There are many modalities available for obtaining CT images, and these vary in radiation dose, cost, and image quality. Historically, conventional and single slice scanning was previously performed with the patient in the prone position with neck extension, with images obtained in the coronal plane from the anterior wall of the frontal sinus through the posterior wall of the sphenoid sinus, so-called “direct coronal” image acquisition. Axially acquired coronal reconstructed images on single slice scanners were previously inferior to direct coronal helical images because of “Stair-step” artifacts [6, 9]. However, current generation multi-slice, multichannel CT scanners can acquire submillimeter axial slices with contiguous coverage of the entire volume, allowing excellent quality reconstructed images from the same data set in any desired plane [9, 16]. Scanning in the axial plane with multiplanar reformations in coronal and sagittal planes is now considered the standard of care, which is preferable, as the supine position is more comfortable for the patient, and reconstructed coronal images are free of dental amalgam artifacts [9]. In recent years, the use of cone-beam CT has been expanded to in-office use for sinonasal evaluation. Advantages include patient convenience, decreased cost, and significant reduction in radiation exposure. However, disadvantages include diminished soft tissue visibility, increased propensity for X-ray scatter/artifact, and a concern for over usage when scanning is so readily available [6, 17].

The appearance of CT images is affected by window width and level that are arranged according to Hounsfield units. A Hounsfield unit (HU) is defined as a unit of X-ray attenuation based on water where air has a value of −1,000, water has a value of 0, and compact bone/metallic density has a value of +1,000. To obtain the highest bony definition of the paranasal sinuses, the windows of the scan are set between 1,500 and 2,000 HU with a center, or level, of +100 to +300 HU, so-called bone windows. Separate images reconstructed with soft tissue window settings are helpful in assessing sinonasal masses, fungal sinus disease, or complications of sinusitis [5, 11].

The three main planes of imaging include coronal, axial, and sagittal. Each plane of imaging can be helpful in defining paranasal sinus anatomy. The coronal view optimally demonstrates the OMC, the ethmoid roof, and the relationship of ocular structures, and it best correlates with the endoscopic surgical approach. Axial views are particularly useful in determining the drainage pathway of the frontal sinus and the course of the sphenoid septum. The parasagittal view is indispensible to a surgeon’s ability to assess the complex frontal recess anatomy and is helpful when assessing the pneumatization pattern of the sphenoid. Therefore, for every sinus CT in patients with chronic rhinosinusitis, the study should contain thin section axial bone window images through the sinuses, with thin reconstructed coronal and sagittal reconstructions, as well as axial soft tissue window images reconstructed with increased slice thickness, for improved soft tissue resolution (Fig. 4.3a–d).

There are multiple radiographic staging systems (Lund–Mackay, Kennedy, Harvard, May) that have been used to assess radiographic disease severity. Notably, they are not designed to be used to diagnose CRS. The Lund–Mackay staging system (Table 4.1) is the most commonly used system. Its use has been recommended by the American Academy of Otolaryngology Task Force on Rhinosinusitis for staging and has been found to be useful in determining radiographic cutoffs for a truly “diseased” paranasal sinus CT scan. This system quantifies disease severity on a scale from 0 to 2 for six regions, which are scored bilaterally, for a potential total score of 24 [4]. Despite a widespread consensus that staging for chronic rhinosinusitis would be helpful, the utility of radiographic staging has been difficult to demonstrate [8]. Symptom scores for CRS typically fail to correlate with CT scan findings and thus with CT scan stage [8]. Investigators have suggested that radiographic cutoffs for an abnormal CT scan should be based on a score of 4 or greater in the presence of appropriately related sinonasal symptoms [2]. These cutoffs have shown a sensitivity of greater than 85 % and a positive predictive value greater than 80 % in the diagnosis of CRS [2].

One of the main concerns associated with CT scanning is radiation exposure and its anticipated effects on other organ systems. Cumulative exposure to radiation correlates with an increase in the incidence of malignancies. With respect to sinus CT scans, the primary concerns are thyroid cancer and cataract formation [5, 7, 18]. The exact dose of radiation to which a patient is exposed is dependent on patient size and is also largely a function of the parameters that are often set by the manufacturer of the scanner or the scanning institution. These settings are often higher than necessary to achieve a comparable image and vary greatly across manufacturers and institutions [9, 16, 19, 20]. Although the risk of malignancy or other adverse effect is perceived to be small, the availability of CT imaging technology usage has greatly increased, and over usage is a concern. It has been suggested that up to 2 % of cancers in the US could be attributable to radiation exposure associated with CT imaging [20]. The age of exposure is highly relevant, with younger patients having an increased risk for cancer incidence and mortality [18, 21].

A standard head CT has an estimated effective radiation dose of 2 mSv (chest X-ray – 0.1 mSv), which is equivalent to approximately the natural background radiation an individual would be exposed to in 8 months [22]. The mean dose to the lens when obtaining images in the axial plane has been shown to be 24.5 mGy compared to 35.1 mGy in the coronal plane [9]. There is a threshold dose of 0.5–2 Gy for detectable lens opacities, with cataracts occurring after dosages of 2–10 Gy [9, 18]. Studies have shown that even after multiple CT examinations, the radiation dose to the lens is still well below the known thresholds of clinical damage [5]. The mean dose for the thyroid gland has been shown to be 1.4 mGy in the axial plane compared to 2.9 mGy in the coronal plane [9]. Various studies have documented maintenance of image quality with a reduction in radiation dose [9]. Multi-slice axial imaging reduces radiation exposure when compared to direct coronal imaging, and cone-beam technology is believed to substantially decrease effective dose [9].

MRI has been used as a complementary exam to CT for the evaluation of the paranasal sinuses. Its advantages include superior soft tissue characterization and no radiation exposure. It is not routinely used for the assessment of CRS due to cost, longer imaging time, and lack of bone detail [5–7, 11]. MRI has also been found to be potentially overly sensitive to mucosal thickening. At least 20–40 % of patients undergoing MRI of the head are found to have edematous tissue of the paranasal sinuses as an incidental finding [3]. However, MRI has been shown to correlate very well with CT when paired CT and MR images of the paranasal sinuses were assessed with the Lund–Mackay system. Further studies are needed to establish the role for MRI in the diagnosis of CRS [2, 6].

Indications for MRI include cases of aggressive sinus infection with ocular and/or intracranial complication, developmental lesions (i.e., meningoencephalocele), or for the evaluation of a sinonasal mass [5, 6]. Any sinonasal mass that does not appear typical for benign sinonasal polyposis should undergo further investigation with MRI, as it can discriminate neoplasm and other nonneoplastic masses from retained mucous or inflammatory disease and can occasionally separate mucosal thickening from retained secretions (Fig. 4.4a, b). It is particularly valuable in assessing for dural or orbital invasion or extension [3, 23]. It has also been shown to be useful for the evaluation of patients who have undergone prior osteoplastic flap with obliteration with ongoing symptoms.

A standard MRI protocol includes obtaining thin section T1-weighted (T1w) images before and after gadolinium administration, with fat saturation on the postcontrast images, and T2-weighted (T2w) images in the axial and coronal planes. MRI signals of sinonasal secretions have been related to multiple properties such as viscosity, fat, metal content, and protein concentration. In general, watery secretions appear hypointense on T1w images and hyperintense on T2w images. As the protein concentration increases and free water decreases, the signal intensity on T2w images decreases, presumably due to cross-linking between glycoprotein molecules [11, 12, 24]. Thus, sinonasal secretions in acute sinusitis, which consist of 95 % water and 5 % proteins, usually demonstrate low intensity on T1w and high intensity on T2w images [12]. In contrast, in chronic sinusitis, secretions often have a higher protein concentration and less free water, and this may result in different patterns of MRI signal intensity [5]. Generally speaking, four signal intensity patterns in CRS are possible:

Notably, inspissated secretions or fungal colonization may produce signal voids on T1w and T2w images, which may result in a signal void and appear identical to normally aerated sinuses [12, 23]. Peripheral, linear enhancement is suggestive of infection, whereas central enhancement is more likely consistent with neoplasm [5, 12]. Additionally, tumor, particularly hypercellular malignant tumors are typically relatively hypointense on T2w images compared to benign sinonasal mucosal disease, allowing the combination of T2 and postcontrast images to help differentiate sinonasal neoplasm from benign mucosal disease (Fig. 4.4a, b).

Ultrasound (US) has been used in the diagnosis of rhinosinusitis, usually acute sinusitis, primarily in Europe. Its advantages include a lack of radiation, low cost, and wide availability; however, it has poor sensitivity compared with CT. US is not recommended for the evaluation of sinusitis [7, 11].

Although not a direct cause of CRS, many anatomic variations of normal structures in and around the sinuses exist that may predispose to sinus obstruction and inflammation [11, 26]. Knowledge of variant sinus anatomy will also help to avoid complications during surgery [27]. Some of the more common variants will be discussed further below.

Septal deviation is one of the most common anatomic variations and is present in 20–79 % of the population, but is often not clinically relevant [26, 28]. However, severe deviation may narrow the middle meatus, impacting the OMC [28] (Fig. 4.5). It may also alter the airflow pattern, which may negatively impact ciliary clearance [11]. Pneumatization of the nasal septum may occur, which is also often not clinically significant, but may narrow the sphenoethmoidal recess [28].

A paradoxical middle turbinate has been shown to be present in approximately 3–26 % of the population [26]. It is believed that this anatomic variant can contribute to the development of sinusitis if particularly large [28, 29]. Concha bullosa, an aerated middle turbinate, is relatively common, is often bilateral, and has been shown to be present in up to 24–55 % of patients [26, 28] (Fig. 4.6). An uncinate bulla, a rare entity with a prevalence of less than 5 %, refers to an extension of air cells into the uncinate process [26, 27, 29]. The ethmoid bulla, the largest and most prominent anterior ethmoid air cell, can be enlarged or extensively pneumatized forming a “giant bulla” [5, 28, 29] (Fig. 4.7). These variants may result in narrowing of the ostiomeatal complex, potentially affecting drainage patterns of the maxillary and frontal sinuses, with a combination of anomalies possibly increasing the pathogenic effect more than any single variant would cause [27]. Pneumatization of the inferior and superior turbinates can also occur, but this is relatively rare with an incidence of <3 % and is usually clinically insignificant because of little impact on drainage patterns [26–28].

Haller cells, or infraorbital ethmoid cells, are located along the anterosuperomedial maxillary sinus, just inferior to the orbital floor, and are typically in close proximity to the maxillary sinus ostium (Fig. 4.8). They are estimated to be present in approximately 10–45 % of patients. Bilateral Haller cells are rare. They may contribute to sinonasal inflammatory disease if they result in obstruction of the ethmoidal infundibulum or ostium of the maxillary sinus [5, 26–28].

Ten to thirty percent of people have an accessory maxillary ostium in the posterior fontanelle (Fig. 4.9a, b). This can cause nasal mucous recirculation and result in recurrent sinusitis. Surgically, if the two ostia are close together, it is important to consider joining the natural and accessory ostia to prevent recirculation [5, 13, 28]. Congenital hypoplasia of the maxillary sinus can occur, which will cause lateralization of the uncinate process. The primary risk of surgery in a case of maxillary hypoplasia is inadvertent penetration of the orbit [5, 28].