Of mice and men: olfactory neuroblastoma among animals and humans



Olfactory neuroblastoma (ONB) is a rare tumour of nasal cavity and paranasal sinuses that arises from the olfactory neuroepithelium and has unpredictable clinical course. As the sense of smell is phylogenetically one of the first senses and olfactory neuroepithelium is evolutionary conserved with striking similarities among different species, we performed an extensive analysis of the literature in order to evaluate the similarities and differences between animals and humans on the clinical, morphological, immunohistochemical, ultrastructural and molecular level. Our analysis revealed that ONB was reported mainly in mammals and showed striking similarities to human ONB. These observations provide rationale for introduction of therapy modalities used in humans into the veterinary medicine. Animal models of neuroblastoma should be considered for the preclinical studies evaluating novel therapies for ONB.

Olfactory neuroblastoma

Olfactory neuroblastoma (ONB), also known as esthesioneuroblastoma, was originally described by French physicians Berger and Luca in 1924.[1] ONB is a relatively uncommon malignant neuroectodermal tumour involving the upper part of the nasal cavity, including the superior concha, the upper part of septum, the roof of nose and the cribriform plate of ethmoid. It accounts for 3% of all intranasal tumours[2] and its incidence is reported at 4 cases per 10 million inhabitants.[3] Since 1924, over 1000 cases in humans have been described in literature.[2] Nevertheless, due to its rarity and complex nature, origin and aetiology of the tumour still remains unclear.

Some indirect evidences suggest that ONB can originate from olfactory stem cells. Mice, in which the SV40T oncogene was inserted under control of the olfactory marker protein gene promoter region, did not develop ONB but adrenal and sympathetic ganglia neuroblastoma instead.[4] Therefore, currently available evidence connects ONB with the basal progenitor cells of the olfactory epithelium. Some other authors have proposed that ONB belonged to the Ewing sarcoma family[5, 6] or the primitive neuroectodermal tumour (PNET).[7] This assumption was based on the identification of translocation t(11;22) in some cases. This alteration is regarded as specific for Ewing sarcoma/PNET.[5] However, further molecular studies[8-10] have not proven this translocation in ONB. Therefore, ONB is now regarded as a distinct entity from Ewing sarcoma/PNET.

Reports of ONB are infrequent in the veterinary literature. Only a few cases of this neoplasm have been described in dogs, cats, cattle, monkeys, horses, fish, axolotl and transgenic mice.[11-20]

Pathological grading and clinical staging

Two major systems of classification of ONBs in humans have been established and are currently in use: Hyams' grading system and Kadish staging system. The former takes into account level of cellular differentiation, the mitotic index and other microscopic features shown in Table 1. Hyams' system classifies ONBs into four grades.[21] There is a continuum from grade I to grade IV and often a definite separation between grades is arbitrary. Briefly, grade I includes the majority of tumours with the most differentiated cells while grade IV neoplasms are the most undifferentiated (anaplastic) tumours.

Table 1. Hayms' grading system of olfactory neuroblastoma
Microscopic featuresGrade IGrade IIGrade IIIGrade IV
  1. FW, Flexner–Wintersteiner rosettes; HR, Homer–Wright pseudorosettes; NF, neurofibrillary.
PleomorphismAbsent to slightPresentProminentMarked
NF matrixProminentPresentMay be presentPresent
GlandsMay be presentMay be presentMay be presentMay be present

Staging system for ONB classification was proposed in 1976 by Kadish et al. It is based on the local spread of the neoplasm, which includes three stages: stage A: tumour limited to the nasal cavity; stage B: tumour within the nasal cavity and paranasal sinuses; and stage C: tumour growing beyond the nasal cavity and sinuses (Table 2).[22] To some extent, grade and tumour stage correlate with prognosis for ONBs. However, in predicting the recurrence of ONB or biological behavior, there are no valuable prognostic indicators other than the Kadish staging system.[23] Reports on the application of these classification systems of ONBs in animals are limited. However, due to the similarities between tumours in man and domesticated animals, both classification systems appear to be applicable and were utilized in a few studies on ONB in animals, which have been conducted recently.[7, 11, 17]

Table 2. Clinical Kadish staging system of olfactory neuroblastoma
Stage       Extent of tumour5-year survival (%)
ATumour confined to nasal cavity75–91
BTumour involves the nasal cavity plus one or more paranasal sinuses68–71
CExtension of tumour beyond the sinonasal cavities41–47

Despite the advances in histopathology and immunohistochemistry, there are still some difficulties in unambiguous diagnosis of ONB. One of the main problems is similarity between this tumour and sinonasal undifferentiated carcinoma (SNUC), malignant melanoma and neuroendocrine carcinoma.[24] Another issue is the diversity of histopathological appearance of these tumours. ONB encompasses a spectrum of morphological subtypes. Also clinical signs are non-specific and may be variable as they largely depend on the extent of tumour involvement.


No clear aetiology agent has been established in the humans so far. The injections of diethylnitrosamine into Syrian hamster[25] and N-nitrosopiperidine into rats[26] have triggered the development of tumours histologically identical to the human ONB. In 1990, there were three cases of ONB reported in cats infected with the feline and murine leukaemia C type retrovirus.[27] Similarly, a pathogenic agent was observed in transgenic mice that developed ONB as a result of an activation of endogenous human adenoviruses.[20] However the role of the virus in human ONB needs further exploration.

Clinical, macroscopic and microscopic features

ONB in human

In humans, ONBs occur at any age, but most commonly in the second and the sixth decades of life[3, 21] and there is neither gender nor racial difference in its occurrence.[28] Its incidence is approximately 2–3% of all tumours of the nasal cavity and paranasal sinuses. The tumour usually causes unilateral nasal obstruction and epistaxis. Less commonly, headache, pain, excessive lacrimation, rhinorrhea, anosmia and visual disturbances are observed.[29-31] In many patients these signs and symptoms are being present for a long time before the tumour diagnosis, which is compatible with slow rate of tumour growth and their non-specific nature at the initial presentation. Generally, the mean time delay between the appearance of the first symptoms and the diagnosis varies from 6 to 12 months.[28]

Taken together, descriptions of ONB cases in humans show some characteristic features of tumours. The tumour is usually a unilateral, polypoid, glistening, soft, red-grey mass with an intact mucosa. Size of tumours ranges from less than 1 cm up to large masses involving the nasal cavity and intracranial region and frequently expanding into paranasal sinuses, orbits and cranial vault. Despite the morphological heterogeneity of ONB, some specific and reliable histological criteria have been established in man. The majority of human ONB tumours have a lobular architecture, which is the most important histological feature. The neuroblastoma tumour cells are small, round, blue cells with high nuclear to cytoplasmic ratio. Lobules and nests formed by neoplastic cells are often separated by a richly vascularized fibrous stroma, or less commonly, show diffuse growth pattern. The nuclei are small and uniform with hyperchromatic, ‘salt-and-pepper’ nuclear chromatin distribution with inconspicuous nucleoli. For the most part (except high grade lesions) nuclear pleomorphism, mitotic figures and necrosis are uncommon. Two types of rosettes are recognized: Homer–Wright pseudorosettes seen in up to 30% of cases, and true Flexner–Wintersteiner rosettes seen in about 5% of cases.[21, 32] On immunohistochemical staining human ONBs are positive for synaptophysin, chromogranin, neuron specific enolase (NSE), neurofilament protein (NF) and S-100 protein. NSE is the most consistently expressed marker, with up to 100% of tumours showing positive staining. The NF immunoreactivity is usually low. Expression of chromogranin is related to the tumour differentiation, the poorer differentiation, the lower immunoreactivity of this marker. Synaptophysin expression is unrelated to the level of differentiation and usually has higher sensitivity comparing to NF and chromogranin. S-100 staining is usually scattered and labelled cells are found at the periphery of the neoplastic lobules. At the same time, these peripheral cells may be positive for glial filament acidic protein (GFAP).[21] Human ONBs are negative for desmin, myogenin, leucocyte common antigen, as well as MIC2 (CD99) antigen and vimentin.[6, 32-34] Also cytokeratin CAM5.2 and AE1/AE3 are usually negative, but in some cases they can show weak immunoreactivity.[6, 32-34] Proliferation marker Ki-67 has shown high range of proliferative index, 10–50%[21] (see Table 3).

Table 3. Immunohistochemistry profiles of ONBs
  1. aDopke et al.[15]
  2. bYamate et al.[16]
NFPresentPresenta/absentbAbsent to slightAbsent to slightAbsentPresentAbsent
GFAPMay be presentPresentMay be presentMay be presentAbsentUnknownAbsent
SynaptophysinPresentPresenta/absentbMay be presentMay be presentAbsentUnknownAbsent
ChromograninPresentAbsentMay be presentMay be presentUnknownUnknownAbsent
Cytokeratin (CAM5.2, AEI/AE3)May be presentAbsentMay be presentMay be presentPresentUnknownPresent

Ultrastructure of ONB

At the ultrastructural level in human ONB 80–230 nm dense core granules are observed within perikarya and in microtubule-containing processes in all tumours.[35] In majority of cases lobules were surrounded by electron-dense sustentacular cells. Few tumours exhibited obviously epithelial features in addition to neuronal differentiation.[35] There are limited data about the ultrastructure of ONB in animals. In domestic medaka (Oryzias latipes) tumour cells had extended cytoplasm in which parallel neurotubules and a few neuroendocrine granules were present. In the perinuclear region, bundles of intermediate filaments and neuroendocrine granules were seen. Single cilia and a pair of centrioles were occasionally found, however, no ciliated cells were found. Some large tumour cells contained electron-dense intracytoplasmic inclusions which showed a crystalloid structure.[18] In the report of Ishikawa et al. tumour demonstrated the presence of numerous ciliated olfactory sensory cells as well as sustentacular, mucous and basal cells within the neoplastic simulations of olfactory membrane, which can be interpreted as a relatively high level of maturation to olfactory neuroepithelium of this tumour.[36] Hara et al. demonstrated tight junction (desmosome-like structures) in an ONB from a dog, however, no dense core secretory granules were found.[37] On the other hand, multiple dense core secretory granules were observed in an ONB from a horse. Additionally, microfilaments, mitochondria with tubular to vesicular cristae, rough endoplasmatic reticulum in single strands or parallel arrays, a few free ribosomes and multiple Golgi complexes were present.[15]

Differential diagnosis of ONB

ONB can be misdiagnosed as another small round blue cell tumours of the sinonasal tract. The first aspect in differential diagnosis of ONB is particular location of the tumour. ONB seems to be derived from olfactory epithelium, which is present in superior turbinate, cribriform plate and superior third of the nasal septum.[28] This is also highly preferential location of the ONB.[28] It should be, however, mentioned that primary ONB was described also in other locations, for example, above the cribriform plate,[38] in the sella turcica,[39] inferior meatus of the nasal cavity,[40] pterygopalatine fossa,[41] sphenoid sinus[42] or ethmoid sinus.[43] Mimickers of ONB can be located in superior turbinate, cribriform plate and superior third of the nasal septum as well.[21]

The diagnosis is relatively easy in case of well-differentiated tumours (Haym's groups I and II), with proliferation of small round cells with fibrillary eosinophilic background material. On the other hand, in high grade tumours (Haym's groups III and IV) marked nuclear pleomorphism, frequent mitoses, areas of necrosis and absence of fibrillary background make the diagnosis challenging.[44] Other tumours of the nasal cavity that come in the differential diagnosis are:

  1. of epithelial origin: poorly differentiated, non-keratinizing squamous cell carcinoma, SNUC, small cell carcinoma, neuroendocrine type (SCCNET);
  2. of melanocytic origin: sinonasal mucosal malignant melanoma;
  3. of mesenchymal origin: rhabdomyosarcoma (RMS), embryonal and alveolar, poorly differentiated synovial sarcoma;
  4. haematolymphoid: extramedullary plasmacytoma and extranodal NK/T cell lymphoma (nasal-type).[45]

ONB is negative or only weakly and focally positive for cytokeratin, which distinguish it from SNUC and SCCNET. Poorly differentiated, non-keratinizing squamous cell carcinoma can be distinguished by presence of in situ component or direct continuity to the overlying surface epithelium and lack of the expression of neuroendocrine markers[45] SNUC shows cells with hyperchromatic to vesicular nuclei, poorly defined high nuclear-to-cytoplasmic ratios and often single, prominent nucleoli.[45] ONB can be distinguished from SNUC in fine needle aspiration (FNA) by the absence of mitoses, frequent occurrence of Homer–Wright rosettes, fibrillary cytoplasm and smooth nuclear contours with indistinct nucleoli. In general, cytological features of SNUC are that of a high grade carcinoma while of ONB are that of neuroendocrine tumour.[46] Sinonasal SCCNETs differs from high grade ONBs by its more primitive oval or round hyperchromatic nuclei and absent or inconspicuous nucleoli. Extensive apoptosis, confluent necrosis, haemorrhages, crush artefact and a high mitotic rate are common.[45] Melanoma, especially amelanotic, can have very variable histology. Crucial in the differential diagnosis (DD) with the ONB is the expression of S-100, HMB45 and vimentin.[45] However, ONB can very rarely produce melanin and one such case was initially misdiagnosed as melanoma of the nasal cavity.[47] Alveolar RMS can express CD56 and synaptophysin. Crucial in the diagnosis is expression of myogenic markers and presence of specific rearrangements involving Pax3 and Pax7.[45] In the DD with poorly differentiated synovial sarcoma is the expression of cytokeratin and epithelial membrane antigen (EMA) and specific translocation t(X;18)(p11.2;q11.2). Extramedullary plasmacytoma is distinguished from ONB by typical pattern of nuclear chromatin and positivity to EMA.

Extranodal NK/T cell lymphoma (nasal-type) share with ONB CD56-positivity. Other neuroendocrine markers, however, are negative. Additionally, unlike ONB, this lymphoma shows very prominent angiocentric and/or angioinvasive growth pattern, expression of NK/T cell markers and presence of Epstein-Barr Virus (EBV).[21, 45] Neuroblastoma metastatic from the adrenal gland to the nasal cavity would have the same morphology, the discriminative feature is lack of MYCN amplification.[21]

Calretinin is positive in ONB but rather negative in other small round blue cell tumours of the sinonasal tract. Immunophenotype p63− calretinin+ (especially when diffuse and intensive) seems to be specific for ONB.[48]

In a small group of patients ONB expressed the olfactory sensory transduction proteins, Olfactory-specific G-protein (Golf), ACIII and CNGCA2, which are known to be specifically expressed in olfactory sensory neurons.[49] However, their usefulness in DD of ONB with other small round cell tumours of sinonasal tract has not been proven yet.

ONB in horse

Up to now, two cases of ONB have been described in a 11- and 17-year-old horse. In the first case the animal was referred to veterinary hospital with a 3-month history of lacrimation and exophthalmos of the right eye and epistaxis from right nasal cavity followed by nasal haemorrhage and proptosis.[16] Second animal was referred to clinic with severe proptosis of the left eye.[15]

In both the cases tumours were unilateral and had grown to large sizes. In the first case the nodule occupied the olfactory region of the right nasal cavity, the right-side nasal turbinates and the right-side frontal sinuses. In the second case mass involved the left ethmoid bone, maxillary sinus, superior turbinate and the left orbit. Equine tumours are composed of highly cellular, uniform cell population. As in human cases, neoplastic cells were round to moderately elongated in shape, with small amount of cytoplasm, and round, hyperchromatic nuclei. Small amount of collagen fibres among neoplastic cells, and intracellular fibrillar background throughout tumour tissue were distinguishable. Rosette-like arrangements of tumour cells were often seen around connective tissue surrounding blood vessels or directly around blood vessels. Also occasionally Homer–Wright true rosettes were seen in tumour tissue. Incidence of pseudorosettes and rosettes depended on grade of the tumour. Mitotic figures, depending on the particular case, were numerous[16] or inconspicuous in equine tumours.[15, 16]

Immunohistochemistry (IHC) studies have shown that equine ONBs are positive for NSE, GFAP and S-100 protein. In case of NSE no characteristic labelling pattern has been identified. There was focal expression of GFAP along cytoplasmic processes, whereas S-100 protein was diffusely distributed. As reported by Dopke et al., NF labelling was particularly intense in perinuclear regions and also along cytoplasmic processes (not described in the report of Yamate et al.). The majority of tumour cells showed weak vimentin labelling, whereas all cells were negative for cytokeratin, desmin and chromogranin[15, 16] (Table 3).

ONB in mouse

According to the study conducted by Koike et al.[20] the development of ONB in mice was associated with an activation of endogenous human adenoviruses in transgenic mice. Mice developed ONB between the ages of 6 and 9 months. Animals developed progressive lethargy with bulging anterior to the cranium. After examination tumours were found in the nasal sinuses (where they originated from), along the cribriform plate and within the olfactory bulbs of the anterior brain. Interestingly, authors of the study confirmed that this association also occurs in domestic cats.[27]

Histological examination supported by electron microscopy performed on tissues from necropsied animals have shown that murine tumours were composed of solid sheets of poorly differentiated, neoplastic cells extending in nasal mucosa and connective tissue. Neoplastic cells varied from round to oval in appearance and occasionally formed characteristic rosettes. As opposed to human ONBs, immunoperoxidase staining for NSE and S-100 were negative in murine neoplastic cells.[20, 53]

ONB in cat

ONB was identified in nine cats so far, with the animal age range between 3 and 10 years. The most common clinical features observed were unilateral epistaxis, nasal discharge and aggressive behavior.

Brosinski et al. in their elegant study examined nine feline ONB cases. Tumours from all samples showed various grades of differentiation. However, the majority of them were of a low grade. Taken together, data obtained by their study indicate that well-differentiated ONBs are highly cellular, and (again as in previous cases) often arranged in lobules or nests separated by a variable amount of fibrovascular stroma. Neoplastic cells are composed of uniform, round to oval hyperchromatic nuclei with dispersed chromatin and inconspicuous nucleoli, which have scant cytoplasm and indistinct cell margins separated by a fine neurofibrillary matrix. Rosette-like arrangements are commonly present in feline ONBs. Comparing to human high-grade ONB tumours (grades III and IV), in feline cases of ONB increased mitotic activity was observed, pleomorphism and extensive necrosis were also present. However, even in low-grade tumours mitotic activity was noticeably higher and this is the major difference between feline ONBs and those occurring in man.[11]

All samples were also examined by IHC. Proliferation marker Ki-67 was expressed in all specimens. Positive cells were located randomly, sometimes showing rosette-associated labelling pattern. It was also shown that the Ki-67 labelling index (LI) strongly correlated with tumour grade. Eight of nine ONBs were microtubule-associated protein-2 (MAP2)-positive. MAP2 expression in tumour cells varied widely (1–80% labelled cells). Nevertheless, there was a strong correlation between expression of MAP2 and grade of differentiation of the tumour. All cases of feline ONBs were positive for NSE labelling with the majority of neoplastic cells showing cytoplasmic NSE expression. Synaptophysin and NF were less frequently detected, whereas chromogranin, GFAP and cytokeratin might be present (Table 3).[11, 13]

Up to 1990, three cases of spontaneous ONB in cats were described, which were related to viral infection. In each case, numerous mature C type retrovirus particles were identified within the tumours. All cats were serologically tested for feline leukaemia virus (FeLV) and two of them were positive for FeLV. Although the relationship between retroviral infection and development of ONB was uncertain at that time, FeLV was proposed as an aetiologic agent of ONB. This thesis was based on data achieved during extensive study which included microscopical, immunocytochemical and polymerase chain reaction (PCR) analysis of tumour specimens.[27] These tumours had all characteristic features of ONBs. Histologically, the tumours were comprised of a highly cellular, uniform population separated by fibrovascular stroma. The neoplastic cells were generally oval, with scant cytoplasm, round nuclei with fine, punctate marginated chromatin and indistinct nucleoli. Cells were predominantly arranged in sheets, however, perivascular palisades, pseudorosettes and true rosettes were occasionally observed.[27]

Two of three cases had random clusters of NSE-positive cells. In each case sporadic individual S-100-positive cells were present, fusiform or elongated in shape, mostly located at the periphery of lobules. There was also cytokeratin staining in all cases. Most often the positive cells occurred in clusters, continuous cords or true rosettes. All cases were negative for vimentin, NF and GFAP staining. Nowadays, we know that diagnosis of feline ONB is not always parallel with the infection with FeLV (see Table 3).[27]

ONB in dog

So far, 13 cases of ONB have been reported in dogs. The age range was 6–15 years. As in case of feline ONBs the most common clinical features observed in dogs were unilateral epistaxis and nasal discharge. Other clinical signs recorded were anorexia, aggression and central nervous system dysfunctions.

Tumours examined in the study of Brosinski et al. showed various grades of differentiation with the prevalence of poorly differentiated tumours. Similar to feline tumours, well-differentiated ONBs are highly cellular, often arranged in lobules or nests separated by fibrovascular stroma. Neoplastic cells have round to oval hyperchromatic nuclei with dispersed chromatin and hardly visible nucleoli. Cells have scant cytoplasm and indistinct cell margins separated by a fine neurofibrillary matrix. Rosette-like arrangements can be observed. However, they are not common in canine ONBs. Similar to feline tumours, canine cases of ONB have shown increased mitotic activity, presence of pleomorphism and extensive necrosis, when compared with human ONB tumours.[11]

IHC study revealed that all canine samples were positive for proliferation marker Ki-67. Ki-67 LI was lower than in feline ONBs, but still it strongly correlated with tumour grade. All ONB samples were MAP2-positive and showed varied MAP2 expression in tumour cells. No statistically significant correlation between expression of MAP2 and grade of differentiation was observed. In all samples, up to 95% of neoplastic cells were NSE-positive. Synaptophysin and NF were less frequently detected, whereas chromogranin, GFAP and cytokeratin might be present (Table 3).[11, 13]

ONB in cattle

In 1981, Anderson and Cordy described the first case of ONB in cattle. Fourteen-month-old heifer was referred with mouth breathing and severe proptosis of the left eye. Macroscopically, the neoplasm was unilateral, extensive irregular mass occupying two-third of nasal cavity and extending into the left nasopharanyx, palatine and maxillary sinus. Proptosis of the eye was due to displacement of medial wall of the left orbit by the tumour mass. ONB involved also the left olfactory bulb and tract. Microscopically it was highly cellular tumour with lobular architecture with fibrovascular septa separating sheets and lobule-like groups of cells. Typically for all ONB cases tumour cells were small and round in shape with sparse cytoplasm, hyperchromatic nuclei with granular chromatin and evident nucleoli.[14]

ONB in fish

ONB was also reported in a few species of fish: Sparus aurata,[50] Coregonus hoyi,[51] Cyprinus carpio,[36] O. latipes[18] and in Carassius auratus.[17] Despite the fact that its occurrence in fish was mentioned as early as in 1932[36] for a long time there was no detailed description of this neoplasm in veterinary literature.

In 1989, spontaneous ONB in a domestic medaka, O. latipes, has been reported by Torikata et al.[18] The tumour was unpigmented, measured 1.5 mm in diameter and occupied both olfactory chambers but was present mainly in the right of them. The tumour was highly cellular and two cell populations were distinguishable, small undifferentiated cells and large polygonal cells with numerous mitoses. Similarly to all described cases, neoplastic cells were organized in nests and separated by delicate fibrovascular tissue. Occasionally, large polygonal cells formed true rosettes.

Both types of neoplastic cells stained positively for NSE and NF. Some large cells were also positive for S-100 protein immunostaining.

In 2011, Vigliano et al. described in detail the first case of ONB in a goldfish C. auratus. As in all animal cases described above, it was unilateral, pinkish tumour emerging from the right nostril of animal. Histologically, the neoplasm presented a well-developed fibrovascular stroma associated with solid cell nests and a large number of Flexner–Wintersteiner rosettes surrounded by myelinated fibres. Neoplastic cells showed a prominent degree of nuclear atypia and low mitotic activity. On the one hand, common occurrence of true rosettes and nuclear pleomorphism are unlike most of human neoplasm. Another difference is the presence of pale, euchromatic nuclei with conspicuous nucleoli. On the other hand, lobular architecture and low mitotic activity are with parallel to cases of ONB described in man. Comparing ONB cases reported in other species described in this article, goldfish ONB differs to higher extent from human and mammal neoplasms (Table 4).[17]

Table 4. Comparison of ONB among species
  1. aOnly one case reported so far.
  2. bDifferences between two cases of ONB in horse are caused by the different grade of tumour in both animals.[2, 11]
Mean age40–45 years (however, the highest incidence in the second and sixth decade of life)11 and 17 years7.5 months9.8 years6.8 yearsUnknown1 yeara14 monthsa
Tumour external appearanceUnilateral, polypoid, soft, red-grey mass with intact mucosaUnilateral, soft, red-grey/yellow nodulebUnilateral tumourUnilateral tumour (various appearance)Unilateral, soft, pink massUnilateral, soft, pink/unpigmented massPapillary, polypoid, pink massUnilateral, irregular mass
SiteUpper nasal cavityNasal cavityNasal sinuses, cribriform plateUpper nasal cavityUpper nasal cavityNostrilOral cavityNasal cavity
PatternLobulesLobules and nestsSheetsLobules and nestsLobules and nestsLobules and nestsNests and sheetsLobules and sheets
CytologySmall, round cells with scant cytoplasm, hyperchromatic nuclei with inconspicuous nucleoliRound cells with scant cytoplasm, hyperchromatic nuclei with inconspicuous nucleoliRound cellsRound cells with scant cytoplasm, indistinct cell borders, hyperchromatic nuclei with inconspicuous nucleoliRound cells with scant cytoplasm, indistinct cell borders, hyperchromatic nuclei with inconspicuous nucleoliSmall cells, pale euchromatic nuclei with conspicuous nucleoliSmall, round cells with scant cytoplasm, small and hyperchromatic nucleiSmall, round cells with scant cytoplasm, elliptical, hyperchromatic nuclei with conspicuous nucleoli
Nuclear pleomorphismUncommonUnknownUnknownOccasionallyOccasionallyCommonUnknownUnknown
AnaplasiaOccasionally and focally (grade dependent)UnknownUnknownUnknownUnknownUnknownUnknownUnknown
Mitotic figuresVariable (grade dependent)Numerous/inconspicuousbUnknownFrequent (especially in high grade tumours)Frequent (especially in high grade tumours)UncommonOccasionallyNumerous
NecrosisOccasionally (grade dependent)Occasionally/inconspicuousbUnknownFrequent (especially in high grade tumours)Frequent (especially in high grade tumours)OccasionallyUnknownFrequent
Neurofibrillary stromaCommonCommonUnknownCommonCommonCommonUnknownCommon
True rosettesUncommonOccasionally/inconspicuousbUnknownOccasionallyCommonCommonCommonInconspicuous

Immunohistochemistry studies of fish ONB sections revealed diffuse pattern of S-100 immunostaining, positive cells were mainly located in rosettes and solid nests. In addition, the majority of cells of rosettes and the solid nests were also positive for cytokeratin staining. However, no specific reactivity was observed for NSE, NF, GFAP and synaptophysin, either in tumour or in goldfish control sections. And this is the major and significant difference between the fish ONB and the rest of human and animal cases of this neoplasm described in this review.[17]

ONB in axolotl

Recently, ONB was also described in a 1-year-old Mexican axolotl. Neoplasm was pink, papillary tumour mass extending from the ethmoturbinate region into the oral cavity. As in all previous cases, tumour mass was composed of a lobular architecture. Small cells with scant cytoplasm and small round nuclei formed nests and sheets. Also rosette-like arrangements were often seen, whereas mitotic figures were occasionally observed.[19]


Given its rarity, treatment options for this neoplasm have not been well evaluated. There are no standard treatment protocols in animal cases. Treatment modalities used for neoplasia of the nasal cavity include surgery, radiotherapy, chemotherapy, immunotherapy, cryotherapy or combination of these options. As presented in work of Ueno et al., treatment with orthovoltage X-ray radiation is the most effective therapy for controlling nasal tumours in dogs.[12]

Therapy for human ONB includes surgery alone, radiation alone and combined approaches. The most widely used therapeutic option in operable ONB is surgical resection followed by adjuvant radiotherapy, locally advanced unresectable tumours are managed with definitive radiotherapy. Palliative latinum-based chemotherapy is used in patients who are not candidates for radical treatment. There are no approved targeted therapies for ONB.[52]

ONB is a tumour whose molecular biology is relatively unknown. Long-term survival at the level of 60–70%[53] make further insight into its molecular biology necessary to determine new targets for treatment. Until now, this search resulted in a limited success; data about target treatment in ONB are restricted to case reports. In a report of Preusser et al. expression of platelet-derived growth factor receptor-beta (PDGFR-β) on the stromal cells of ONB widely disseminated after surgery and radiotherapy was an inspiration for introduction of multi-tyrosine kinase inhibitor (TKI) sunitinib mesylate. The patient responded well to the therapy with significant improvement of symptoms, including recovery of Karnofsky's Prognostic Score (KPS) from 40 to 70%, disease stabilization for 15 months and no significant toxicity.[54] There is also a report of durable response in a patient with heavily pretreated ONB treated with monoclonal antibody against vascular endothelial growth factor (VEGF), Bevacizumab. Radiographic response included a partial response by month 2, followed by Stable Disease (SD) for approximately 22 months, followed by asymptomatic progression of the parotid, cervical and supraclavicular LNs by approximately month 24. Additionally, there was an increase of KPS from 80 to 90% and palliation of his fatigue, neuro-cognitive deficits, aphasia and seizures. The patient was also able to discontinue dexamethasone.[55] Another possible treatment option is temozolomide, an alkylating agent with significant activity against human malignant gliomas. There is a description of 23-months stable disease in heavily pretreated ONB with multiple metastases to the central nervous system (CNS).[56]

Another promising therapy in ONB is blockage of sonic hedgehog signalling pathway. In a study of Mao et al. Patched1, Gli1 and Gli2 was detected in 70, 70 and 65% of human ONB specimens, respectively, but not in normal human olfactory epithelium.[57] Treatment with cyclopamine, a selective inhibitor of the Shh pathway, inhibited the proliferation and colony formation of ONB cells, induced ONB cell cycle arrest and apoptosis, and down-regulated the expression of Patched1, Gli1 and cyclin D1 while up-regulating p21 expression in vitro. These regulatory effects of cyclopamine were partially or completely erased by exogenous Shh.[57] Despite these promising results no study or case report of clinical cyclopamine usage has been described in the literature.

It seems that Tumor Necrosis Factor (TNF)-related apoptosis-inducing ligand (TRAIL) and Bortezomib, an proteasome inhibitor can be potentially effective in ONB. Freshly isolated primary esthesioneuroblastoma cells are completely resistant to TRAIL-induced apoptosis despite expressing all components of the TRAIL pathway; however, they can be sensitized by subtoxic doses of Bortezomib. Bortezomib cause up-regulation of TRAIL death receptor expression, enhancement of the TRAIL death-inducing signalling complex (DISC), and down-regulation of anti-apoptotic proteins of the TRAIL pathway. From the clinical point of view it is important that TRAIL-resistant primary tumour cells be repeatedly sensitized by Bortezomib, providing the basis for repeated clinical application schedules.[58]

Bcl-2 positivity can be potentially useful in predicting response to neoadjuvant chemotherapy[59] and strongly correlates with Hyam's grading.[60] Additionally, a significant positive correlation between the expression levels of Hypoxia-inducible factor 1 alfa (HIF-1α); bcl-2 and the microvessel density was found. This suggests that bcl-2 may act as a stimulator of angiogenesis in ONB, and thus represents a novel target and/or novel marker for anti-angiogenic treatment strategies in the therapy of ONB.[61]

New sequencing techniques can be extremely helpful in finding new efficient target drugs in ONB. Nice study of whole genome sequencing showed presence of mutations of MAP4K2, SIN3B, TAOK2, KDR, TP53, MYC and NLRC4 in disseminated ONB. Among them TP53, MAP4K2 and TAOK2 were present in primary and lymph node metastases, while KDR, MYC, SIN3B and NLRC4 only in the tumour specimen from nodal metastasis. This observation suggests that KDR, MYC, SIN3B and NLRC4 can be associated with lymph node dissemination in ONB. However, clinical usefulness of this observation is so far sparse.[62]


ONB shows striking similarities on the morphological, immunohistochemical, ultrastructural and molecular level in humans and animals. Future detailed studies conducted on animal ONB cases can be helpful not only in understanding the pathophysiology of ONB but may also help elucidate the origin of this neoplasm, which still remains to be investigated.

In conclusion, combining the knowledge and experience of veterinarians and oncologists can give us bilateral benefits. As the microscopical appearance and molecular phenotype of animal ONBs appears similar to those in humans, animal ONB models would appear useful for developing new approaches to the treatment of these tumours in humans. On the other hand, a detailed clinical staging, prognostic factor systems and existing treatment protocols for ONB in humans would be invaluable for treating animals in the future.