Between 30 and 50 million people in the United States suffer from inflammatory sinus disease. By comparison, less than 1% of all tumors occur in the sinonasal cavities. Between 10 and 33% of patients with acute sinus inflammatory disease will develop chronic disease, which is largely unresponsive to conservative management.1,2

This paper reviews the anatomy of the sinuses from an imaging perspective, paying special attention to the osteomeatal complex, a key area in the pathogenesis of chronic sinusitis. Normal sinus development and the pathophysiology of inflammatory disease also will be discussed. Computed tomographic (CT) and magnetic resonance imaging (MRI) indications, especially with regard to their role in guiding endoscopic sinus surgery, will be surveyed as well.

Sinus anatomy

The mucosa between the sinuses and nasal cavity is continuous. On average, the mucus blanket of the sinus epithelium turns over every 20 to 30 minutes. The function of the sinuses is to humidify air and capture inspired dust particles. Cilia propel the mucus towards the natural ostium of the sinus despite the presence of any surgically created orifices. As mucus moves towards the pharynx, it is exposed to polymorphonuclear leukocytes and other body defenses.3

The ethmoid sinuses are divided into anterior, middle, and posterior air cells. The nasal cavity is bisected by the nasal septum, comprised of hard and soft components. The hard midline components of the nasal septum are the perpendicular plate of the ethmoid bone, the septal cartilage, and the vomer bone. The lateral wall of the nose is divided by three undulating projections-the superior, middle, and inferior turbinates; only the inferior turbinate is a separate bone from the ethmoid bone. These projections curl inferolaterally and divide the nasal cavity into three distinct regions-the inferior, middle, and superior meati.2,4,5

The osteomeatal complex (OMC) conducts communications between the frontal, anterior and middle ethmoid, and maxillary sinuses, and the middle meatus of the nasal cavity via multiple ostia and infundibula. The frontal sinus drains via a nasofrontal communication to the middle meatus through the frontal recess, a part of the anterior ethmoid complex. The maxillary ostium, at the superior medial antral wall, drains to the posterior ethmoid infundibulum. The basal (ground) lamella divides the anterior and middle from the posterior ethmoids.2,4,5

The infundibulum is a funnel-shaped passage bounded by the uncinate process medially, the maxillary sinus inferiorly, the orbit and ethmoid bulla laterally, and the hiatus semilunaris superiorly. The hiatus semilunaris is the final segment for drainage of the maxillary sinus, just inferior to the ethmoid bulla, following the infundibulum. It is best appreciated in the sagittal or coronal plane (figure 1).

The middle turbinate (part of the ethmoid bone) of the nasal cavity defines the space of the middle meatus (figure 2). It attaches to the cribriform plate superiorly, and to the lamina papyracea (the thin ethmoid bone forming the medial orbital wall) laterally.

The posterior ethmoid air cells are the only structures which drain into the superior meatus. The sphenoid sinus drains into the sphenoethmoidal recess, posterior to the superior turbinate and at the level of the sphenopalatine foramen. The nasolacrimal duct is the only structure which drains into the inferior meatus, just below the inferior turbinate.

A noteworthy anatomic variant in patients undergoing endoscopic sinus surgery is the concha bullosa, or pneumatized middle turbinate, which is present in 4 to 15% of the population (figure 3). There also are four important ethmoid air cell variants. These are the agger nasi, the most anterior ethmoid air cells, which are located near the lacrimal duct; the haller cells, located outside the medial orbital floor on the lateral wall of the infundibulum and on the roof of maxillary sinus (figure 4); the ethmoid bulla air cells at the mid ethmoids, located outside the lamina papyracea at the lateral wall of middle meatus; and the Onodi cells, the most posterior ethmoid air cells; these can surround the optic canals posterolaterally. In 25% of all people, the turbinate curls convex laterally, not medially, forming what is referred to as "paradoxical" turbinates.2-5

Sinus development

Sinus development follows a predictable pattern with age. The maxillary sinus is typically present at birth and completely developed by age 15. Approximately 9% of maxillary sinuses are hypoplastic and 0.4 % are aplastic. The ethmoid sinus usually is very small at birth, but is completely developed by puberty. The sphenoid sinus appears by age three and continues to grow, both posteriorly and inferior to the sella, into adulthood. The frontal sinus appears after age six. Frontal sinuses are aplastic in up to 4% of the general population, and up to 90% in individuals with Down's syndrome.4

In the very young, opacification of an otherwise pneumatized sinus does not necessarily represent pathology. Children under one year of age may have "redundant mucosa" and their normal sinuses may be opaque on imaging.4 For children between the ages of one and three, the significance of soft-tissue opacification in the sinuses is equivocal. However, sinus opacification found in children over age three may be reliably interpreted as abnormal.4

Two common developmental anomalies worthy of note in a discussion of sinus development are choanal atresia and abnormal closure of the anterior neuropore. The oronasal membrane normally perforates by 7th week of fetal development. Choanal atresia, or failure of this membrane to normally perforate, can result in either a bony (85%) or membranous (15%) septation (figure 5). Half of all cases of choanal atresia are associated with other anomalies. The anterior neuropore, located near the optic recess, normally closes at 4-weeks gestation. Failure of this structure to close properly can result in fronto-

nasal, frontoethmoid, or frontosphenoid encephaloceles. Trapping of fetal ectoderm during closure can give rise to ectopic (nasal) glioma, nasal dermoid cysts, and/or sinus tracts along the foramen cecum. These are benign lesions, without cerebrospinal fluid (CSF) collections or true neural tissue.4,6

Inflammatory sinusitis

The osteomeatal complex (OMC) is the central drainage point for the maxillary, ethmoid, and frontal sinuses. Infection can travel to and from the other sinuses by direct spread, or to and from adjacent structures through a system of valveless veins. Examples of the latter include spread of infection from the ethmoid sinus to the subperiosteal lamina papyracea, as well as from a molar tooth abscess to the maxillary sinus (figures 6,7).

The key imaging features in the differential diagnosis of inflammatory sinus disease are as follows. Acute sinusitis is suggested by the presence of air-fluid levels. Chronic sinus disease is more commonly associated with bony sclerosis of the sinus walls. Osteo-

myelitis also can appear sclerotic, but with patchy foci of destruction. Mucoceles, as well as benign processes such as polyps, can cause the bone to appear both expanded and demineralized. The presence of tumor is suggested by a destructive process with a superimposed tissue density mass.4

Of special note, fungal sinus disease frequently can appear markedly dark on T2-weighted MR images, secondary to high fungal mycelial iron, magnesium, and manganese content from amino acid metabolism. Fungal sinusitis also commonly demonstrates punctate calcification related to Ca++PO4 and Ca++SO4 deposition near the mycelium. In contrast, esthesioneuroblastoma is less likely to present with calcific concretions, although this is somewhat controversial in the literature.1,2,4 It is important to distinguish "routine" fungal sinusitis, which represents fungal colonization of the sinuses causing mucosal inflammation and increased secretions, from invasive fungal disease, such as is commonly seen with aspergillosis or mucormycosis in diabetics or immunocompromised individuals (figure 8). Invasive fungal disease is a far more serious condition, often with a poor prognosis despite treatment with antifungal agents or surgery. It is typically characterized by vascular microinvasion, bony and soft-tissue destruction with enhancement on CT or MR imaging, and an insidious, relentless course (figure 9).2

Fungal sinus disease often is diagnosed in patients who fail to respond to routine antibiotic treatments. Imaging, although nonspecific, can be highly suggestive of this diagnosis. When confronted with a high CT attenuation, T1/T2-dark soft-tissue mass in a sinus, the major differential diagnostic possibilities worth considering are inspissated sinonasal secretions or potential polypoid disease. Chronic hyperplastic sinonasal polyposis with inspissated secretions can mimic fungal disease (figure 10). Tumors, such as inverted papilloma, are far less common. A useful rule of thumb is that polyps and acute sinusitis will enhance with the administration of intravenous contrast, whereas inspissated secretions will not.

Normally, sinonasal secretions are composed of 95% water and 5% other proteinaceous macromolecules; therefore, these secretions appear hyperintense on routine T2-weighted MR imaging (figure 11). With chronic obstruction, however, as virtually all free water is eliminated and the secretions become inspissated, both the T1 and T2-weighted MR signal intensity can drop precipitously, causing signal voids that may be indistinguishable from air (figure 12). Thus, MR has the potential to underestimate severe chronic sinus disease.1,2

Mucoceles present as airless, mucoid-filled, expanded paranasal sinuses (figures 13 and 14). On imaging, these expand and demineralize but do not destroy bone. Pathologically, mucoceles are expanding cysts lined by mucosa, with accumulated secretions and desquamation. They are believed to be due to ostium obstruction; howver, a few researchers believe that they are primarily of cystic origin. Infected mucoceles are known as mucopyoceles. Approximately two-thirds of mucoceles occur in the frontal sinuses, one-fourth in the ethmoid sinuses, and one-tenth in the maxillary sinuses; only a small percentage occur in the sphenoid sinuses.2 Mucoceles can be thought of as the end stage of a chronically obstructed sinus; formation causes bony distortion and a remodeling of the osseous structures. On CT and MR imaging, only a thin, uniform rim of normal mucosa should enhance with intravenous contrast administration.2

A polyp is a benign sinonasal mucosal lesion. The etiology of polyps is poorly understood, but all theories regarding their origin have in common repeated bouts of inflammation. IGF-1, insulin growth factor, and many other agents also have been implicated in the creation of polyps. Polyps have both edematous and fibrous stages, and can expand and erode bone when chronic. They have been associated with allergic sinusitis, asthma, cystic fibrosis, Kartagener's syndrome, and nickel exposure, as well as with non-neoplastic hyperplasia of inflamed mucous membranes. Polyposis is demonstrated on CT by enlargement of the sinus ostia due to rounded masses within the nasal cavity. This commonly occurs at the ostia to the maxillary antrum, with extension to the choanal region of the nasal cavity, hence the term "antral-choanal polyps." Expanded sinuses, thinning of the bony trabeculae, and erosive bone changes at the skull base are additional features. Polyps usually enhance peripherally but may also enhance solidly like neoplasms; specific CT and MR findings do not reliably distinguish polyposis from cancer. Fungal infections may not be differentiated from polyps by imaging; tumors, however, may sometimes be differentiated because they show homogeneous MR signal, whereas polyps are more often heterogeneous, with multiple components (figures 15, 16).2

Imaging issues: functional endoscopic sinus surgery

Despite MRI's clear advantages in evaluating soft tissue, coronal CT scanning is the method of choice for visualizing the fine cortical bony structures of the osteomeatal complex and their relationship to the adjacent sinuses. Optimal technique consists of thin (1 to 3 mm) contiguous imaging through this region. Ideal positioning is with the patient prone, head extended, so that maxillary sinus secretions will layer at the antral floor (figure 17). Premedication with antihistamines or steroids will minimize the effects of any reversible soft-tissue disease. Administration of IV contrast agents is not required.3,7

Several surgical options are available for the treatment of chronic sinusitis that is unresponsive to conservative medical treatment. Establishing a nasoantral window can restore drainage of blocked sinus ostia. This allows drainage at the inferior meatus, beneath the inferior turbinate; however, care must be taken during surgery to avoid damage to the nasolacrimal duct. The Caldwell Luc approach (transbuccal maxillorhinostomy), once standard, is no longer commonly used in this era of functional endoscopic sinus surgery. In such surgery, a nasoantral window can be created above the level of the teeth roots via an oral-antral approach. Turbinectomy, though still not an uncommonly performed operation, can cause nasal dehumidification, which predisposes patients to epistaxis. Finally, endoscopic sinus surgery, discussed in detail below, can be used to resect the uncinate process via a hiatus semilunaris approach.3,7,8

The development of "functional" endoscopic sinus surgery (FESS), which allows direct access to the OMC (the key area in pathogenesis of chronic sinusitis), has revolutionized the management of chronic inflammatory sinus disease. Endoscopic surgery has the advantage of restoring the natural drainage pattern of the various sinuses. The principle underlying "functional" sinus surgery is that if the blockage to a sinus ostium is eliminated, mucosal edema and inflammation will resolve, allowing cilia to beat normally, resulting in the re-establishment of normal ciliary clearance. In FESS, therefore, the mucosa is not resected during surgery. The mucociliary flow pattern after the creation of a nasoantral window, for example, remains towards the OMC, above the inferior turbinate. Therefore, a surgical strategy of enlarging the natural ostium, as in FESS, is theoretically more appropriate than alternative surgical approaches.

FESS was first described by Messerklinger9 and Wigand;10 the "functional" concept and the importance of mucociliary clearance evolved from the work of Hilding, Proctor, and Messerklinger.9,11 Some clinicians continue to advocate middle turbinectomy in addition to initial uncinectomy for more complete visualization of the posterior ethmoids and sphenoid rostrum during surgery.3,8

Some indications for endoscopic sinus surgery include repeated bouts of sinusitis that is unresponsive to antibiotics, obstructive or erosive mucoceles, chronic hyperplastic rhinosinusitis, and polyp-related asthma, as well as more acute inflammatory processes such as periorbital cellulitis secondary to ethmoiditis. FESS is not indicated for the treatment of asymptomatic retention cysts.

Possible complications of endoscopic sinus surgery relate either to recurrent inflammatory disease (10 to 20%) or damage to structures bordering the sinus cavities. Bony areas that are at risk include the lamina papyrecea, cavernous carotid arteries near the medial walls of the sphenoid sinus, and cribriform plates. Damage to these structures can result in orbital, vascular, or intracranial injury. Specific examples of such complications include synechia formation, orbital abscess, blindness, injury to the nasolacrimal duct, carotid cavernous fistula, cerebrospinal fluid (CSF) leak, cerebral abscess, or subarachnoid hemorrhage (figure 18). Many of these can potentially be avoided if the anatomy of the region is well understood prior to surgery.

In the context of inflammatory sinus disease, the role of the radiologist is not primarily to help determine the need for endoscopic sinus surgery, but to provide a roadmap for the endoscopist should surgery be performed.

There are no definite radiologic indications for routine postoperative imaging. Surgical success is gauged by the clinical response. Thus, in imaging the sinonasal cavities, the radiologist must attempt to answer the following questions:

Is there an abnormality (e.g., mucosal thickening) or congenital variation which interferes with the normal mucociliary flow out of the ostium? Structures which can be anatomically adjacent to the middle meatus include a concha bellosum, paradoxical turbinate, or deviated nasal septum.

Is there a potential problem that may arise during surgery? Is the carotid artery in the normal position in the sphenoid region? Is the lateral wall of the infundibulum formed by the orbital wall? Are the lamina papyracea and cribriform plate intact? Are the fovea ethmoidalis asymmetric or low?

Are there unusual air cells which could remain obstructed after a "routine" endoscopic approach to the middle meatus? These may include haller cells, agger nasi, and concha bullosum (table 1).

Noninfectious sinusitis: a brief note

Noninfectious inflammatory processes can mimic tumor. Although nonspecific, the differential diagnosis of midline nasal inflammation with associated septal destruction is limited. Wegener's granulomatosis is characterized by necrotizing vasculitis and granulomas of the upper and lower respiratory tract and kidneys (figure 19). Idiopathic midline granuloma occurs in the nasal septum and is a lymphoreticular, pre-lymphomatous condition; the treatment of choice is radiation.4 Cocaine-induced nasal perforation often is accompanied by a nasal septal granuloma and destructive mass (figure 20); it may resemble idiopathic midline granuloma but, unlike that entity, is typically treated with antibiotics.4 Also included in the differential diagnosis of inferior meatus sinus lesions are lacrimal sac lesions, such as Wegener's granulomatosis and nasolacrimal duct mucoceles. AR

References

1. Som PM: Head and neck imaging, Part I: An overview. In: RSNA Special Course in Neuroradiology, pp 139-146. Oak Brook, RSNA Publications, 1994.

2. Som PM, Brandwein M: Sinonasal cavities: Inflammatory diseases, tumors, fractures, and postoperative findings. In: Som PM, Curtin H (eds): Head and Neck Imaging, ed 3, pp 126-318. St. Louis, Mosby, 1996.

3. Mafee MF: Preoperative imaging anatomy of nasal-ethmoid complex for functional endoscopic sinus surgery. In: Mafee MF (ed): The Radiologic Clinics of North America, pp 1-20. Philadelphia, WB Saunders Co., 1993.

4. Harnsberger H: Head and Neck Imaging, pp 377-420. Chicago, Year Book Medical Publishers, Inc., 1990.

5. Valvassori G, Mafee MF, Carter B, (eds): Imaging of the Head and Neck, pp 248-328. New York, Thieme Medical Publishers Inc., 1995.

6. Naidich TP, Zimmerman RA, Bauer BS, et al: Midface: Embryology and congenital lesions. In: Som PM, Curtin H (eds): Head and Neck Imaging, ed 3, pp 3-60. St. Louis, Mosby, 1996.

7. Zinreich SJ, Kennedy DW, Rosenbaum AE, et al: Paranasal sinuses: CT imaging requirements for endoscopic surgery. Radiology 163:769-775, 1987.

8. Hudgins PA: Complications of endoscopic sinus surgery: The role of the radiologist in prevention. In: Mafee MF (ed): The Radiologic Clinics of North America, pp 21-32. Philadelphia, WB Saunders Co., 1993.

9. Messerklinger W: Endoscopy of the nose. Baltimore, Urban and Schwartzenberg, 1978.

10. Wigand ME, Stein W, Jaumann MT: Endonasal sinus surgery with endoscopical control: From radical operations to rehabilitation of the mucosa. Endoscopy 10:255-260, 1978.

11. Hilding AC: Physiological basis of nasal operation. Calif Med 72:103-107, 1950.