Multimodality imaging of sinonasal undifferentiated carcinoma
Click Here to Manage Email Alerts
A 53-year-old female nonsmoker presented to an outpatient facility in November 2017 with 6 weeks of progressive nasal congestion, facial pressure and headaches.
She was empirically treated with antibiotics without relief.
She subsequently developed severe headaches — retro-orbital, frontal and maxillary — with nausea, photophobia and phonophobia, and she went to the ED.
Sinus CT obtained at an outside institution was concerning for acute sinusitis with polyposis.
The patient began IV antibiotics, in addition to a steroid taper. After minimal relief, her symptoms recurred in January. The patient also experienced an episode of double vision, at which point she was referred to neurology and ophthalmology at our medical center for further evaluation.
Repeat sinus CT was obtained.
Imaging and pathology
CT of the sinuses without contrast obtained on Jan. 10 appeared significant for a hyperattenuating mass centered in the superior nasal cavity and ethmoid sinus, eroding the anterior skull base, including olfactory recesses, cribriform plates and inferior crista galli, with intracranial extension.
It also showed significant erosion of the superior nasal septum and ethmoid septae, with marked thinning and remodeling of the bilateral lamina papyracea (Figure 1).
The mass attenuation (max, 55 Hounsfield units [HU]) was considered highly suggestive of a solid neoplasm and, less likely, an aggressive mucocele (Figures 2a and 2b).
Contrast-enhanced magnetic resonance evaluation was recommended to better delineate the intracranial extent of disease. MRI of the sinuses performed on Jan. 14 revealed a heterogeneously enhancing, intrinsically T1 isointense, T2 hyperintense mass involving the superior nasal cavity, bilaterally measuring 4.9 cm by 4.4 cm by 5.1 cm (anteroposterior by transverse by craniocaudal [Figures 3a and 3b]).
The mass inferiorly extended between the nasal septum and right inferior turbinate, and also to the superior margin of the left inferior turbinate within the nasal cavity. Tumor also bulged into the superomedial aspect of the right maxillary sinus. Posteriorly, the mass extended into the anterior aspects of the sphenoid sinuses.
Tumor filled the ethmoid sinuses and superiorly extended into the anterior cranial fossa — eccentric to the right — through the cribriform plates, lateral lamella and fovea ethmoidalis. The intracranial component measured 2 cm by 1.2 cm in maximum area in the coronal plane (Figure 4a).
There were smaller discrete nodules along the dura further posteriorly, the larger measuring 0.8 cm by 0.5 cm in the region of the anterior right planum sphenoidale (Figure 4b). There also was adjacent smooth dural enhancement and thickening along the left anterior cranial fossa floor (Figure 4c).
There was lateral bulging of the lamina papyracea bilaterally. Posteriorly, there was focal protrusion of tumor into the superomedial orbits bilaterally, larger on the right, measuring 1.2 cm by 0.9 cm in the coronal plane (Figures 4a and 5).
There was superolateral displacement of the bilateral superior oblique muscles and inferolateral displacement of the bilateral medial rectus muscles. Tumor focally abutted the medial margin of the optic nerve sheath complexes bilaterally in the posterior orbit, without evidence of encasement. Further anteriorly, the periorbita appeared intact.
There was postobstructive opacification of the right frontal and maxillary sinuses and bilateral sphenoid sinuses, with central T2 hypointense signal compatible with inspissated secretions versus fungal colonization.
There was partial opacification of the left frontal and maxillary sinuses. There was no abnormal enhancement in the brain parenchyma.
The remainder of the extraocular muscles appeared intact. There was no signal abnormality of the optic nerves, chiasm or tracts. There was no cervical lymphadenopathy. These findings were nonspecific, considered to reflect esthesioneuroblastoma or sinonasal malignancy, such as squamous cell carcinoma, sinonasal undifferentiated carcinoma or, less likely, lymphoma.
The patient proceeded to biopsy of the nasal cavity mass, which revealed high-grade carcinoma with necrosis, consistent with sinonasal undifferentiated carcinoma (SNUC).
Specifically, the biopsy demonstrated diffusely invasive high-grade carcinoma with widespread tumor necrosis. Cytologically, the tumor was composed of highly pleomorphic neoplastic cells with high nucleus-to-cytoplasm ratio, mitotic figures (> 5/high-powered field), enlarged vesicular nuclei and inconspicuous nucleoli.
The tumor showed no discrete glandular or squamous differentiation. Immunohistochemical studies revealed the tumor cells to be variably positive for epithelial markers, including cytokeratin AE1/AE3, Cam 5.2, CK19, EMA, and weakly or focally positive for neuron-specific enolase, synaptophysin and CD56.
The tumor was negative for S-100, Melan-A, calretinin, CD45, CD3, desmin, TTF-1, CK7, CK5/6, p40 and chromogranin.
INI-1 was positive and p16 showed less than 50% tumor cell staining. In situ hybridization for Epstein-Barr virus was negative. The overall histopathologic and immunohistochemical features of the tumor were most consistent with SNUC.
The patient was considered unresectable at the time of presentation given the degree of intracranial and intraorbital extension. An initial staging PET/CT evaluated for regional and metastatic spread of disease prior to initiation of concurrent cisplatin chemotherapy and external beam radiation treatment.
Vertex-to-thighs protocol fluorodeoxyglucose (FDG) PET/CT performed on Jan. 23 redemonstrated a markedly hypermetabolic superior nasal cavity and ethmoid sinus mass (standard uptake value [SUV], 54) extending toward the medial orbits and into the intracranial compartment along the anterior cranial fossa, eccentric to the right.
The medial orbital wall involvement abutted the bilateral medial rectus and superior oblique muscles; however, there was no extension of FDG activity into the orbit proper. The intraconal fat planes otherwise were clear on both sides.
The intracranial fossa floor was again seen to be diffusely demineralized with multifocal dehiscence. However, metabolic imaging did not demonstrate any evidence of intracranial extent beyond the dural disease noted on earlier MRI sinus evaluation (Figures 6a and 6b).
PET/CT evaluation for nodal disease revealed a hypermetabolic 1-cm right retropharyngeal lymph node (SUV, 19 [Figures 7a, 7b and 7c]). No lateral or central neck metastatic nodal disease was identified.
PET/CT evaluation for metastatic disease demonstrated no evidence to suggest liver or lung metastases. In addition, there was no distant bone metastases to the axial or appendicular skeleton.
Discussion
SNUC is a rare and highly aggressive neoplasm arising in the paranasal sinuses, initially described as a separate entity by Frierson in 1986.
SNUC is believed to originate from the Schneiderian epithelium or from the nasal ectoderm of the paranasal sinuses, and it is considered a distinct entity from other poorly differentiated neoplasms of the nasal cavity and sinuses.
It is associated with cigarette smoking and prior radiation therapy, but there is no known relation with prior Epstein-Barr virus infection. The most common clinical symptoms at presentation include nasal obstruction, proptosis, cranial nerve palsies, periorbital swelling, diplopia, epistaxis and periorbital pain.
As was the case with our patient, SNUCs tend to be rather advanced on initial imaging evaluation. Therefore, SNUC typically is described as a rapidly enlarging mass involving multiple sites, often with clear evidence of extension beyond the anatomic confines of the sinonasal tract, including invasion of the anterior cranial fossa, adjacent paranasal sinuses and orbits.
CT evaluation usually reveals a large primary tumor mass with poorly defined margins, most commonly arising within the ethmoid sinuses and superior nasal cavity. In addition, aggressive features — such as bone destruction and dehiscence — often are seen, as is obstruction of adjacent sinuses.
Contrast-enhanced MRI better delineates soft tissue spread of disease. Magnetic resonance findings include heterogeneous enhancement of the tumor, which is intrinsically T1 isointense and T2 iso- to hyperintense compared with adjacent muscle.
Hybrid imaging studies — such as FDG PET/CT and PET/MRI — demonstrate avid radiotracer uptake, as well as highly metabolic regional and distant metastatic spread of disease. The major role of functional imaging is clear delineation of local, regional and metastatic disease extent due to the high rate of both local-regional recurrence and distant metastasis in SNUC.
Of note, it is important to remain cognizant of the differential diagnosis for masses that arise in the nasal cavity and paranasal sinuses, which includes both benign and malignant entities. Primary intraosseous lesions in this region include central giant-cell granulomas, developmental odontogenic cysts (dentigerous cyst, odontogenic keratocyst) and primary odontogenic neoplasms, such as ameloblastoma.
Further, odontogenic infection is the most common cause of soft tissue swelling involving the paranasal sinuses and may, at times, be mistaken for a neoplastic process. Less common reactive or neoplastic lesions of connective tissue origin — such as inflammatory myofibroblastic tumor, nodular fasciitis, myofibroma and desmoplastic fibroma — also should be included in the differential diagnosis.
The common malignant lesions in this region include esthesioneuroblastoma, nasopharyngeal carcinoma, neuroendocrine carcinoma, rhabdomyosarcoma, lymphoepithelioma, lymphoma, melanoma and poorly differentiated adenoid cystic carcinoma. However, it is nearly impossible to differentiate SNUC from other neoplasms that arise in this region on the basis of imaging characteristics alone, with the possible exception of sinonasal mucosal melanoma.
Therefore, given the undifferentiated nature of SNUC, immunohistochemical analysis is extremely useful. Specifically, positive staining for neuron-specific enolase and chromogranin, cytokeratins 7, 8 and 19, nonreactive to S-100 and nonexpression of vimentin is fairly characteristic for this malignancy.
These findings suggest that the tumor is of epithelial origin and, in addition, lacks any evidence of neuroendocrine, muscle, melanocyte or leukocyte differentiation.
As SNUC generally is a diagnosis of exclusion, the list of negative staining results is very helpful to rule out lymphoma, melanoma, olfactory neuroblastoma, rhabdomyosarcoma, SMARCB1-deficient carcinoma, large cell neuroendocrine carcinoma and nasopharyngeal carcinoma, the major malignancies included in the differential diagnoses with SNUC.
Overall, the prognosis for patients with SNUC is poor. Since the initial recognition of SNUC as a distinct clinicopathological entity, treatment regimens have evolved to include combined craniofacial resection, radiotherapy and chemotherapy. Despite many advances in oncologic therapeutic regimens for SNUC, the mean survival time remains less than a year following diagnosis.
References:
Eggesbo HB. Cancer Imaging. 2012;10.1102/1470-7330.2012.0015.
Ejaz A and Wenig BM. Adv Anat Pathol. 2005;12:134-143.
Frierson HF Jr, et al. Am J Surg Pathol. 1986;10:771-779.
Goel R, et al. J Nat Sci Biol Med. 2012;doi:10.4103/0976-9668.95986.
Gorelick J, et al. Neurosurgery. 2000;47:750-754.
Jeng YM, et al. Am J Surg Pathol. 2002;26:371-376.
Kim BS, et al. Am J Otolaryngol. 2004;25:162-166.
Mendenhall WM, et al. Am J Clin Oncol. 2006;29:27-31.
Phillips CD, et al. Radiology. 1997;202:477-480.
Pitman KT, et al. Skull Base Surg. 1995;5:269-272.
Righi PD, et al. Am J Otolaryngol. 1996;17:167-171.
Tanzler ED, et al. Head Neck. 2008;doi:10.1002/hed.20748.
For more information:
Ana M. Franceschi, MD, is a neuroradiology fellow at NYU Langone Medical Center.
Munir Ghesani, MD, FACNM, is assistant professor of radiology and director of PET/CT fellowship at NYU Langone Medical Center in New York. He also is a HemOnc Today Editorial Board Member. He can be reached at munir.ghesani@nyumc.org.
Disclosures: Franceschi and Ghesani report no relevant financial disclosures.