Nasopharyngeal carcinoma. A “different” head and neck tumour. Part A: from histology to staging| ACTA Otorhinolaryngologica Italica

Abstract

As in some other sites of the head and neck, a broad range of tumours can arise in the nasopharynx: epithelial, mesenchymal, lymphoid, and neuro-ectodermal. However, nasopharyngeal carcinoma (NPC) is the most interesting and intriguing type, because it is a “peculiar” malignancy, and “different” from almost all other head and neck tumours according to several points of view. In fact, NPC is a unique disease whose aetiology, clinical behaviour, epidemiology, and histopathology are different from those of all other squamous cell carcinomas of the head and neck. The most distinguishing characteristics of this malignancy are the consistent association with Epstein-Barr virus and the striking geographical differences in its incidence. The aim of this narrative review is to analyse the very large number of studies (sometimes contradictory) on NPC. In this first part, the histopathology, aetiology, epidemiology, clinical behaviour, natural history, diagnostic work up, and staging will be examined.

Introduction

It is well-known that nasopharyngeal carcinoma (NPC) is rare in most parts of the world, but is endemic in North Africa and Southeastern Asia, and most notably in South China. This epidemiologic characteristic may be enough to define NPC as a peculiar malignancy among head and neck tumours. Furthermore, NPC has several other characteristics that have fascinated generations of oncologists, pathologists, scientists, and epidemiologists ¹. In this review, I will discuss the various, and sometimes not yet solved, problems concerning NPC.

Histopathology

In 1921 Regaud ² and Schmincke ³, independently, used the term “lymphoepithelioma” to designate a highly radiosensitive neoplasm of the nasopharynx showing nests of non-keratinising squamous cells embedded in a lymphoid stroma. The unusual histological features of NPC have generated controversies over the nature of the tumour. There have been more names coined for the various histological subtypes of NPC than any other tumour type. The World Health Organization (WHO) Classification of Tumours Pathology and Genetics of 2005 ¹ reports as many as 12 synonyms (Tab. I). The correct histologic setting of this tumour remained a controversial question for many years, as one can see in a study published by Scanlon et al. in 1967 where the author stated “There is probably no other single group of cancers of the respiratory system about which there is less agreement in the world literature with regards to the correct and proper pathological classification” ⁴. The 1978 World Health Organization (WHO) classification ^5,6 recognised three histological subtypes of NPC: squamous cell carcinoma (WHO type 1), non-keratinising carcinoma (WHO type 2), and undifferentiated carcinoma (WHO type 3). The 1991 WHO classification ⁷ retained the squamous cell carcinoma subtype (keratinising squamous cell carcinoma, K-NPC), while the last two subtypes were combined under a single category of “non-keratinising carcinoma” (NK-NPC), which was further subdivided in “differentiated” (DNK-NPC) and “undifferentiated” (UNK-NPC); lymphoepithelioma-like carcinoma was considered a morphologic variant of undifferentiated carcinoma. The use of numerical designation of WHO types 1, 2, and 3 was eliminated, although many authors continue to use it. The 2005 ¹ and 2017 ⁸ WHO classifications of head and neck tumours maintained the terminology of the 1991 classification, with the addition of the category basaloid squamous cell carcinoma (Tab. II). However, the difficulties in making an undisputable histological diagnosis have been clearly stated in the 2005 WHO classification ¹: “The wide ranging reported figures on the frequencies of various subtypes indicate that the boundaries between the categories are not always clear, sampling error is a significant problem due to the small size of the biopsies, and intra- and inter-observer reproducibility of the classification is sub-optimal”. However, beyond the aforementioned diagnostic difficulties, the 2005 WHO classification reported that the proportion of K-NPC among all NPC was higher in low-incidence areas compared with high-incidence areas (1% in Hong Kong, 17% in Singapore, 8% in Tunisia, 13% in Japan, and 25% in USA) ¹. The rate of K-NPC was 18% in a series of 171 NPC treated at a single Italian institution between 1990 and 1999 ⁹. The proportion of K-NPC appears to be increasing in non-endemic areas. In fact, a 2013 study done using the Surveillance, Epidemiology, and End Results (SEER) data base, provided a comprehensive description of racial differences among 5,868 NPC patients in the USA and reported 2,456 cases of K-NPC (42.2%). The distributions of K-NPC were also significantly different across races: non-Hispanic white 1520 (52.4%), Hispanic white 89 (42.0%), black 226 (42.9%), Asian 549 (28.2%), other 72 (31.0%) ¹⁰. In a more recent study ¹¹, the authors conducted a retrospective cohort study of 9,995 patients with NPC from the National Cancer Database (NCDB). The proportion of each histological subtype varied substantially between ethnicities. Asian patients had nearly equal distributions across the three subtypes. This varied significantly from black and white patients, among whom K-NPC comprised most cases (60 and 68%, respectively), while UNK-NPC comprised the minority of cases (18 and 12%, respectively). Hispanic patients fell in between, with 51% of cases comprised of the K-NPC type and 19% of cases being of UNK-NPC subtype ¹¹. With regards to NK-NPC, in the 2017 WHO classification there is the following statement: “Subclassification into undifferentiated and differentiated subtypes has no clinical and prognostic value” ⁸. This categorical statement was refuted by the study of Stepan et al. ¹¹ which showed that 5-year overall survival (OS) for UNK-NPC was 69%, while that for DNK-NPC was 62% (the adjusted hazard ratio in comparison with keratinising histology was 0.67 for non-keratinising differentiated and 0.55 for non-keratinising undifferentiated NPC).

In another study, Wu et al. ¹² analysed non-metastatic NPC patients between 2004 and 2014 using the SEER data base. These authors collected information on 2,845 patients, including 1,218 (42.8%), 849 (29.8%), and 778 (27.3%) with K-NPC, DNK-NPC and UNK-NPC, respectively. Within the follow-up period of over 5 years, patients with DNK-NPC had poorer NPC-specific survival (NPC-SS) compared to UNK-NPC, and had comparable NPC-SS between the two subtypes after more than 5 years of follow-up.

Aetiology and epidemiology

Unlike almost all other head and neck carcinomas (HNCs), tobacco smoking and alcohol consumption are contributing factors only for K-NPC. On the contrary, NK-NPC has a complex and unique aetiology that is not yet completely understood. NPC is uncommon in most countries but it is endemic in a few well-defined populations, including natives of southern China, Southeast Asia, the Arctic, and the Middle East/North Africa. Batsakis stated in this regard that “The racial predisposition for NPC is present among these populations whether they are in their native country or have been “transplanted” to other lands” ¹³. Buell, in 1965, carried out a noteworthy epidemiologic study in the Chinese population of California ¹⁴. He found that immigrants from China have 30 to 40 times mortality from NPC than the Caucasian population of California, and that Chinese who were born in the United States have about 20-fold greater mortality. In 1981, Dickson presented a series of 209 patients with NPC and found that “Chinese born in China had a 117.9 times greater incidence of the disease, while North American born Chinese showed only a 7.3 times greater incidence than Caucasians” ¹⁵.

This characteristic ethnic and geographic distribution of NPC worldwide suggests a multifactorial aetiology, including genetic traits, environmental and dietary factors, and infection by Epstein-Barr virus (EBV). Hundreds of studies have tried to elucidate how these aetiologic factors interact in the development of the disease. A thorough study by Chang and Adami defined the epidemiology of NPC as “enigmatic” ¹⁶ (Tab. III).

Regarding EBV infection, a 1966 study by Old et al. ¹⁷ demonstrated the association of EBV and two human malignancies, and its possible oncogenic potential. These authors found that the incidence of EBV positivity was high in two groups of patients: those with Burkitt’s lymphoma (56% of 55 African cases) and those with “epidermoid carcinoma of the postnasal space” (85% of 39 African and American cases) ¹³. Later, other pioneering studies investigated the close association between NPC and EBV ^18-21. Today, querying PubMed using the keywords “nasopharyngeal carcinoma” and “Epstein-Barr virus”, you can find as many as 3,810 results. The oncogenic role of the virus in the genesis of NPC, especially in endemic regions, is universally recognised. The International Agency for Research on Cancer (IARC) classified EBV as carcinogenic in 1997. The association of EBV with NPC is very different between NK-NPC and K-NPC. In fact, while there is a near constant association of EBV with NK-NPC, EBV is generally absent in K-NPV, especially in non-endemic regions. However, Chang stated that “EBV alone is not a sufficient cause of NPC, because virtually all adults worldwide are infected with the virus, yet only a small proportion of individuals develop NPC” ¹⁶. In fact, Kangro et al. found that over 90% of children in Hong Kong were infected with EBV by 8 years of age ²². “Therefore, it is apparent that environmental and/or genetic cofactors also contribute to NPC risk” ¹⁶.

Few studies have investigated the association of human papillomavirus (HPV) in NPC and its possible aetiologic role. The results are contradictory. Dogan et al. ²³ reported on 9/67 patients with HPV-positive NPC. However, 3 cases were shown to be of non-nasopharyngeal origin and were excluded from further analysis because the bulk of the disease was in the oropharynx. Furthermore, the K-NPC/NK-NPC ratio was not reported in the article. Nevertheless, the authors concluded that “The aetiologic role of HPV in NPC is confirmed. The favourable prognostic significance of HPV positivity is similar to that of EBV positivity”. Verma et al. ²⁴ queried the NCDB for NPC with known HPV status. In all, 956 patients met the study criteria; this included 308 patients (32%) with HPV-positive disease and 648 patients (68%) who were HPV-negative. In both the overall and matched cohorts, there were no differences in OS between HPV-positive and HPV-negative patients, and HPV status did not independently predict for OS. Wu et al., analysing a mono-institutional series of 78 NPC patients, reported the same results: “OS was not significantly different between the EBV+ group and the HPV+ group, and double-negative EBV^-/HPV^- tumours had worse OS on univariate analysis compared to EBV+ or HPV+ tumours: however, there was no difference in OS on multivariate analysis after adjusting for covariates of age and baseline health” ²⁵.

Regarding environmental and dietary factors, the most studied non-viral factor involved in NPC aetiology is consumption of salt-preserved fish, a traditional staple food in several NPC-endemic areas. According to Chang et al. “In southern China, intake of salted fish and other preserved foods is particularly high among boat-dwelling fishermen and their families. Furthermore, salted fish is a traditional weaning food, resulting in early and frequent feeding of infants, especially in the Cantonese population and in families of lower socioeconomic status. Childhood exposure, especially at weaning, seems more strongly related to NPC risk than adulthood exposure” ¹⁶. Furthermore, Chang wrote “The process of salt preservation is inefficient, allowing fish and other foods to become partially putrefied. As a result, these foods accumulate significant levels of nitrosamines, which are known carcinogens in animals” ¹⁶. In fact, the carcinogenic potential of salt-preserved fish has been demonstrated by experiments in rats, which develop malignant nasal and nasopharyngeal tumours after salted fish consumption ^26,27. Zou et al. analysed 145 samples of cooked salted fish collected from various areas in China ²⁸. The samples from areas with higher NPC risk showed a higher average level of total volatile N-nitrosamines than those from areas of lower NPC risk. In contrast with salted fish, the frequent consumption of fresh fruits and/or vegetables has been associated with a lower risk of NPC ²⁹.

Regarding genetic traits, a large number of studies have analysed the role of various genetic factors associated with NPC, and in particular polymorphisms in the histocompatibility locus antigens (HLA), cytochrome P450 2E1 (CYP2 E1) and alteration of p53 codon 72 Arg > Pro. Those who wish to investigate thoroughly this item may read the paramount paper by Chattopadhyay et al. ³⁰. The importance of genetic predisposition is demonstrated by an indisputable familial aggregation of NPC. Many epidemiologic studies have demonstrated that the excess risk was generally 4- to 10-fold among individuals with a first-degree relative with NPC, compared with those without a family history. However, such clustering can result from shared genetic susceptibility, shared environmental risk factors, or both. Multiple genetic and environmental factors, rather than a single major susceptibility gene, most likely explain the observed pattern of inheritance ¹⁶.

Regarding sex and age distributions, the incidence of NPC is 2-to 3-fold higher in men than in women in almost all populations ¹. In low-risk populations, NPC incidence increases progressively with increasing age in both genders, and the peak incidence occurs in the 65-69 year age group. In contrast, in endemic countries the incidence peaks around 50 to 59 years and declines thereafter, suggesting the involvement of exposure to carcinogenic agents early in life.

NPC is a very rare disease in childhood. Its incidence varies widely in different regions, reflecting the aforementioned interactions between genetic and environmental factors. Italy is a low-incidence area. The annual incidence of NPC has been estimated at 0.73 per million children (ages 10-14 years) and at 1.08 per million adolescents (ages 15-17 years), with an incidence of nil among children aged _< 10 years ³¹. Similar data emerged from the SEER database. From 1988 to 2006, 129 children/adolescents (ages 0-19 years) were registered, for an incidence of 0.5 per million person-years. In contrast, 6,000 adult cases were collected in the SEER database during the same period ³².

The incidence of NPC is declining in all countries, “in Hong Kong since the 1970s, in Taiwan since the 1980s, and in Singapore Chinese since the late 1990s” ¹⁶. Females had a significantly lower age-standardised incidence (male/female ratio 2.5-2.6, p < 0.01) and mortality (male/female ratio 3.0-3.5, p < 0.01) throughout the period 1980-1999 ³³. This lag in trend may be attributable to economic development. Especially in Hong Kong, the lifestyle for most citizens changed progressively from a traditional Southern Chinese style to a more western style, particularly in terms of diets. Preserved salted fish was no longer a common food for most citizens. Although it remains difficult to single out the exact aetiological factor(s), one can at least confidently postulate that the declining incidence of NPC in Hong Kong is mainly attributed to changing environmental risk factors ³³.

Clinical behaviour and natural history

The nasopharynx is a roughly cubical space. It is bound superiorly and posteriorly by bone walls (basi-sphenoid, basi-occiput, and atlas). These structures may be infiltrated and eroded by the NPC, but this is a late occurrence. The nasopharynx tapers inferiorly because it is bound by the upper surface of the soft palate. Anteriorly, it communicates with the nasal cavity. The lateral wall (clinically the most important as regards NPC) is mostly occupied by the opening of the Eustachian tube with a cartilaginous elevation (torus tubarius). Below this opening, the mucosa bulges because of the underlying levator veli palatini muscle. Above and behind this cartilage there is the pharyngeal recess, called the fossa of Rosenmüller. A recently published editorial commentary showed the original wonderful images and the text of Rosenmüller’s description in 1805 ³⁴. This recess is of paramount importance in the local spread of NPC, because it is closely related with the upper parapharyngeal space and the skull base (anterior foramen lacerum and foramen ovale). Chong et al. ³⁵ studied 114 patients with NPC and demonstrated that these foramina are the most common route of early spread to the cavernous sinus and middle cranial fossa, even without macroscopic skull base erosion. In this case, the sixth cranial nerve is the most commonly involved with resulting diplopia.

Carcinoma in situ of the nasopharynx (NPCIS) has rarely been reported. The revision of a large series of specimens of nasopharyngeal biopsies performed in 10 years in Hong Kong identified three cases of NPCIS ³⁶. Two of the three cases subsequently developed into invasive NPC after initial presentation. The interval of transformation varied from 40 to 48 months. In all three cases, the specimens showed abnormal findings on light microscopy and positive staining for EBV-encoded RNA (EBER) with in situ hybridisation (ISH). Elevated anti-viral capsid antigen (anti-VCA) titres were present in two of the pre-invasive lesions. No cell-free EBV DNA was detected in the sera of these patients during the pre-invasive phase of the disease (Tab. IV).

In the nasopharynx there is a rich lymphatic network, making cervical lymph node (LN) involvement from NPC early and frequent. A painless enlargement of the upper cervical LNs was the most common first symptom in several series (75-80%) ³⁷. Patients with NPC often present with locally advanced disease, hence it is sometimes impossible to ascertain the precise area of origin of the tumour. However, Lederman analysed 346 cases of NPC and reported significant differences in the frequency with which the various areas of the nasopharynx were primarily involved by the tumour ³⁸. The lateral wall was the most common site, followed by the superior wall. These findings are coherent with the rates of early presenting symptoms. In a Californian series of 101 patients presented by Wang et al. ³⁹, the most common symptoms were ear-related, and neck masses were the second. Lee et al. ³⁷ studied a very large series of 4,768 patients with NPC treated at the Queen Elizabeth Hospital, Hong Kong. In this endemic area, the most common symptoms presented by patients were the painless enlargement of the upper cervical LN(s) (74.5%), nasal symptoms (discharge, bleeding, and/or obstruction, 73.4%) and otological symptoms (62.4%). These “tubal occlusive” symptoms ¹³ are often followed by otitis media with effusion (OME) (62.4%) ³². While OME is a common condition in the paediatric population, it is less prevalent in adults. Some authors emphasised that a lasting OME causing deafness is a recognised indicator of nasopharyngeal obstruction and the possibility of a nasopharyngeal malignancy must be considered in all adults, especially in non-endemic countries. Glynn et al. ⁴⁰ found a nasopharyngeal mass in 55/85 patients (69%) who presented with unilateral or bilateral OME. All patients underwent an examination under anaesthesia of the ears and a nasopharyngeal biopsy. The four patients with suspicious-looking masses were all found to have malignancies (two squamous cell carcinomas, one B-cell non-Hodgkin lymphoma and one adenocarcinoma). All other patients with masses were found to have benign lymphoid hyperplasia. Finkelstein et al. ⁴¹, in another non-endemic country, found head and neck tumours, mainly NPC, in 8/167 patients with OME (4.8%). The authors concluded that nasopharyngeal and parapharyngeal space-occupying lesions should be ruled out in all cases.

In spite of these “alarm-bells”, the diagnosis of NPC is too often very delayed, mainly in Western countries. Also in Hong Kong, Lee et al. found a mean symptom duration before diagnosis of 7.4 months ³⁷. This delay was partially justifiable in the pre-flexible fibrescope era, when the visualisation of the nasopharynx with transoral small mirrors was a difficult enterprise in some patients; nowadays, it is not understandable. Lee et al. ³⁷ wrote: “This disease is often managed solely by radiotherapists and information about the disease from general textbooks is scant due to its rarity in the West. Consequently, doctors from other specialties (except otolaryngologists) are often unfamiliar with the natural course of NPC. It is not surprising therefore that very few patients present with Stage I disease”. Unfortunately, in my long experience at the Istituto Nazionale Tumori of Milan, I have seen several patients, although examined by otolaryngologists for persistent OME and cervical LN enlargement, who had been treated for a long time with antibiotics without examination of the nasopharynx. Lee et al. confirms that “the association of otological symptoms with NPC is inadequately recognised” ³⁷.

This incorrect interpretation of these symptoms is even less understandable because the levels of cervical metastases from NPC are often peculiar. Unfortunately, while many studies stressed the frequency of neck LN metastases, few analysed the topographical distribution in each level. Lee et al. ³⁷ used the generic term “upper cervical lymph node(s)”. Lindberg ⁴², in a paramount study published in 1972, reviewed the records of 2,044 patients with squamous cell carcinoma of the head and neck seen from 1948 to 1965 at the University of Texas in Houston, in order to define the preferred areas of spread for each site of the primary tumour. He divided each side of the neck into 9 nodal regions (submental, submaxillary, subdigastric, midjugular, low jugular, upper posterior cervical, mid posterior cervical, low posterior cervical, and supraclavicular). Overall, 1,155/2,044 patients (57%) presented with clinical evidence of cervical lymph node (CLN) metastasis, but the N+ patients with NPC were 147/169 (87%). Moreover, the rates of metastasis in the posterior chain (LN along the spinal accessory nerve) were very low for all other sites of the primary tumour. On the contrary, according to Lindberg “nasopharynx lesions have the highest incidence of bilateral metastases and posterior cervical chain involvement”; the uppermost nodes, which are “beneath the sternocleidomastoid muscle at the tip of the mastoid process”, are the most involved ones ⁴². Lindberg underlined another peculiarity of NPC. While “the incidence of cervical node metastasis on admission increases sharply as the size of the primary increases” in almost all sites of the head and neck, “in nasopharyngeal lesions the percentage of patients with nodal metastasis is really the same for all stages, reflecting the aggressiveness of small primaries” ⁴².

Another very large study on this topic was made by Molinari et al. ⁴³ These authors started from the opposite point of view and analysed the records of 3,100 patients with cervical metastasis seen at the Istituto Nazionale Tumori of Milan in 7 consecutive years in order to discover the probable site of the primary tumour for each involved cervical level(s). They divided the neck in levels similar to those of Lindberg’s ⁴² and Robbins’ ⁴⁴; it should be noted that Robbins’ levels were published 14 years later (Fig. 1). The authors made a series of maps in which they calculated the rates of involvement of each level(s) of the cervical lymphatic area by a tumour of a given site. They used Bayes’ formula. In biomedical statistics, it allows to establish how many probabilities there are that a particular symptom or combination of symptoms are caused by a particular illness. In this case, the symptom was adenopathy and the illness was a tumour localised in a particular site. Each oval in each cervical level is proportional to probability that it is involved by metastasis of a given tumour, and it is wider when it is frequently involved by a rare tumour (e.g. NPC). NPC was the most probable primary when there were metastases in the spinal chain, either in case of an isolated adenopathy or in case of multiple adenopathies (Figs. 2,3).

It should be noted that if the distribution of the node metastases represented in Figure 2 is bilateral, the probabilities of nasopharyngeal primary tumour are 76% and lymphoma 22%.

The peculiarity of the NPC to metastasise in the more posterior lymph nodes of the neck, demonstrated by Lindberg and Molinari, has apparently been questioned by a long series of more recent studies. Ho et al. ⁴⁵ performed a systematic review of original articles and abstracts analysing the cervical nodal metastasis status of patients with NPC by searching electronic databases PUBMED (January 1990 to December 2009), CANCERLIT (January 1990 to December 2009), and the Cochrane Library (January 1980 to July 2007). Studies were eligible if the cervical (CLN) and/or retropharyngeal node (RLN) positivity rate in NPC was reported. Search strategy included the key words “nasopharyngeal cancer”, “lymph nodes”, “nasopharyngeal carcinoma”, “lymphatic metastasis”, “cervical nodes”, and “retropharyngeal nodes”. The initial literature search resulted in 411 studies, but after excluding those who did not report the incidence of lymph node metastases, 13 original research reports and a total of 2920 NPC cases evaluated with MRI were included in the analysis. Collectively, 84.9% of NPC cases presented with regional lymphadenopathy. Metastases to neck nodes follow an orderly pattern and the probability of “skip” metastasis between regional nodes vary from 0.5% to 7.9%. The two most commonly involved regions at staging were the RLN (69.4%) and level II (70.4%). However, it must be emphasised that, with regards to level II, none of the studies reviewed made a separation between levels IIa and IIb. The importance of this distinction has recently been underlined by a long series of studies on thousands of cases. These studies have shown that in carcinomas of the oral cavity, larynx, hypopharynx, etc. the incidence of metastases at level IIb are very unlikely when there are no involved LN at level IIa ^46-48. For oral carcinoma, Chinn et al. ⁴⁶ wrote: “Recently, the extent of nodal dissection to include level IIb has been questioned. Results of several prospective studies support the exclusion of level IIb because of a 3.9% incidence of positive nodes in oral cavity carcinoma”. The conclusions of Jia et al. ⁴⁷ for supraglottic carcinoma were: “The distribution of pathologically positive lymph nodes by levels was 35 in ipsilateral level IIa; 4 in ipsilateral level III, and 8 in contralateral level IIa. There was no positive lymph node at level IIb pathologically. When selective neck dissection was carried out for patients with SCC of the supraglottic larynx with N0 neck, superselective neck dissection removing lymph nodes in levels IIa and III was adequate. Level IIb lymph node pads may be left undissected so that spinal accessory nerve dysfunction is decreased and operative time reduced”. For hypopharyngeal cancer, Sakai et al. ⁴⁸ wrote: “The results suggest that preservation of level IIb during neck dissection was possible in N0 cases of hypopharyngeal cancer”.

As for the level IIb, Grégoire et al. ⁴⁹, in their worldwide consensus guideline on classification and delineation of the node groups of the head and neck region, reflecting “in-depth discussions of a panel of European, Asian, Australian/New Zealander and North American experts from various disciplines, i.e. radiation oncologists, head and neck surgeons, and anatomists, representing the major relevant cooperative groups in radiation oncology (DAHANCA, EORTC, HKNPCSG, NCIC CTG, NCRI, RTOG, TROG)” wrote: “Level IIb is more likely associated with primary tumours of the oropharynx or nasopharynx, and less frequently with tumours of the oral cavity, larynx or hypopharynx”, and, regarding level V, “Level V lymph nodes are at high risk for harbouring metastases from cancers of the nasopharynx, oropharynx and thyroid gland”, and: “Level Vc (supraclavicular nodes) receives efferent lymphatics from the posterior triangle nodes (level Va and Vb) and is more commonly associated with nasopharyngeal tumours” ⁴⁹. Wang et al. ⁵⁰ analysed the levels involved by metastases in their series of 3,100 cases of newly diagnosed NPC between January 2010 and January 2013. A total of 2,679 (86.4%) cases had involved lymph nodes. The detailed distribution was: level IIa 1,798 (67.1%), level IIb 2,341 (87.4%)”. Jiang et al. ⁵¹ analysed the records of 960 NPCs in their institution and found that the top four levels with the highest rates of lymph node metastasis were: RLN, (86.35%), IIb (84.06%), IIa (62.29%), and III (47.29%). Regarding level V, they demonstrated that LN metastasis in the cervical region posterior to level V can occur in patients with NPC, and that 4.95% of patients with LN metastases had involvement of that region.

Both the studies by Lindberg ⁴² and Molinari et al. ⁴³ were carried out before the era of CT and MRI. Hence, they were unable to report the rates of RLN involvement. Since then, several studies demonstrated that RLNs represent the first-echelon lymph nodes typically involved in NPC ^52,53.

Distant metastasis at presentation has been reported in approximately 5% of patients ⁸. After loco-regional treatment, distant metastases are the major cause of mortality. In the series by Huang et al. ⁵⁴, 125 of 629 (19.8%) cases of NPC were found to have distant metastasis. The frequency of metastases in the various sites was as follows: bone (75% of total metastatic patients), lung (46%), liver (38%), retroperitoneal lymph nodes (10%). Multiple organ involvement was common (57%). Most of the distant metastases (95%) occurred within 3 years after completion of radiotherapy; in the first year (52%), in the second year (23%), and in the third year (20%). QU et al. stated that “Tumours are highly likely to metastasize if they have a larger volume, the regional LNs are relatively large, or the regional LNs are biopsied but not removed” ⁵⁵.

Diagnostic workup

According to the recent ESMO-EUROCAN guidelines for the NPC, routine staging procedures include a medical history, physical examination (including cranial nerve examination), complete blood count (CBC), serum biochemistry [including liver and renal function tests and lactate dehydrogenase (LDH)], nasopharyngoscopy, CT scan or MRI of the nasopharynx and base of the skull and neck (up to the clavicle) and 18F-fluorodeoxy-glucose (FDG)-PET/CT imaging ⁵⁶ (Tab. V). Medical history and physical examination are even more suitable for NPC because the diagnosis of this tumour is late in most cases, either for the heedlessness by the patient or for the scant awareness of physicians, who misinterpret the related symptoms. In the 1970s, when I was a trainee at the Istituto Nazionale Tumori of Milan, most patients presented after a cervical LN biopsy with the diagnosis of metastasis from undifferentiated carcinoma of unknown primary origin, without any examination of the nasopharynx. As mentioned above, it is worth noting that in the pre-fibrescope era the visualisation of the nasopharynx was a difficult enterprise in some patients. Fortunately, nowadays the flexible fibrescopes have changed the scenario, making possible both thorough examination of the nasopharynx and a biopsy. Nevertheless, even in more recent years, too many patients presenting with symptoms revealing a possible NPC underwent a cervical node biopsy before careful inspection of the nasopharynx. When the primary tumour has already been histologically diagnosed, a lymph node biopsy is nearly always unnecessary. However, a lymphadenectomy may rarely be requested by the pathologist to clarify possible diagnostic doubts between NPC and lymphoma.

According to the European Society for Medical Oncology (ESMO) guidelines “In terms of imaging, a contrast enhanced CT and/or MRI from the skull base to the clavicles is recommended” ⁵⁶. CT has long been used for staging NPC, especially for the detection of skull base involvement with lytic or sclerotic lesions, but it has been now largely replaced by MRI for primary and nodal staging ⁵⁷. MRI has been shown to be superior to CT in the delineation of extent of perineural spread along the skull base foramina, intracranial involvement, retropharyngeal nodal disease, and invasion of adjacent deep neck spaces ^56,58.

Moreover, two studies from the same centre demonstrated that MRI imaging depicts subclinical cancers missed at endoscopy and endoscopic biopsy, and helps identify the majority of patients who do not have NPC and therefore do not need to undergo invasive sampling biopsies ^59,60. However, CT is still used for radiotherapy planning and, in some centres, together with positron emission tomography (PET). 18F-fluorodeoxyglucose PET/CT has been shown to have a high diagnostic accuracy for detection of distant metastases and is superior to CT alone ⁶¹. It is also used to monitor patients after therapy and to detect NPC recurrence. A thorough synthesis have been made by a study from Taiwan ⁶². These authors demonstrated that “MRI appears to be superior to PET/CT for the assessment of locoregional invasion and RLNs. PET/CT is more accurate than MRI for determining cervical nodal metastasis and should be the better reference for the neck status. PET/CT has an acceptable diagnostic yield and a low false-positive rate for the detection of distant malignancy and can replace conventional work-up to this aim”.

The NCCN guidelines ⁶³ also suggest performing EBV DNA testing. “For non-keratinising or undifferentiated histology, consider testing of EBV in tumour and blood. Common means for detecting EBV in pathologic specimens include in situ hybridisation for EBER or immunohistochemical staining for latent membrane protein (LMP). The EBV DNA load within the serum or plasma may be quantified using polymerase chain reaction (PCR) targeting genomic sequences of the EBV DNA such as BamHI-W, Epstein-Barr nuclear antigen (EBNA), or LMP; these tests vary in their sensitivity. The EBV DNA load may reflect prognosis and change in response to therapy”. The possibility to detect the presence of EBV genomes in fine needle aspirate (FNA) from neck lymph node metastases was tested in non-endemic and endemic countries ^64,65. Both these studies demonstrated the utility of EBV gene amplification in FNA samples of neck metastases and suggested that the presence of the EBV genome in FNA samples of neck nodes is predictive of the presence of NPC.

Circulating EBV DNA analysis has been shown to be valuable in the detection, prognostication and monitoring of NPC patients. Lo et al. ⁶⁶ studied the kinetics of circulating EBV DNA in the plasma of NPC patients undergoing radiation therapy and demonstrated a rapid decline in plasma EBV DNA concentration after treatment with a median half-life of 3.8 days. They concluded that kinetic analysis of circulating tumour-derived DNA during treatment may be a powerful tool to evaluate the in vivo response of NPC to antineoplastic treatment. To et al. ⁶⁷ analysed plasma EBV DNA concentrations in 21 patients with either local recurrent NPC (17 cases) or disease persistence (4 cases). All patients underwent salvage surgery. The in vivo elimination of EBV DNA was very rapid after surgical resection. The median half-life of plasma EBV DNA was 139 min. The failure of complete and rapid elimination of EBV DNA from circulation predicted disease recurrence.

While circulating cancer-derived EBV DNA in plasma has been well established as a tumour marker in patients with diagnosed NPC for prognostication, predicting treatment response, and disease surveillance, Chan et al. stated that “there is much less information on the use of analysis of circulating DNA to screen for early cancers. A fundamental question is whether a small tumour would release sufficient amounts of tumour DNA into the circulation to allow sensitive detection of the cancer-associated changes” ⁶⁸. A paramount prospective study published in the New England Journal of Medicine investigated this topic in an endemic country (Hong Kong). Chan et al. ⁶⁸ analysed EBV DNA in plasma specimens of 20,174 men between the ages of 40 and 62 years who did not have symptoms of NPC. Participants with initially positive results were retested approximately 4 weeks later, and those with persistently positive EBV DNA in plasma underwent nasal endoscopic examination and MRI. EBV DNA was detectable in plasma samples obtained from 1,112 participants (5.5%), and 309 (1.5% of all participants and 27.8% of those who initially tested positive) had persistently positive results on the repeated sample. Among these 309 participants, 300 underwent endoscopic examination, and 275 underwent both endoscopic examination and MRI; of these participants, 34 had NPC. The proportion of participants with NPC identified by screening with stage I or II disease was significantly higher than in a historical cohort (71% vs 20%, p _< 0.001 by the chi-square test). Moreover, these patients had superior 3-year progression-free survival (97% vs 70%). Nine participants declined to undergo further testing, and one presented with advanced NPC at 32 months after enrolment. NPC developed in only one participant with negative EBV DNA in plasma samples within one year after testing. The sensitivity and specificity of EBV DNA in plasma samples in screening for NPC were 97.1% and 98.6%, respectively. The costs for each EBV DNA analysis, endoscopic examination, and MRI were $30, $80, and $1,000 (U.S. dollars), respectively. The authors concluded: “Considering the potential decrease in mortality and morbidity, as well as treatment cost savings associated with the shift in stage distribution, screening for NPC appears to be a feasible practice in regions with a high incidence of this disease” ⁶⁸. A subsequent study examined the cost-effectiveness of screening for NPC with plasma EBV DNA among 50-year-old Asian American men in the United States, and concluded that, although screening for NPC with EBV DNA for 50-year-old Asian American men may result in earlier detection, it was unlikely to be cost-effective ⁶⁹.

Narrow-band imaging

To my knowledge, the first reports of a new endoscopic imaging technique, the narrow-band imaging (NBI), appeared about 20 years ago ⁷⁰. NBI is a diagnostic tool that enhances the endoscopic view of the upper aero-digestive tract (UADT) to reveal small mucosal lesions that would otherwise escape diagnosis even by experienced endoscopists ⁷¹. This technique, which has previously been used for the detection of adenomas in the gastrointestinal tract, has the potential to reduce the false negative rates associated with conventional white light endoscopy. NBI is an imaging technique that uses two specific wavelengths of light that are strongly absorbed by haemoglobin, allowing improved visualisation and delineation of mucosal microvascular patterns ⁷². “Thus, superficial mucosal lesions that would previously have been missed by regular white light during endoscopy would be identified by the blue light of NBI, based on the increased vascularity and neoangiogenesis of the tumour” ⁷¹. In 2008, the literature on the use of NBI in head and neck oncology was very sparse ⁷¹. Today, hundreds of studies have been published on this topic in the diagnostic work-up of head and neck cancers, especially those at oropharyngeal and hypopharyngeal sites. As for the nasopharynx, the number of studies is smaller but some studies are very indicative for the high number of patients examined. Most of the studies were done in endemic areas where NBI was used as screening. For example, Wen et al. ⁷³ compared the diagnostic accuracy of conventional white-light (WL) imaging endoscopy with that of NBI endoscopy in patients at high risk for NPC and found that the sensitivity and negative predictive values of NBI in NPC screening were significantly higher than those of WL (93.9% vs 71.2%, p = 0.001; and 98.1% vs 91.7%, p = 0.003; respectively). Yang et al. ⁷⁴ examined a total of 1,854 patients by means of an electronic nasopharyngolaryngoscope equipped with conventional WL and NBI system. Of these, 62 cases (3.34%) were pathologically confirmed as NPC. The sensitivity, specificity, positive predictive value and negative predictive value for detecting NPC significantly increased from 90.3%, 75.4%, 11.3% and 99.6% with WL up to 100%, 99.2%, 81.6% and 100% with NBI, respectively. The authors concluded that these findings suggested that NBI endoscopy might serve as an ideal tool in the detection of NPC. In this regard, it should be emphasised that NBI can be useful to locate the primary tumour in the case of cervical metastases from an unknown primary.

However, it should be noted that, among most of the conclusions in favour of NBI, there are also some that are against it. In the study by Vlantis et al. ⁷⁵ “The nasopharynx of 156 patients who failed serological screening for or presented with symptoms of nasopharyngeal carcinoma was graded under WL and NBI endoscopy and a biopsy taken”. The authors concluded that “Narrow band imaging endoscopy of vasculature alone for suspected nasopharyngeal carcinoma is not more useful than white light endoscopy of nasopharyngeal morphology, nor does it add to or surpass the diagnostic accuracy of white light endoscopy in this regard”. Another study from the same centre demonstrated that: “Intraobserver and interobserver agreement for nasopharyngeal malignancy was strong on WL endoscopy but only moderate on NBI endoscopy. Agreement may be improved by adopting a standard set of assessment guidelines, including an objective detailed morphological analysis under WL and vasculature analysis under NBI” ⁷⁶.

Radiomics

The most recent tool for the diagnosis and staging of NPC is the so-called ‘radiomics’. Radiomics, which was first proposed by Lambin ⁷⁷ in 2012, is a relatively ‘young’ concept and is considered a natural extension of computer-aided diagnosis and detection systems. It converts imaging data into a high-dimensional mineable feature space using a large number of automatically extracted data-characterisation algorithms to reveal tumour features that may not be recognised by the naked eye and to quantitatively describe the tumour phenotype. These extracted features are called radiomic features and include first-order statistics features, intensity histograms, shape- and size-based features, texture-based features and wavelet features. Conceptually speaking, radiomics belongs to the field of machine learning, although human participation is needed. The basic hypothesis of radiomics is that the constructed descriptive model (based on medical imaging data, sometimes supplemented by biological and/or medical data) can provide predictions of prognosis or diagnosis ⁷⁸.

A complete and in-depth review of the characteristics and possibilities of radiomics can be found in the article by Scapicchio et al. ⁷⁹: “Diagnostic imaging is going through an epoch-making moment of profound transformation, to which radiologists must adapt. This is a transformation from a discipline based on the visual interpretation of the images toward a new type of radiology, which must integrate the quantitative data (biomarkers) coming from the images with the interpretative modality. In fact, since are formed by the interaction of radiation or ultrasounds with tissues or organs, medical images are not simple images, but they reflect various physical properties of the body. Medical images can be converted into meaningful and mineable data through a quantification process. The extracted quantitative features can be analysed to reflect the underlying pathophysiology. However, quantitative data are not easily interpretable by the human mind, they can only be extracted from a computer and analysed through complex algorithms. Despite promising results in research, such applications of radiomics still necessitate a deep exploration, refinement, standardisation and validation to achieve routine clinical adoption, but they may be of great help in the clinical management of specific diseases in the near future. Radiomics is a piece in the puzzle of precision medicine, its final goal being to build models able to classify the disease and/or to predict its outcome or the answer to a therapy” ⁷⁹. A meta-analysis carried out by Park et al. ⁸⁰ found that 91% of radiomics studies concern oncological applications, and that for the most part (81%) radiomics is studied for diagnostic purposes. They stated that “To date, radiomics has been applied in various oncology research (mainly brain, prostate, breast and lung tumours) for precise clinical decision, treatment response and prognosis predication.” “The evolving branch of radiomics linking imaging features to gene expression is today known as radiogenomics. By combining quantitative imaging features with clinical, genomic and other information in a multi-omics study, it is possible to mine these data to detect and validate radiomics biomarkers and better understand their function and biological significance” ⁸⁰.

For NPC, Gao et al. wrote: “In current clinical practice, the TNM staging system is the most used prognostic tool based on tumour size, nodes, and distant metastasis. However, patients with the same clinical stage display totally different therapeutic response and clinical outcome. Hence, accumulated omic-studies including genomics, proteomics, and immunomics have been applied to deepen our understanding on tumour heterogeneity and malignant progression of NPC, seeking for potential prognostic biomarkers and therapeutic molecular targets” ⁸¹. Li et al. wrote: “Although radiomics has been applied to medical imaging since 2012 and DL (deep learning) has started to be applied to medical imaging around 2015, their application in NPC has only gradually begun in 2017. Many attempts have been made to apply AI (artificial intelligence) tools for NPC imaging in clinical settings. However, there are still some limitations in this field” ⁷⁸. Spadarella et al. performed a systematic search in PubMed, Web of Science and Scopus using “MRI, magnetic resonance imaging, radiomic, texture analysis, nasopharyngeal carcinoma, nasopharyngeal cancer” in all possible combinations. They found 24 studies and evaluated them according to the Radiomic quality score. They concluded that: “Radiomic articles in NPC are mostly of low methodological quality. The greatest limitations are the lack of external validation, biological correlates, prospective design and open science” ⁸².

Staging

As many as 18 different classifications of NPC have been developed since the first proposed by Geist and Portman (1952) ⁸³. However, only two have been widely used, that of the American Joint Committee on Cancer-International Union Against Cancer (AJCC-UICC) staging system and the Chinese one (Tab. VI). China developed independent criteria for clinical staging of NPC since 1959 and the first TNM staging system was established in Shanghai in 1965. It was revised in 1979, 1981, 1992 and 2008 ⁸⁴. Since 1979, the Chinese classification used levels and fixity of LN metastases for staging, and divided the lower LNs from those in the supraclavicular fossa (SCF). N2 stage was: “Upper cervical LNs, mobile or fixed 3-8 cm in greatest dimension, lower cervical LNs, mobile, > 8 cm in greatest dimension”. N3 stage was: “Lower cervical LNs, fixed, > 8 cm in greatest dimension, or extension to the SCF, or supraclavicular skin nodule” ⁸⁴. In 1990, from the analysis of a series of 410 cases treated at the Istituto Nazionale Tumori of Milan the authors found that: “With regards to nodal extent, only the level of the involved nodes was a significant variable. Both AJCC and UICC classifications, when applied to the entire series of patients, appeared to be unsatisfactory” ⁸⁵. It is important to note that this study used the AJCC and UICC classifications in force in those years (4^th edition) for staging, both of which did not provide for the subdivision of lymph node metastases into levels. In 1991, a study from Hong Kong analysed 564 patients with NPC and compared the 1979 Chinese classification, the UICC, and the AJCC classifications, for their efficacy in predicting prognosis. Teo et al. stated that the Chinese classification was superior to the other two “because its overall stages differed from one another more significantly in the actuarial survival, disease-free survival (DFS), and freedom from distant metastasis (FDM) rates, and its N staging was more accurate in predicting FDM” ⁸⁶.

The AJCC and UICC classifications for NPC were different from each other in the first four editions, both for T and N. Moreover, the definition of N categories was the same for all head and neck sites in both classifications. The TNM Staging System 5^th edition, jointly adopted by the AJCC and the UICC in 1997, was a milestone. Regarding N classification, both systems reported the following statement: “The distribution and the prognostic impact of regional LNs spread from nasopharynx cancer, particularly of the undifferentiated type, is different than that of other head and neck mucosal cancers and justifies use of a different N classification scheme”. The cut-off between N1-N2 and N3 was the level above the SCF or extension to it. The 7^th edition (2010) introduced an important new parameter for LN metastases, namely unilateral or bilateral involvement of RLNs. The parameter of RLNs remained in the 8^th edition (2017), even if the term “supraclavicular fossa” was changed into “above or below the caudal border of cricoid cartilage”. Therefore, patients with a tumour T1 and with only metastases to the RLNs were classified “N0” till 2010 and “N1” after then, which has an important consequence for the therapy. Actually, according to NCCN guidelines, the therapeutic indication for T1N0 is radiotherapy alone (RT), while for T1N1 is chemo-radiotherapy (CT-RT). In 2015, a total of 816 patients with untreated NPC who underwent magnetic resonance imaging of the nasopharynx and neck were studied retrospectively and restaged according to the Chinese 2008 staging system and the AJCC staging system. The AJCC T classification was better in predicting the 5-year local relapse-free survival, whereas the Chinese 2008 N classification was superior in predicting the 5-year distant metastasis-free survival. However, 5-year overall survival was comparable in both systems ⁸⁷.

Some studies analysed the connection between pre-treatment plasma EBV DNA load and prognosis and demonstrated that patients in which the tumour tissues harbour EBV have a better prognosis than those without EBV-related NPC ⁸⁸. For this reason, a recent study proposed a new staging system with the incorporation of the measurement of plasma EBV DNA in the 8^th AJCC-UICC classification. The authors demonstrated that the proposed stage groupings have better prognostic performance than the 8^th edition of the TNM staging system ⁸⁹.

At the conclusion of this first part of the review, it is appropriate to point out that the underestimation and misinterpretation of the first mild symptoms of a possible NPC have serious consequences for the patient, and reduce the chances of cure.

Conflict of interest statement

The author declares no conflict of interest.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Figures and tables

Figure 1.Levels of the lymph nodes of the neck according to Molinari et al. 1. Submental region; 2. Submandibular region; 3. Subdigastric region; 4. Middle jugular region; 5. Low jugular region; 6. Upper spinal region; 7. Middle spinal region; 8. Low spinal region; 9. Lateral supraclavicular region (from Molinari et al., 1977, mod.) ⁴³.

Figure 2.Probability of NPC with this type of multiple lymph node metastases. Monolateral: nasopharynx 59%; lymphoma 22%; tonsil 16%; other 3%. Bilateral: nasopharynx 76%; lymphoma 22%; other 2%. Each oval in each cervical level is proportional to probability that it is involved by metastases of a given tumour, and it is wider when it is frequently involved by a rare tumour (e.g. NPC) (from Molinari et al., 1977, mod.) ⁴³.

Figure 3.Probability of NPC with this type of isolated adenopathy. Nasopharynx 58%; tonsil 9.6%; oropharynx 8.5%; base of tongue 7.7 %; hypopharynx 7.7%; larynx 4.6%; retromolar trigone 2.9%; other 1% (from Molinari et al., 1977, mod.) ⁴³.

References

1. Pathology and Genetics of Head and Neck Tumours. IARC Press: Lyon; 2005.
2. Regaud C, Reverchon L. Sur un cas d’epithelioma epidermoide developé dans le massif maxillaire supérieur, et étendu aux téguments de la face, aux cavités buccale, nasale et orbitaire, ainsi qu’aux ganglions du cou. Rev Laryngol Otol Rhinol. 1921; 42:369.
3. Schmincke A. Uber limpho-epitheliale geschwulste. Beitr Pathol Anat Allg Pathol. 1921; 68:161.
4. Scanlon PW, Rhodes RE, Woolner LB. Cancer of the nasopharynx: 142 patients treated in the 11 year period 1950-1960. Am J Roentgenol Radium Ther Nucl Med. 1967; 99:313-225.
5. Shanmugaratnam K, Sobin LH. Histological typing of upper respiratory tract tumours. WHO: Geneva; 1978.
6. Shanmugaratnam K. Histological typing of nasopharyngeal carcinoma. IARC Sci Publ. 1978; 20:3-12.
7. Shanmugaratnam K. Histological typing of tumours of the upper respiratory tract and ear. Springer-Verlag: Berlin; 1991.
8. WHO Classification of head and neck tumours. IARC: Lyon; 2017.
9. Palazzi M, Guzzo M, Tomatis S. Improved outcome of nasopharyngeal carcinoma treated with conventional radiotherapy. Int J Radiat Oncol Biol Phys. 2004; 60:1451-1458. DOI
10. Wang Y, Zhang Y, Ma S. Racial differences in nasopharyngeal carcinoma in the United States. Cancer Epidemiol. 2013; 37:793-802. DOI
11. Stepan KO, Mazul AL, Skillington SA. The prognostic significance of race in nasopharyngeal carcinoma by histological subtype. Head Neck. 2021; 43:1797-1811. DOI
12. Wu SG, Lian CL, Wang J. The effect of histological subtypes on survival outcome in nasopharyngeal carcinoma after extensive follow-up. Ann Transl Med. 2019; 7:768. DOI
13. Batsakis JG. Clinical and pathological considerations. Williams &amp; Wilkins: Baltimore; 1979.
14. Buell P. Nasopharynx cancer in chinese of California. Br J Cancer. 1965; 19:459-470. DOI
15. Dickson RI. Nasopharyngeal carcinoma: an evaluation of 209 patients. Laryngoscope. 1981; 91:333-354. DOI
16. Chang ET, Adami HO. The enigmatic epidemiology of nasopharyngeal carcinoma. Cancer Epidemiol Biomarkers Prev. 2006; 15:1765-1777. DOI
17. Old LJ, Boyse EA, Oettgen HF. Precipitating antibody in human serum to an antigen present in cultured Burkitt’s lymphoma cells. Proc Natl Acad Sci USA. 1966; 56:1699-1704. DOI
18. Henle W, Henle G. Evidence for an oncogenic potential of the Epstein-Barr virus. Cancer Res. 1973; 33:1419-1423.
19. Klein G. The Epstein-Barr virus and neoplasia. N Engl J Med. 1975; 293:1353-1357. DOI
20. Nadol JB. Viral particles in nasopharyngeal carcinoma. Laryngoscope. 1977; 87:1932-1937. DOI
21. Ziegler JL, Magrath IT, Gerber P. Epstein-Barr virus and human malignancy. Ann Intern Med. 1977; 86:323-336. DOI
22. Kangro HO, Osman HK, Lau YL. Seroprevalence of antibodies to human herpesviruses in England and Hong Kong. J Med Virol. 1994; 43:91-96. DOI
23. Dogan S, Hedberg ML, Ferris RL. Human papillomavirus and Epstein-Barr virus in nasopharyngeal carcinoma in a low-incidence population. Head Neck. 2014; 36:511-516. DOI
24. Verma V, Simone CB, Lin C. Human papillomavirus and nasopharyngeal cancer. Head Neck. 2018; 40:696-706. DOI
25. Wu SS, Chen B, Fleming CW. Nasopharyngeal cancer: Incidence and prognosis of human papillomavirus and Epstein-Barr virus association at a single North American institution. Head Neck. 2022; 44:851-861. DOI
26. Yu MC, Nichols PW, Zou XN. Induction of malignant nasal cavity tumours in Wistar rats fed Chinese salted fish. Br J Cancer. 1989; 60:198-201. DOI
27. Zheng X, Luo Y, Christensson B. Induction of nasal and nasopharyngeal tumours in Sprague-Dawley rats fed with Chinese salted fish. Acta Otolaryngol. 1994; 114:98-104. DOI
28. Zou XN, Lu SH, Liu B. Volatile N-nitrosamines and their precursors in Chinese salted fish – a possible etiological factor for NPC in china. Int J Cancer. 1994; 59:155-158. DOI
29. Lin C, Cao SM, Chang ET. Chinese nonmedicinal herbal diet and risk of nasopharyngeal carcinoma: a population-based case-control study. Cancer. 2019; 125:4462-4470. DOI
30. Roy Chattopadhyay N, Das P, Chatterjee K, Choudhuri T. Higher incidence of nasopharyngeal carcinoma in some regions in the world confers for interplay between genetic factors and external stimuli. Drug Discov Ther. 2017; 11:170-180. DOI
31. Casanova M, Bisogno G, Gandola L. Rare tumors in pediatric age group. A prospective protocol for nasopharyngeal carcinoma in children and adolescents: the Italian Rare Tumors in Pediatric Age (TREP) project. Cancer. 2012; 118:2718-2725. DOI
32. Sultan I, Casanova M, Ferrari A. Differential features of nasopharyngeal carcinoma in children and adults: a SEER study. Pediatr Blood Cancer. 2010; 55:279-284. DOI
33. Lee AW, Foo W, Mang O. Changing epidemiology of nasopharyngeal carcinoma in Hong Kong over a 20-year period (1980-99): an encouraging reduction in both incidence and mortality. Int J Cancer. 2003; 103:680-685. DOI
34. Mudry A. Pharyngeal recess or Rosenmüller’s fossa: its first description revisited. Head Neck. 2021; 43:2295-2296. DOI
35. Chong VF, Fan YF, Khoo JB. Nasopharyngeal carcinoma with intracranial spread: CT and MR characteristics. J Comput Assist Tomogr. 1996; 20:563-569. DOI
36. Pak MW, To KF, Lo YM. Nasopharyngeal carcinoma in situ (NPCIS) – pathologic and clinical perspectives. Head Neck. 2002; 24:989-995. DOI
37. Lee AW, Foo W, Law SC. Nasopharyngeal carcinoma: presenting symptoms and duration before diagnosis. Hong Kong Med J. 1997; 3:355-361.
38. Lederman M. Cancer Of The nasopharynx. Proc R Soc Med. 1964; 57:740-742.
39. Wang KH, Austin SA, Chen SH. Nasopharyngeal carcinoma diagnostic challenge in a nonendemic setting: our experience with 101 patients. Perm J. 2017; 21:16-180. DOI
40. Glynn F, Keogh IJ, Ali TA. Routine nasopharyngeal biopsy in adults presenting with isolated serous otitis media: is it justified?. J Laryngol Otol. 2006; 120:439-441. DOI
41. Finkelstein Y, Ophir D, Talmi YP. Adult-onset otitis media with effusion. Arch Otolaryngol Head Neck Surg. 1994; 120:517-527. DOI
42. Lindberg R. Distribution of cervical lymph node metastases from squamous cell carcinoma of the upper respiratory and digestive tracts. Cancer. 1972; 29:1446-1449. DOI
43. Molinari R, Cantù G, Chiesa F. A statistical approach to detection of the primary cancer based on the site of neck lymph node metastases. Tumori. 1977; 63:267-282.
44. Robbins KT, Medina JE, Wolfe GT. Standardizing neck dissection terminology. Official report of the Academy’s Committee for Head and Neck Surgery and Oncology. Arch Otolaryngol Head Neck Surg. 1991; 117:601-605. DOI
45. Ho FC, Tham IW, Earnest A. Patterns of regional lymph node metastasis of nasopharyngeal carcinoma: a meta-analysis of clinical evidence. BMC Cancer. 2012; 12:98. DOI
46. Chinn SB, Myers JN. Oral cavity carcinoma: current management, controversies, and future directions. J Clin Oncol. 2015; 33:3269-3276. DOI
47. Jia S, Wang Y, He H. Incidence of level IIB lymph node metastasis in supraglottic laryngeal squamous cell carcinoma with clinically negative neck – a prospective study. Head Neck. 2013; 35:987-991. DOI
48. Sakai A, Okami K, Sugimoto R. Evaluating the significance of level IIb neck dissection for hypopharyngeal cancer. Head Neck. 2013; 35:1777-1780. DOI
49. Grégoire V, Ang K, Budach W. Delineation of the neck node levels for head and neck tumors: a 2013 update. DAHANCA, EORTC, HKNPCSG, NCIC CTG, NCRI, RTOG, TROG consensus guidelines. Radiother Oncol. 2014; 110:172-181. DOI
50. Wang X, Hu C, Ying H. Patterns of lymph node metastasis from nasopharyngeal carcinoma based on the 2013 updated consensus guidelines for neck node levels. Radiother Oncol. 2015; 115:41-45. DOI
51. Jiang C, Gong B, Gao H. Correlation analysis of neck node levels in 960 cases of Nasopharyngeal carcinoma (NPC). Radiother Oncol. 2021; 161:23-28. DOI
52. Zhang GY, Liu LZ, Wei WH. Radiologic criteria of retropharyngeal lymph node metastasis in nasopharyngeal carcinoma treated with radiation therapy. Radiology. 2010; 255:605-612. DOI
53. Pan XB, Huang ST, Qu S. Retropharyngeal lymph node metastasis on N stage of nasopharyngeal carcinoma. PLoS One. 2021; 16:e0253424. DOI
54. Huang CJ, Leung SW, Lian SL. Patterns of distant metastases in nasopharyngeal carcinoma. Kaohsiung J Med Sci. 1996; 12:229-234.
55. Qu W, Li S, Zhang M. Pattern and prognosis of distant metastases in nasopharyngeal carcinoma: a large-population retrospective analysis. Cancer Med. 2020; 9:6147-6158. DOI
56. Bossi P, Chan AT, Licitra L, ESMO Guidelines Committee. Nasopharyngeal carcinoma: ESMO-EURACAN Clinical Practice Guidelines for diagnosis, treatment and follow-up†. Ann Oncol. 2021; 32:452-465. DOI
57. Abdel Khalek Abdel Razek A, King A. MRI and CT of nasopharyngeal carcinoma. AJR Am J Roentgenol. 2012; 198:11-8. DOI
58. Chen J, Luo J, He X. Evaluation of Contrast-Enhanced Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) in the detection of retropharyngeal lymph node metastases in nasopharyngeal carcinoma patients. Cancer Manag Res. 2020; 12:1733-1739. DOI
59. King AD, Vlantis AC, Bhatia KS. Primary nasopharyngeal carcinoma: diagnostic accuracy of MR imaging versus that of endoscopy and endoscopic biopsy. Radiology. 2011; 258:531-537. DOI
60. King AD, Vlantis AC, Yuen TW. Detection of nasopharyngeal carcinoma by MR imaging: diagnostic accuracy of MRI compared with endoscopy and endoscopic biopsy based on long-term follow-up. AJNR Am J Neuroradiol. 2015; 36:2380-2385. DOI
61. Xu C, Zhang Y, Peng L. Optimal modality for detecting distant metastasis in primary nasopharyngeal carcinoma during initial staging: a systemic review and meta-analysis of 1774 patients. J Cancer. 2017; 8:1238-1248. DOI
62. Ng SH, Chan SC, Yen TC. Staging of untreated nasopharyngeal carcinoma with PET/CT: comparison with conventional imaging work-up. Eur J Nucl Med Mol Imaging. 2009; 36:12-22. DOI
63. National Comprehensive Cancer Network. Head and Neck Cancers.Publisher Full Text
64. Macdonald MR, Freeman JL, Hui MF. Role of Epstein-Barr virus in fine-needle aspirates of metastatic neck nodes in the diagnosis of nasopharyngeal carcinoma. Head Neck. 1995; 17:487-493. DOI
65. Lee WY, Hsiao JR, Jin YT. Epstein-Barr virus detection in neck metastases by in-situ hybridization in fine-needle aspiration cytologic studies: an aid for differentiating the primary site. Head Neck. 2000; 22:336-340. DOI
66. Lo YM, Leung SF, Chan LY. Kinetics of plasma Epstein-Barr virus DNA during radiation therapy for nasopharyngeal carcinoma. Cancer Res. 2000; 60:2351-2355.
67. To EW, Chan KC, Leung SF. Rapid clearance of plasma Epstein-Barr virus DNA after surgical treatment of nasopharyngeal carcinoma. Clin Cancer Res. 2003; 9:3254-3259.
68. Chan KCA, Woo JKS, King A. Analysis of plasma epstein-barr virus dna to screen for nasopharyngeal cancer. N Engl J Med. 2017; 377:513-522. DOI
69. Harris JP, Saraswathula A, Kaplun B. Cost-effectiveness of screening for nasopharyngeal carcinoma among Asian American Men in the United States. Otolaryngol Head Neck Surg. 2019; 161:82-90. DOI
70. Gono K, Yamazaki K, Doguchi N. Endoscopic observation of tissue by narrowband illumination. OPT REV. 2003; 10:211-215. DOI
71. Piazza C, Dessouky O, Peretti G. Narrow-band imaging: a new tool for evaluation of head and neck squamous cell carcinomas. Review of the literature. Acta Otorhinolaryngol Ital. 2008; 28:49-54.
72. Yeung DC, Vlantis AC, Wong EW. A meta-analysis of narrow-band imaging for the diagnosis of primary nasopharyngeal carcinoma. F1000Res. 2018; 7:759. DOI
73. Wen YH, Zhu XL, Lei WB. Narrow-band imaging: a novel screening tool for early nasopharyngeal carcinoma. Arch Otolaryngol Head Neck Surg. 2012; 138:183-188. DOI
74. Yang H, Zheng Y, Chen Q. The diagnostic value of narrow-band imaging for the detection of nasopharyngeal carcinoma. ORL J Otorhinolaryngol Relat Spec. 2012; 74:235-239. DOI
75. Vlantis AC, Woo JK, Tong MC. Narrow band imaging endoscopy of the nasopharynx is not more useful than white light endoscopy for suspected nasopharyngeal carcinoma. Eur Arch Otorhinolaryngol. 2016; 273:3363-3369. DOI
76. Vlantis AC, Wong EWY, Ng SK. Narrow band imaging endoscopy of the nasopharynx for malignancy: an inter- and intraobserver study. Laryngoscope. 2019; 129:1374-1379. DOI
77. Lambin P, Rios-Velazquez E, Leijenaar R. Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer. 2012; 48:441-446. DOI
78. Li S, Deng YQ, Zhu ZL. A comprehensive review on radiomics and deep learning for nasopharyngeal carcinoma imaging. Diagnostics (Basel). 2021; 11:1523. DOI
79. Scapicchio C, Gabelloni M, Barucci A. A deep look into radiomics. Radiol Med. 2021; 126:1296-1311. DOI
80. Park JE, Kim D, Kim HS. Quality of science and reporting of radiomics in oncologic studies: room for improvement according to radiomics quality score and TRIPOD statement. Eur Radiol. 2020; 30:523-536. DOI
81. Gao Y, Mao Y, Lu S. Magnetic resonance imaging-based radiogenomics analysis for predicting prognosis and gene expression profile in advanced nasopharyngeal carcinoma. Head Neck. 2021; 43:3730-3742. DOI
82. Spadarella G, Calareso G, Garanzini E. MRI based radiomics in nasopharyngeal cancer: Systematic review and perspectives using radiomic quality score (RQS) assessment. Eur J Radiol. 2021; 140:109744. DOI
83. Lee AW, Foo W, Law SC. Staging of nasopharyngeal carcinoma: from Ho’s to the new UICC system. Int J Cancer. 1999; 84:179-187. DOI
84. Zong J, Huang Q, Guo Q. Evolution of the Chinese staging system for nasopharyngeal carcinoma. Chin Clin Oncol. 2016; 5:19. DOI
85. Grandi C, Boracchi P, Mezzanotte G. Analysis of prognostic factors and proposal of a new classification for nasopharyngeal cancer. Head Neck. 1990; 12:31-40. DOI
86. Teo PM, Leung SF, Yu P. A comparison of Ho’s, International Union Against Cancer, and American Joint Committee stage classifications for nasopharyngeal carcinoma. Cancer. 1991; 67:434-439. DOI
87. Pan J, Xu Y, Qiu S. A comparison between the Chinese 2008 and the 7th edition AJCC staging systems for nasopharyngeal carcinoma. Am J Clin Oncol. 2015; 38:189-196. DOI
88. Ma BB, King A, Lo YM. Relationship between pretreatment level of plasma Epstein-Barr virus DNA, tumor burden, and metabolic activity in advanced nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys. 2006; 66:714-720. DOI
89. Guo R, Tang LL, Mao YP. Proposed modifications and incorporation of plasma Epstein-Barr virus DNA improve the TNM staging system for Epstein-Barr virus-related nasopharyngeal carcinoma. Cancer. 2019; 125:79-89. DOI