in with standard clinical practice within the UK, where chemotherapy and radiotherapy are given concurrently.
Buglione and colleagues investigated the presence and importance of circulating tumour cells in patients presenting with primary head and neck SCC from a variety of anatomical subsites.Reference Buglione, Grisanti, Almici, Mangoni, Polli and Consoli28 They concluded that circulating tumour cells were most commonly found in patients with oropharyngeal tumours (HPV status was not mentioned in the article), hypopharyngeal tumours and undifferentiated sinonasal tumours. Higher circulating tumour cell counts correlated with later stage (stage IV) disease, and counts declined in response to treatment. The authors did not state which, if any, treatment option (chemotherapy, radiotherapy, surgery or multimodal treatment) had a greater effect on the reduction in circulating tumour cell load.
Tinhofer and colleagues assessed circulating tumour cell load in patients treated with chemoradiotherapy, or induction chemotherapy followed by radiotherapy combined with cetuximab, for locally advanced (stage IV) head and neck SCC.Reference Tinhofer, Hristozova, Stromberger, Keilhoiz and Budach29 Blood samples from the study patients were collected at the following time points: before treatment; after three cycles of docetaxel, cisplatin, and 5-fluorouracil induction chemotherapy; at the end of treatment with radiotherapy plus cetuximab or chemoradiotherapy; and at six weeks and three months following completion of treatment. The authors observed a rise in circulating tumour cell numbers within the peripheral blood post-radiotherapy. This effect was reduced with the use of concurrent cetuximab but not with chemotherapy. The concern following this observation is that although radiotherapy has well established benefits against the primary tumour, there may be radioresistant clones which disseminate following treatment with radiotherapy.Reference Camphausen, Moses, Beecken, Khan, Folkman and O'Reilly30, Reference Biswas, Guix, Rinehart, Dugger, Chytil and Moses31
Conclusion
From the evidence presented in this article, circulating tumour cells and disseminated tumour cells may be clinically useful as prognostic markers or in the assessment of response to treatment in head and neck SCC.
Head and neck cancer distant metastases are typically diagnosed through radiological assessment, and are seldom biopsied or excised to allow laboratory analysis. The paucity of metastatic tissue to analyse in the laboratory means that the biology and biochemistry of head and neck metastases is rarely characterised. The possibility exists that head and neck SCC distant metastases arise from circulating and disseminated cells that have become activated to seed in organs such as the lungs, liver and bone. If this is the case, laboratory analysis of these circulating and disseminated cells could give us a biological continuum of distant metastases, allowing determination of how the metastatic process occurs.
There is plenty of evidence demonstrating that patients with low-volume primary tumours of the head and neck, without radiological or pathological evidence of lymph node involvement or distant spread, have a good prognosis following treatment.Reference Habbous, Harland, La Delfa, Fadhel, Xu and Liu32, Reference Brockstein, Haraf, Rademaker, Kies, Stenson and Rosen33 However, a number of these patients will have circulating tumour cells and disseminated tumour cells. Though these cells have spread throughout the peripheral blood and are being harboured within organs such as the bone marrow, these patients have good overall survival rates.Reference Möckelmann, Laban, Pantel and Knecht16 A hypothetical mechanism for this improved survival is early dissemination of cancer cells in an immune-competent patient, which may activate host immunity against the disseminated cancer cells and therefore versus cells of the primary tumour. Hence, it is conceivable that a survival advantage may be conferred in the context of low-volume locoregional disease. In contrast, the dissemination of tumour cells may enhance the development of tumour-related immune tolerance, thereby providing favourable conditions for tumour progression.Reference Freiser, Serafini and Weed34 Whether either or both of these outcomes are possible has yet to be established, but this is a compelling angle for future research.
If circulating tumour cells and disseminated tumour cells are the cells from which distant metastases in head and neck cancer arise, studies into new treatment options which eradicate these cells is of paramount importance.
