Fax: (212) 427-8137
Imaging the Paranasal Sinuses: Where We Are and Where We Are Going
Article first published online: 24 OCT 2008
Copyright © 2008 Wiley-Liss, Inc.
The Anatomical Record
Special Issue: The Paranasal Sinuses: The Last Frontier in Craniofacial Biology
Volume 291, Issue 11, pages 1564–1572, November 2008
How to Cite
Fatterpekar, G. M., Delman, B. N. and Som, P. M. (2008), Imaging the Paranasal Sinuses: Where We Are and Where We Are Going. Anat Rec, 291: 1564–1572. doi: 10.1002/ar.20773
- Issue published online: 24 OCT 2008
- Article first published online: 24 OCT 2008
- Manuscript Accepted: 23 APR 2008
- Manuscript Received: 22 APR 2008
- paranasal sinus;
As has happened in all facets of neuroimaging, cross-sectional imaging has dramatically changed our approach and understanding of the anatomy and pathology of paranasal sinuses. We have moved away from plain film radiographs to modern high-resolution sinus computerized tomography (CT) and magnetic resonance imaging (MRI) that helps us better depict underlying normal anatomy and evaluate pathology. Recent advances in PET/CT imaging have introduced a physiologic aspect to anatomical imaging and holds promise to better stage and restage head and neck tumors. In this article, we describe the various imaging techniques, concerns, advantages and disadvantages of the individual techniques, and provide an overview of the various pathologies involving the paranasal sinuses. Anat Rec, 291:1564–1572, 2008. © 2008 Wiley-Liss, Inc.
Of primary concern to the radiologist evaluating the paranasal sinuses and nasal fossae is identification of any osseous changes or variations, noting the presence of abnormal soft-tissue disease and its possible extension beyond the sinonasal cavities, and the characterization of this disease. Available imaging techniques include computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography/computed tomography (PET/CT). Each of these modalities offers certain advantages, and each has disadvantages when compared with the other techniques.
Notably, plain films are no longer considered to be a part of the primary imaging armamentarium. At best, they give only an overview of the anatomy and underlying pathology, as they are limited to displaying three-dimensional structures in a two-dimensional plane. CT and MR imaging have the advantage of being able to show fine anatomic detail in serial tomographic sections, and thus eliminating the gross volume averaging inherent in plain films (Dodd and Jing, 1977; Harnsberger, 1990; Som, 2003; Standring et al., 2004). In fact, in most cases, after a plain film study shows that disease is probably present, a CT or MR imaging study is then routinely obtained.
In the discussion later, we describe the imaging techniques used to evaluate paranasal sinuses with a brief description of the pathologies affecting the sinonasal cavities.
MATERIALS AND METHODS
At our institution, paranasal sinuses are primarily evaluated with CT. MR is used to evaluate tumors and to assess for extension of an infectious process beyond the paranasal sinuses into the adjacent soft tissues. PET/CT is used for staging and restaging of head and neck tumors.
Traditionally, CT imaging of the sinus has been acquired in the axial and coronal planes, using noncontrast high-resolution 3-mm thick contiguous scans. Axial images are obtained with the patient supine on the scanning table and maintaining neutral position of the scanning gantry. This differs from the coronal scans, which are enabled by extension of the patient's neck in either prone or supine position and angling of the scanning gantry to approximate the sinus coronal plane. An increasing number of institutions have abandoned the separate coronal acquisition, because the very thin overlapping sections obtained on newer multidetector scanners can be reformatted to nearly the same quality as a native coronal acquisition. The coronal imaging plane offers best visualization of the drainage pathways of the sinuses, whereas some drainage pathways (such as sphenoid sinus ostia) and sinus walls oriented close to the coronal plane are better seen on axial images.
The initial scanning data are typically reconstructed with two different imaging algorithms (Figs. 1 and 2). The bone, or edge, algorithm enhances the interface between tissues of substantially differing densities, so that osseous margins and intact bone are easily distinguished from demineralized or eroded bone. However, this bone algorithm causes artifactual noise in structures of similar density, such as mucosal thickening of the sinus margin. Therefore, soft-tissue algorithm images are also generated to eliminate this artifactual noise in homogeneous structures and allow better visualization of soft-tissue structures and abnormalities. Because evaluation of both bone and soft tissue is crucial in the assessment of sinuses, both algorithms are scrutinized for evidence of pathology.
Scans enhanced with iodine-based intravenous contrast are obtained only in patients who are acutely ill and suspected of having a complication of acute sinusitis, such as subperiosteal abscess, epidural or subdural empyema, or osteomyelitis, or when malignant disease of the paranasal sinuses or nasal cavity is suspected.
Of particular concern in any CT imaging of the head and neck region is the radiation dose delivered to the lens and to the thyroid gland. The most important determinant of the radiation exposure is the milliampere-second (mAs), which correlates with the quantity of X-ray photons projected through the patient. Depending on the imaging protocol, several studies have reported lens dose to vary widely, from 1.88 to 64 mGy. These single scan measured doses are still much lower than the 0.5–2 Gy threshold for lens damage, and very much lower than the cumulative dose of 6–14 Gy (equivalent of about 200 scans) that has been shown to significantly increase the risk of cataract formation. Thus, even patients who undergo multiple scans are at a very low risk for premature cataract formation. Also, with newer multidetector CT scanners, the dose has further been lowered. This is because the newest generation of CT scanners has intrinsically better spatial and contrast resolution and more dose-efficient detectors, thereby maintaining image quality at lower mAs (International Commission on Radiological Protection, 1990; Sillers et al., 1995; Bernhardt et al., 1998; Nishizawa et al., 1998; Bassim et al., 2005). Although imaging literature reveals no clear threshold dose above which the risk for thyroid cancer increases, the average dose of ∼4 mGy obtained per paranasal sinus study is not considered to cause any increased incidence (Zammit-Maempel et al., 2003). In children, however, where a linear dose-response relationship is speculated, it is recommended that scans be obtained at a lower mAs setting. Further, in these patients, one may lower the dose further by reconstructing coronal images from multidetector helical data, eliminating the radiation exposure of the direct coronal acquisition (Babbel et al., 1991; Mulkens et al., 2005).
Although CT is ideal for assessing the osseous margins of the paranasal sinuses, the inherently superior soft-tissue resolution and multiplanar capabilities render MRI superior for assessment of soft-tissue masses and extension of infectious/malignant disease processes beyond the paranasal sinuses. MR imaging of the paranasal sinuses must include high-resolution (3 mm) T1-weighted and T2-weighted images, not only of the sinonasal cavity but also of the orbit, skull base, and the adjacent intracranial compartment. Images should be acquired in both the axial and coronal planes, although sagittal or arbitrary oblique planes can be added as necessary. In addition, contrast-enhanced T1-weighted images are routinely obtained (Fig. 3). Gadolinium chelate contrast agents (instead of the iodine-based agents in CT) are used in MRI to evaluate enhancement on MRI. The seven unpaired electrons in gadolinium's outer orbital ring yield a very high magnetic moment, and, consequently, areas of extravasation or high vascularity appear hyperintense (“bright”) on T1-weighted imaging. Because fat will also appear hyperintense on T1-weighted images, fat-saturated T1-weighted techniques are usually included to increase the sensitivity for enhancement, and thus improve detection of local disease extent and presence of disease beyond the paranasal sinuses (e.g., perineural tumor spread or intracranial extension) (Hudgins, 1996).
As sinonasal inflammatory disease is probably the most common disease affecting man, it is most likely that, when a patient comes for a sinonasal imaging study, the referring clinician wants to assess extent of disease and identify any complicating factors such as bone involvement or normal anatomic variants that might affect surgery (Laine and Smoker, 1992; Yousem, 1993; Zeifer, 2000). As a result, the radiologist is equally interested in imaging bone and soft tissue structures. As mentioned, this is best accomplished with the use of noncontrast CT imaging, as the same image data set can be easily manipulated to highlight either bone detail or the adjacent soft tissues.
On noncontrast CT, certain soft-tissue collections have high attenuation (density) when compared with the appearance of muscle (Fig. 4). The primary diagnoses to consider in this situation are desiccated secretions within an obstructed sinus, a fungal mycetoma (aspergilloma), and hemorrhage within a sinus cavity. Importantly, tumor does not have such high attenuation on routine noncontrast scans.
Of these diagnostic possibilities, history and identification of an associated fracture can almost always confirm whether the density indicates blood (Fig. 5) (Zilkha, 1982; Zinreich, 1990). In addition, history compatible with hemorrhage may include not only recent trauma but also bleeding diseases such as acute leukemia and Von Willebrand's disease that tend to bleed from mucosal surfaces. At the present time, it is virtually impossible to distinguish between desiccated secretions and fungal disease when using noncontrast CT, contrast-enhanced CT, or MR imaging, with or without contrast. Infiltrative changes in adjacent bone may be the only helpful imaging sign to aid in identification of the more aggressive fungal disease (Fatterpekar et al., 1999).
Notably, as mentioned, tumor will not have high attenuation comparable to desiccated secretions, fungal ball, or blood. Thus, if soft tissue fullness has high attenuation on a noncontrast CT study, in almost every case, tumor can be eliminated from the differential diagnosis. If one obtains a contrast-enhanced CT study, inflammatory tissues will enhance. Also, some physiologic enhancement will be seen, such as in the frequently changing mucosal congestion pattern known as the normal nasal cycle. In addition, tumor will enhance. Thus, the preliminary distinction of inflammatory tissues from tumor is ordinarily difficult or impossible (Som et al., 1987).
If one is interested in potential intracranial extension of sinonasal disease, this will be better seen, in most cases, on contrast-enhanced MR images (Som et al., 1989). If the patient cannot have an MR study (because of a pacemaker, implanted metallic fragments such as shrapnel, metallic implants, vascular clips, or claustrophobia), a contrast-enhanced CT scan then becomes the best available modality for assessing intracranial structures.
In general, when MR imaging is compared with CT, MR better distinguishes normal and inflamed soft tissues and better differentiates between these tissues and tumor. This is primarily based on the fact that compared with the normal sinonasal (Schneiderian) mucosa, inflamed mucosa is associated with increased submucosal edema and increased surface secretions, both of which are initially 95% water. Thus, on T1-weighted MR sequences, water has a long relaxation time that is seen as low signal intensity (“dark”). On T2-weighted images, water also has a long relaxation time, which is seen as high signal intensity (“bright”). That is, on MR imaging, inflamed sinonasal soft tissues are characterized by low T1-weighted and high T2-weighted signal intensities (Som et al., 1990). The thin, uniformly smooth appearance of normal sinonasal mucosa can also help distinguish normal mucosa from the thicker, often polypoid configuration of inflamed mucosa (Fig. 6) (Som, 2003).
By comparison, tumor is more cellular than normal or inflamed tissue. Also, in general, tumor is less differentiated than normal tissue and this is reflected as a lower serous and mucinous content. Thus, on MR imaging, tumors tend to have low T1-weighted signal intensity and low to intermediate T2-weighted signal intensity (Fig. 7). It is because of this that T2-weighted images are best to distinguish inflamed mucosa from adjacent tumor (Loevner and Sonners, 2004). Unfortunately, early infiltration of bone may be impossible to identify on MR in areas where the bone is thin (as is the case in most facial bone structures). In thick bone, where a good medullary cavity can be better seen on MR imaging, tumor infiltration is more easily resolved. Thus, subtle early bone erosion, or erosion of thin bone, will be better seen on CT, on which the bone is directly imaged.
Although bone in the floor of the anterior cranial fossa may appear grossly intact on CT and noncontrast MR images, these modalities have limited sensitivity for early or mild intracranial spread of tumor and/or inflammation (Fig. 8). Such disease is best identified on contrast-enhanced, fat-suppressed, T1-weighted images. Fat suppression is needed to eliminate any high signal intensity from adjacent fat (i.e., fatty marrow, etc.) that might be confused with or obscure actual enhancement. In addition, secretions and CSF, both of which appear hyperintense on T2-weighted imaging and could thus obscure adjacent pathologic high signal, are both lower in signal intensity on T1-weighted images and will not obscure enhancement (Kaufman et al., 1983). Thus, as mentioned, in cases of sinonasal tumor, a thorough MR examination includes T1-weighted, T2-weighted, and T1-weighted, contrast-enhanced, fat-suppressed sequences. Usually axial and coronal images are obtained, but sagittal and oblique image planes can be easily obtained if they are thought to highlight the disease.
As with most imaging, perfection is rarely obtained, and distinction between tumor and some inflammatory tissues, such as granulation tissue, may still elude the clinician and radiologist on even the most thoroughly performed CT and MR studies. In such cases, a combined PET/CT examination may provide additional useful information (Schoder et al., 2004).
At present, clinical PET imaging relies on [18F] fluoro-deoxy-glucose (FDG) uptake. The preferential uptake of FDG by tumor cells results from the increased transport of FDG into the cell because of an increase in the glucose transporter GLUT1. Cancer cells have markedly elevated the levels of hexokinase, which quickly phosphorylates the FDG, preventing it from leaving the cell. As the hydroxyl group is not present on carbon-2 of FDG, the next enzyme in the normal glycolytic chain (glucose phosphate isomerase) cannot metabolize FDG as it would ordinary glucose. As a result, FDG accumulates within the cell. Thus, FDG accumulation is related to glycolytic activity, which is dramatically increased in cancer cells compared with normal cells. This increased glycolysis in tumor cells is referred to as the Warburg effect.
After injection with radioactive FDG, the patient waits for 45–60 min in a darkened room with minimal external stimuli to allow for differential accumulation of the FDG within the tumor compared with normal cells. Acquisition of CT images at the same session as the PET scan allows accurate registration of the two data sets, and areas of increased FDG activity may be accurately mapped on the CT study for precise localization.
Unfortunately, inflammatory cells also exhibit increased glycolytic activity. As a result, one cannot be certain that areas of FDG uptake necessarily represent tumor. If the increased uptake can be localized within a morphologically identifiable mass on the CT scan, tumor can be diagnosed with more certainty (Fig. 7).
There are several new agents being studied that have been shown to accumulate only within tumor. One such agent, fluoro-deoxy-thymidine (FLT), has shown considerable promise in studies performed at multiple centers. When DNA is damaged, thymidine can be generated by phosphorylation of deoxythymidine by thymidine kinase 2 (TK2), a nuclear enzyme that is always present. However, if the cell is about to replicate, an entirely new set of DNA is required. To meet this increased demand, cytoplasmic thymidine kinase 1 (TK1) is made by the cell. FLT gets into the cell by both passive diffusion and a carrier-mediated nucleoside transport mechanism. Once inside the cell, it is rapidly phosphorylated by TK1 and, after further phosphorylation, it can be utilized in the production of DNA. A pathway such as this that uses so-called “salvage” enzymes, which are produced only to meet cell cycle demand, is referred to as the salvage pathway. Thus, as with FDG, the accumulation of the radioactive marker is a reflection of the phosphorylating enzyme activity, which in this case is TK1. The imaging significance of this pathway is that FLT should accumulate only in actively replicating cells and not in inflammatory cells.
Today, the radiologist can provide the clinician with more accurate and thorough information than at any prior time. The precise identification of normal variants that could cause operative complications is routinely made. The localization of disease can be accurately accomplished and in almost all cases, distinction can be made between inflammatory and malignant disease. In those cases that still require biopsy, the imaging studies provide the clinician with easy visualization of the best biopsy route. As newer agents become available, even quicker and more precise identification of tumor will be available utilizing PET/CT.
- 1991. Optimization of techniques in screening CT of the sinuses. AJNR Am J Neurorad 12: 849–854. , , , , .
- 2005. Radiation dose to the eyes and parotids during CT of the sinuses. Otolaryngol Head Neck Surg 133: 531–533. , , , .
- 1998. CT scanning of the paranasal sinuses: axial helical CT with reconstruction in the coronal direction versus coronal helical CT. Br J Radiol 71: 846–851. , , , , , .
- 1977. Radiology of the nose, paranasal sinuses and nasopharynx. Baltimore: Williams & Wilkins. p 59–65. , .
- 1999. Fungal diseases of the paranasal sinuses. Semin Ultrasound CT MR 20: 391–401. , , , , .
- 1990. Imaging for the sinus and nose. In: HarnsbergerR, editor. Head and neck imaging handbook. St. Louis: Mosby-Year Book. p 387–419. .
- 1996. Sinonasal imaging. Neuroimaging Clin N Am 6: 319–331. .
- International Commission on Radiological Protection. 1990. Recommendations of the ICRP. Oxford: Pergamon Press, Publication 60; Vol. 21.
- 1983. Sinusitis induced subdural empyema. Neurology 33: 123–132. , , .
- 1992. The ostiomeatal unit and endoscopic surgery: anatomy, variations, and imaging findings in inflammatory diseases. AJR Am J Roentgenol 159: 849–857. , .
- 2004. Imaging of neoplasms of the paranasal sinuses. Neuroimaging Clin N Am 14: 625–646. , .
- 2005. Comparison of effective doses for low-dose MDCT and radiographic examination of sinuses in children. AJR Am J Roentgenol 184: 1611–1618. , , , , .
- 1998. Dose evaluation and effective dose estimation in radiological studies of paranasal sinuses. Radiat Prot Dosim 82: 271–276. , , , et al.
- 2004. Head and neck cancer: clinical usefulness and accuracy of PET/CT image fusion. Radiology 231: 65–72. , , , , .
- 1995. Radiation exposure in paranasal sinus imaging. Otolaryngol Head Neck Surg 112: 248–251. , , .
- 2003. Sinonasal cavities. In: SomPM,CurtinHD, editors. Head and neck imaging,4th ed. Mosby. St. Louis, Missouri. p 1–438. .
- 1990. Hypointense paranasal sinus foci. Differential diagnosis with MR imaging and relation to CT findings. Radiology 176: 777–781. , , , , .
- 1989. Benign and malignant sinonasal lesions with intracranial extension: differentiation with MR imaging. Radiology 172: 763–766. , , , , .
- 1987. CT appearance distinguishing benign nasal polyps from malignancies. J Comput Assist Tomogr 11: 129–133. , , , .
- StandringS,HealyJ,JohnsonD,WilliamsA,EllisH, editors. 2004. Nose, nasal cavity and paranasal sinuses: head and neck. Gray's anatomy: the anatomical basis of clinical practice (Gray's anatomy),39th ed. Churchill Livingstone, Edinburgh, UK.
- 1993. Imaging of sinonasal inflammatory disease. Radiology 188: 303–314. .
- 2003. Radiation dose to the lens of eye and thyroid gland in paranasal sinus multislice CT. Br J Radiol 76: 418–420. , , .
- 2000. Pediatric sinonasal imaging: normal anatomy and inflammatory disease. Neuroimaging Clin N Am 10: 137–160. .
- 1982. Computed tomography in facial trauma. Radiology 144: 545–548. .
- 1990. Paranasal sinus imaging. Otolaryngol Head Neck Surg 103: 863–868. .