Key points
- •
Preoperative computed tomography (CT) is necessary to evaluate the anatomy of the nasal cavity and paranasal sinuses, identify anatomic variations, determine the extent of disease, and identify prior surgical changes, if present.
- •
Surgical complications are often identified at the time of surgery. Some major complications are only evident postoperatively, and many require detailed imaging evaluation. The most common, although overall rare, complications necessitating imaging evaluation are cerebrospinal fluid (CSF) leak, vascular injury, intracranial infection, and orbital injury.
- •
β 2 -transferrin assay use has limited the utility of radionuclide cisternography for the identification of CSF leak.
- •
High-resolution CT is the choice imaging modality in the preoperative evaluation as well as in postoperative complications. CT cisternography is the most widely accepted imaging examination performed to localize and define CSF leaks.
- •
MRI is used for the evaluation of the extent of orbital injury, intracranial injury, characterization of postoperative skull-base defects, and as the imaging modality to further clarify equivocal CT findings. Catheter angiography with possible treatments to include embolization, stenting, or balloon occlusion remains to be the definitive imaging study when vascular injury is suspected, although CT angiography may be an alternative in selected cases.
Application of imaging to reduce functional endoscopic surgery complications
Introduction
In the early 1980s, endoscopic sinus surgery (ESS), now referred to as functional endoscopic surgery (FESS), replaced the external approach procedures and is the treatment of choice for a long list of sinonasal pathologic abnormalities, including inflammatory and neoplastic pathologic abnormalities, as well as skull-base and orbital lesions. The introduction of advanced endoscopes, cutting instruments, imaging techniques, including image guidance, facilitated the surgery and aimed to improve the safety of the procedure. Over the past 3 decades, there has been a steady increase in the number of FESS procedures. Preoperative computerized tomography (CT) is widely accepted as a necessity before entering the surgical theater, because it provides an anatomic roadmap and guidance for the endoscopic procedure.
The close proximity of the surgical site to the orbit and the intracranial compartment, however, continued to associate FESS with a variety of complications. As early as 1929, Mosher opined that intranasal ethmoidectomy was “the easiest way to kill a patient”. One would have expected that the introduction of advanced surgical instrumentation, advances in imaging, and the introduction of image-guided surgery would avoid these complications. However, because of individual anatomic variations and the lack of correlation between subjective and objective findings, FESS remains an unstandardized procedure; this in turn results in a broad range of surgical strategies. The serious complications have significantly decreased throughout the past decades, but continue to occur ( Table 1 ).
| Author [Reference], Year | Number of Patients | Major (%) | Minor (%) |
|---|---|---|---|
| Freedman & Kern, 1979 | 1000 | 0.6 | 2.3 |
| Stankiewicz, 1987 | 90 | 8 | — |
| Stankiewicz, 1989 | 90 | 2 | — |
| Friedman & Katsantonis, 1989 | 255 | 0.9 | 2.1 |
| Levine, 1990 | 250 | 0.7 | 8.3 |
| Wigand & Hosemann, 1991 | 1000+ | 0.1 | — |
| Stammberger, 1991 | 6000 | 0.6 | 1.2 |
| Kennedy, 1992 | 120 | — | — |
| May et al, 1994 | — | — | — |
| Their cases | 2108 | 0.85 | 6.9 |
| Reviewed cases | 2583 | 1.1 | 5.4 |
| Gross et al, 1997 | 1106 | 13.9 | — |
| Li and Xu, 1998 | 1089 | 1.2 | 4.5 |
| Danielsen & Olofsson, 2006 | 1915 | 0.47 | 5.6 |
| Eviatar et al, 2014 | 1190 | 0.31 | 1.37 |
| Krings et al, 2014 | — | — | — |
| Primary FESS | 78,944 | 0.36 | — |
| Revision FESS | 4151 | 0.46 | — |
The objective of this article is to address how the imaging information might be better used to further reduce FESS complications and avoid serious complications. The emphasis is on identifying, defining, and discussing the most pertinent anatomic variations, and postsurgical/traumatic changes that may predispose to complications if encountered during FESS. Also discussed is imaging-related workup for select complications.
Complications
Ever since the introduction of FESS, a multitude of publications has dealt with the complications associated with this procedure. Table 1 summarizes the type and rate of complications since the publication by Stankiewicz in 1987. At that time, he reported a complication rate of 8% in his first 90 patients who underwent FESS, including cerebrospinal fluid (CSF) leaks, meningitis, temporary blindness, and significant hemorrhage. These and other complications generally occur when there is manipulation beyond the confines of the nasal cavity and sinuses. These complications include intracranial and orbital injury. Factors resulting in nonoptimal outcomes, such as recurrence of disease, cephalocele, CSF leak, mucocele formation, and optic nerve injury, are discussed.
In 1989, Stankiewicz reported a 2% complication rate on the following 90 patients who underwent FESS. His experience and the experience of many to follow showed that a learning curve (experience) as well as more focused attention on the imaging information and the knowledge afforded by imaging with respect to anatomic landmarks can reduce complications associated with FESS. In the time interval shown in Table 1 , major complications were reduced from 8% to 0.31%. Furthermore, Krings and colleagues showed that the complication rate was relatively the same in the primary and revision FESS: 0.36% versus 0.46%.
There are multiple other factors discussed in the literature that may result in surgical complications aside from the structural variants of the sinus cavity and the paranasal sinuses. For instance, patients with connective tissue disorders, allergic fungal rhinosinusitis, and cardiac diseases are more likely to suffer from complications than the young and otherwise healthy. The experience of the surgeon has been discussed as a potential risk factor as has the handedness of the surgeon. These factors are discussed in the literature but are beyond the scope of this presentation.
Application of imaging to reduce functional endoscopic surgery complications
Introduction
In the early 1980s, endoscopic sinus surgery (ESS), now referred to as functional endoscopic surgery (FESS), replaced the external approach procedures and is the treatment of choice for a long list of sinonasal pathologic abnormalities, including inflammatory and neoplastic pathologic abnormalities, as well as skull-base and orbital lesions. The introduction of advanced endoscopes, cutting instruments, imaging techniques, including image guidance, facilitated the surgery and aimed to improve the safety of the procedure. Over the past 3 decades, there has been a steady increase in the number of FESS procedures. Preoperative computerized tomography (CT) is widely accepted as a necessity before entering the surgical theater, because it provides an anatomic roadmap and guidance for the endoscopic procedure.
The close proximity of the surgical site to the orbit and the intracranial compartment, however, continued to associate FESS with a variety of complications. As early as 1929, Mosher opined that intranasal ethmoidectomy was “the easiest way to kill a patient”. One would have expected that the introduction of advanced surgical instrumentation, advances in imaging, and the introduction of image-guided surgery would avoid these complications. However, because of individual anatomic variations and the lack of correlation between subjective and objective findings, FESS remains an unstandardized procedure; this in turn results in a broad range of surgical strategies. The serious complications have significantly decreased throughout the past decades, but continue to occur ( Table 1 ).
| Author [Reference], Year | Number of Patients | Major (%) | Minor (%) |
|---|---|---|---|
| Freedman & Kern, 1979 | 1000 | 0.6 | 2.3 |
| Stankiewicz, 1987 | 90 | 8 | — |
| Stankiewicz, 1989 | 90 | 2 | — |
| Friedman & Katsantonis, 1989 | 255 | 0.9 | 2.1 |
| Levine, 1990 | 250 | 0.7 | 8.3 |
| Wigand & Hosemann, 1991 | 1000+ | 0.1 | — |
| Stammberger, 1991 | 6000 | 0.6 | 1.2 |
| Kennedy, 1992 | 120 | — | — |
| May et al, 1994 | — | — | — |
| Their cases | 2108 | 0.85 | 6.9 |
| Reviewed cases | 2583 | 1.1 | 5.4 |
| Gross et al, 1997 | 1106 | 13.9 | — |
| Li and Xu, 1998 | 1089 | 1.2 | 4.5 |
| Danielsen & Olofsson, 2006 | 1915 | 0.47 | 5.6 |
| Eviatar et al, 2014 | 1190 | 0.31 | 1.37 |
| Krings et al, 2014 | — | — | — |
| Primary FESS | 78,944 | 0.36 | — |
| Revision FESS | 4151 | 0.46 | — |
The objective of this article is to address how the imaging information might be better used to further reduce FESS complications and avoid serious complications. The emphasis is on identifying, defining, and discussing the most pertinent anatomic variations, and postsurgical/traumatic changes that may predispose to complications if encountered during FESS. Also discussed is imaging-related workup for select complications.
Complications
Ever since the introduction of FESS, a multitude of publications has dealt with the complications associated with this procedure. Table 1 summarizes the type and rate of complications since the publication by Stankiewicz in 1987. At that time, he reported a complication rate of 8% in his first 90 patients who underwent FESS, including cerebrospinal fluid (CSF) leaks, meningitis, temporary blindness, and significant hemorrhage. These and other complications generally occur when there is manipulation beyond the confines of the nasal cavity and sinuses. These complications include intracranial and orbital injury. Factors resulting in nonoptimal outcomes, such as recurrence of disease, cephalocele, CSF leak, mucocele formation, and optic nerve injury, are discussed.
In 1989, Stankiewicz reported a 2% complication rate on the following 90 patients who underwent FESS. His experience and the experience of many to follow showed that a learning curve (experience) as well as more focused attention on the imaging information and the knowledge afforded by imaging with respect to anatomic landmarks can reduce complications associated with FESS. In the time interval shown in Table 1 , major complications were reduced from 8% to 0.31%. Furthermore, Krings and colleagues showed that the complication rate was relatively the same in the primary and revision FESS: 0.36% versus 0.46%.
There are multiple other factors discussed in the literature that may result in surgical complications aside from the structural variants of the sinus cavity and the paranasal sinuses. For instance, patients with connective tissue disorders, allergic fungal rhinosinusitis, and cardiac diseases are more likely to suffer from complications than the young and otherwise healthy. The experience of the surgeon has been discussed as a potential risk factor as has the handedness of the surgeon. These factors are discussed in the literature but are beyond the scope of this presentation.
Radiographic evaluation of complications in functional endoscopic surgery
Complications of FESS have been divided by severity into minor and major categories. If complications result in additional or more extensive surgery, blood transfusion, a new deficit, or death, they are considered major. These major complications also include CSF leak, optic nerve injury, oculomotor deficits, nasolacrimal duct injury, permanent anosmia, and perioperative hemorrhage.
Multiple factors have been implicated in higher morbidity related to FESS, although accurate correlations have been unclear. The reported incidence of major complications is listed in Table 1 . Although the exact rate is unknown, it is accepted that the most common major complication aside from recurrent symptoms is CSF leak.
CSF leak is a result of skull-base injury during FESS and may result in or be related to rhinorrhea, meningitis, intracranial abscess, CSF fistula, and meningoencephalocele. Clinically, symptoms may initially include headache, nasal discharge, or nasal obstruction. Usually patients with a CSF leak are identified intraoperatively, and identified leaks are successfully repaired. Otherwise, leaks may develop days to years after surgery.
The CT information highlights the anatomic risk factors that may facilitate CSF leak. These factors include asymmetry or low-lying ethmoid roof, a steep ethmoid skull-base angle, and a deep olfactory sulcus. The lateral lamella of the cribriform is the thinnest bone of the skull base and is increasingly prone to injury the longer it is. Increasing angulation of the ethmoid skull base has been shown to increase the risk of intracranial penetration during ethmoidectomy, uncinectomy, and middle turbinectomy. The middle turbinate is anatomically contiguous with the lateral lamella of the cribriform, and therefore, in the course of middle turbinectomy, aggressive manipulation may result in disruption of this bone. Furthermore, when evaluating the CT scan of a patient with prior middle turbinectomy, two areas need more focused attention: (1) the integrity of the lamina papyracea at the site of attachment of the basal lamella (also known as the third lamella), which is a coronally oriented lamella extending from the middle turbinate medially and to the lamina papyracea laterally, and (2) the integrity of the skull base at the attachment of the middle turbinate to the lateral lamella of the cribriform plate.
Prior surgery and the associated anatomic distortions are the highest risk factors for CSF leak. The changes associated with the skull base should be identified before revision surgery. These changes may include dehiscences or disruption in the ethmoid roof, cribriform plate, and the planum sphenoidale.
Recurrent sinonasal inflammation may result in rhinorrhea. The appearance of the fluid, its physical consistency, rate of drainage, and other clinical findings do not allow the surgeon to discriminate CSF leak from various other forms of rhinorrhea. Second, because CSF leaks may occur on a delayed basis, recurrent sinonasal inflammation may present concomitantly with a leak. Therefore, a high clinical suspicion should be present when a post-FESS patient presents with persistent or recurrent rhinorrhea.
If CSF rhinorrhea is suspected after FESS, nasal secretions should be tested for β 2 -transferrin to confirm the presence of a CSF leak. β 2 -Transferrin is a protein highly specific for human CSF and should be absent in other bodily secretions. Only a few milliliters of nasal secretions are needed to detect the protein.
If a CSF leak is suspected, a high-resolution noncontrast CT of the sinuses should be obtained. Modern, multidetector thin-section imaging obtained with the patient supine may be best as the patient is usually more comfortable and less likely to move, which may degrade the image quality. Multiplanar reformations should be evaluated carefully with added attention to the coronal series (the coronal images being perpendicular to the hard palate). The objective is to identify a bony defect. A CSF leak is suspected at a site of the bony defect and is usually associated with a fluid level or focal opacification of the adjacent airspace. The suspected finding should be confirmed on multiple orthogonal planes, and if these changes are present, the defect is likely the location of the leak. One should be vigilant in seeking additional possible locations, because multiple locations of CSF leak may be present. Note that there are mimickers of a CSF leak, such as blood products, recurrent mucosal inflammatory thickening, and congenital or previously acquired dehiscences at the skull base.
In the absence of soft tissue in the sinus cavity at the site of the suspected leak, and when the size of the defect and side of the leak can be correlated, an endoscopic-guided surgical approach may be performed with confidence.
Should the endoscopic evaluation fail to correlate with the CT suspicion, or a soft tissue masslike density is detected at the bone defect, an MRI examination should be the next step. Simply using T1- and T2-weighted MRI may provide a diagnosis; however, high-resolution imaging with thin isovolumetric precontrast and postcontrast images are excellent in identifying the appearance of CSF as well as associated soft tissue, which would indicate the presence of an encephalocele. In the case of an encephalocele, tell-tale signs of this finding include when soft tissue is nondependent and abutting the skull base, when meninges are directly visualized extending into the sinonasal cavity, or when there is identification of brain parenchyma extending contiguously into the sinuses. In cases of meningoenchephalocele, encephelomalacia can be appreciated in the inferior frontal lobe. Enhancement of the meninges may be present but does not always indicate infection. MRI alone is not currently recognized as the primary imaging examination because the bony detail afforded by high-resolution CT is superior to that of MRI.
Although not approved currently for use in the United States, some reports of gadolinium MRI cisternography show high sensitivity for detection of CSF leaks. The limitations of CT and radionuclide cisternography are similar to those encountered with MRI, although MRI may afford increased sensitivity.
In the situation of persistent and active CSF leak, CT cisternography ( Fig. 1 ) may help identify the site of CSF leakage, although may be negative in slow, positional, or low-volume CSF leaks. From a technical point of view, the intrathecal injection of contrast should be monitored with fluoroscopy to make sure that the contrast column flows cephalad at the time of injection and that it is not diluted within the CSF in the lumbosacral thecal sac. Alternatively, a cervical puncture may be performed, recognizing the risks. After injection, the patient should be tilted with the head down at a 60° angle and held in this position for 90 seconds. Subsequently, the patient should be immediately transported to the CT scanning area for an immediate evaluation. The images after injection should be compared with precontrast images, both obtained with high-resolution, thin-section protocols. Detection of bony defects, opacified sinus, tract of CSF from the subarachnoid space to the sinonasal cavity, or contrast within a bony defect, air cell, or nasal cavity indicates a leak. Attention should be paid to the attenuation as measured in opacified airspaces that were present before injection as a slow leak or dilution of contrast-laden CSF into a volume of CSF that may result in more subtle increases in density. If suspected, delayed images should be obtained. All positive findings should be repeatedly compared with the precontrast series because commonly found bony sclerosis and osteitis as well as blood products may be mistaken as CSF, although with careful inspection, may be identified correctly. Because many surgeons are using intraoperative guidance systems for FESS and endoscopic CSF repair, it is advised to consider this before CT cisternography to avoid the necessity of repeat examinations for the navigation protocols. The CT images following injection should be obtained with the technique used for navigation software, including placement of fiducial markers if necessary.
