Coronal computed tomographic scan showing extensive unilateral left nasal cavity soft tissue thickening of the septum, turbinates, and nasal floor, with complete opacification of the left nasal cavity, in a patient (patient 9) with invasive fungal sinusitis. Right nasal cavity does not display soft tissue thickening.
Coronal computed tomographic scan showing unilateral nasal cavity and sinus soft tissue thickening of the septum, nasal floor, and lateral wall (arrows) in a patient with invasive fungal sinusitis. The patient (patient 23) had a maxillary antrostomy previously, not associated with his diagnosis of invasive fungal sinusitis.
Bone erosion of the left lamina papyracea (arrow) with soft tissue thickening of the lateral nasal and medial canthal areas in a patient with invasive fungal sinusitis (patient 4). Also note the extensive unilateral nasal cavity soft tissue thickening on the left side.
Axial soft tissue window demonstrates thickening of the left retroantral fat pad (posterior arrow) and premaxillary soft tissue (anterior arrow) in a patient (patient 23) with invasive fungal sinusitis.
Coronal computed tomographic scan of a control patient demonstrating bilateral maxillary air-fluid levels, ethmoid soft tissue thickening, and very mild thickening of the mucosa of the nasal floor (black arrows) and septum (white arrow). Note that the degree of nasal cavity soft tissue thickening is much less than in the patients with invasive fungal sinusitis.
Coronal computed tomographic scan of control patient demonstrating bilateral ethmoid and maxillary opacification but minimal thickening of the mucosa of the lateral nasal wall (white arrow) and floor (black arrow) and septum.
DelGaudio JM, Swain RE, Kingdom TT, Muller S, Hudgins PA. Computed Tomographic Findings in Patients With Invasive Fungal Sinusitis. Arch Otolaryngol Head Neck Surg. 2003;129(2):236-240. doi:10.1001/archotol.129.2.236
Copyright 2003 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2003
To determine the radiographic findings of computed tomographic (CT) imaging most suggestive of invasive fungal sinusitis (IFS) in an immunocompromised patient population.
A retrospective review of patients with a diagnosis of IFS reached with CT and confirmed by histopathologic evaluation.
An academic tertiary care hospital.
Twenty-three immunocompromised patients with confirmed IFS and preoperative CT imaging. Controls were 10 patients with acute myelocytic leukemia and CT evidence of sinusitis but no history of IFS.
The CT scans were reviewed to identify factors predictive of invasive fungal disease. Parameters evaluated were nasal cavity and sinus soft tissue thickening, the presence of air-fluid levels, bone erosion, extrasinus extension, and unilateral or bilateral nasal cavity and sinus involvement.
The CT findings included severe soft tissue edema of the nasal cavity mucosa (turbinates, lateral nasal wall and floor, and septum) in 21 of the 23 patients, sinus mucoperiosteal thickening in 21, bone erosion in 8, orbital invasion in 6, facial soft tissue swelling in 5, and retroantral fat pad thickening in 2. Two patients had air-fluid levels. No patients had intracranial involvement. Unilateral involvement was found in 21 patients, and bilateral involvement in 2. Review of the control group revealed only mild soft tissue edema of the nasal cavity in 2 (P<.001), unilateral involvement in 2 (P<.001), and evidence of bone erosion or extrasinus soft tissue involvement in none.
Most patients do not have classic CT findings of bone erosion or extrasinus extension in the early course of IFS. We found that severe unilateral thickening of the nasal cavity mucosa was the most consistent finding on CT suggestive of underlying IFS, occurring much more frequently in immunocompromised patients with IFS than without IFS. Even though severe nasal cavity soft tissue thickening is much more common in IFS, this is a nonspecific finding that can be seen, to a lesser degree, in all forms of rhinosinusitis. Therefore, the clinician cannot rely solely on CT imaging and must maintain a high index of suspicion when evaluating immunocompromised patients to establish a prompt diagnosis. Early nasal endoscopy with biopsy and initiation of appropriate therapy are necessary to improve prognosis.
INVASIVE FUNGAL infection of the sinonasal cavities is a disease requiring urgent attention and treatment by the otolaryngologist–head and neck surgeon, but it is difficult to diagnose and treat. Essentially, the patients are immunocompromised as a result of systemic chemotherapy and/or bone marrow transplantation, immunosuppressive agents used after solid organ transplantation, acquired immunodeficiency syndrome, or long-term systemic steroid use. Poorly controlled diabetes mellitus, usually with diabetic ketoacidosis, is another common underlying factor in the development of invasive fungal sinusitis (IFS). Protein calorie malnutrition and excessive stored iron can also be predisposing factors in the development of IFS.1 Although uncommon, IFS has been documented in otherwise healthy individuals.2,3
Most patients with IFS are already in poor overall physical health as a result of their underlying condition or treatment, and are expected to have a higher mortality rate and poorer prognosis. Leukopenia, especially neutropenia, significantly reduces the inflammatory response, which increases the risk of IFS and may mask some of its early symptoms. These factors contribute to the difficulty in diagnosing and treating the disease, which can progress rapidly and with devastating results. Therefore, early diagnosis is necessary. Treatment usually consists of aggressive surgical debridement and intravenous antifungal therapy.
The symptoms of IFS, which was first described in the French medical literature in 1885,4 overlap with those of bacterial sinusitis. They include nasal congestion or obstruction, nasal discharge, facial pain, facial swelling, headache, dental pain, and visual complaints. Anesthesia of the nasal mucosa and/or facial skin is more suggestive of an invasive process such as IFS than of bacterial infection. On physical examination, the most common rhinologic finding is ischemic mucosa. Black eschar is classically described as the hallmark of IFS, but this is a late finding resulting from vascular thrombosis and tissue necrosis. Earlier findings include pale, ischemic mucosa that does not bleed normally with manipulation, mucosal edema, and pain out of proportion to physical findings, likely due to tissue ischemia. Extrasinus extension is usually a late finding. Orbital involvement can present with limitation of extraocular movement, proptosis, and decreased vision. Rapid progression of the disease can occur with involvement of the facial skin, hard and soft palates, retroantral area, and intracranial extension.
A classification system with criteria for the diagnosis of IFS has been proposed.1 de Shazo et al1 describe 3 forms of IFS, which they classify as follows: granulomatous, acute fulminant, and chronic invasive. They propose that IFS is diagnosed by radiographic evidence of sinusitis and by histopathologically proven hyphal forms within sinus mucosa, submucosa, blood vessels, or bone.1 Multiple fungal species have been identified in patients with IFS, most commonly Aspergillus and fungi of the order of Mucoraceae, including Rhizopus and Mucor.1,5
Review of the otolaryngology literature reveals that computed tomographic (CT) findings such as edema of nasal cavity soft tissue, sinus mucoperiosteal thickening, bone erosion, orbital invasion, facial soft tissue swelling, and periantral soft tissue infiltration are all findings suggestive of IFS,1,5- 7 but no radiographic findings have been found to be pathognomonic of IFS. The purpose of this study was to retrospectively review CT scans of patients with histologically proven IFS to determine if specific radiologic characteristics could establish the diagnosis.
A retrospective analysis of patients with a diagnosis of IFS between 1995 and 2001 in the Emory University Hospital system was performed. Only patients with a tissue diagnosis of IFS who, prior to biopsy or debridement, had undergone a CT scan of the sinuses that was available for evaluation were included in this review. Twenty-three patients (12 men and 11 women) met these criteria. Demographic data, underlying diagnosis, identified fungal organism, absolute neutrophil counts, and patient outcomes were reviewed (Table 1). Most of the patients with leukemia were in the early postoperative period from bone marrow transplantation. All CT studies were reviewed by 1 head and neck radiologist and 2 otolaryngologists. Of the initial 23 scans performed, 11 were coronal views alone, 11 were both axial and coronal views, and 1 was only an axial view. All of the scans were performed with 3-mm-thick sections. Three of the CT scans were performed with contrast, the remaining without contrast. The CT scans were evaluated with both bone and soft tissue window settings.
All radiographic and clinical material was retrospectively evaluated. Each CT scan was evaluated for nasal cavity mucosal and soft tissue thickening, sinus mucoperiosteal thickening, osseous erosion, extrasinus extension (ie, orbital and intracranial invasion), facial soft tissue thickening, retroantral fat pad thickening, and unilateral vs bilateral involvement. All 23 patients had been taken to the operating room for nasal endoscopy, histopathologic confirmation of IFS, and debridement.
A control group of 10 immunocompromised patients was chosen from an acute myelogenous leukemia database. All of them had a sinus CT scan that showed evidence of sinusitis, but none had a history of IFS. Each was treated for bacterial sinusitis and improved. Their ages ranged from 28 to 70 years, with a mean of 51.8 years.
All 23 patients in the study group were immunocompromised and diagnosed as having acute fulminant IFS. The fungal pathogens identified were Aspergillus or Mucormycosis (Table 1). Absolute neutrophil counts were less than 500/µL in 11 patients and greater than 500/µL in 12 patients. Four patients with persistent IFS died of their underlying medical condition during their hospitalization, and each of them had an absolute neutrophil count of less than 500/µL (Table 1). Nineteen of the 23 patients were discharged from the hospital after resolution of IFS, with appropriate follow-up from an otolaryngologist. One patient who survived had permanent visual loss.
Of the CT findings evaluated, severe thickening of the nasal cavity mucosa and soft tissues, including the turbinates, septum, and nasal floor, was the most common finding, noted in 21 (91%) of the 23 patients (Figure 1 and Figure 2). The nasal cavity soft tissue thickening was unilateral in 21 patients (91%), and was greater than the degree of thickening expected in acute bacterial rhinosinusitis. Sinus mucoperiosteal thickening or opacification was present in 21 (91%), but only 2 had air-fluid levels. Bone erosion was seen in only 8 (35%) (Figure 3). Orbital invasion was found in 6 patients (26%), and 5 (22%) had soft tissue thickening of the face overlying the maxillary sinus. Three of these 5 patients had CT scans with contrast while the other 2 had inflammatory changes visualized with a soft tissue algorithm (Figure 4). Two patients were identified with infiltration of the retroantral fat pad without evidence of bone erosion (Figure 4). No patient in this series had intracranial involvement.
All patients were taken to the operating room for biopsy and/or debridement. The most common findings at surgery were mucosal edema and/or pale mucosa. The edematous tissue was hypervascular due to acute inflammation, whereas the pale mucosa had little or no bleeding as a result of tissue ischemia. The presence of necrotic tissue and eschar was an uncommon finding.
In the control group, only 2 of 10 patients had nasal cavity soft tissue thickening (Figure 5 and Figure 6), and in these 2 patients the degree of thickening was not nearly as severe as seen in the patients with IFS (Figure 1, Figure 2, Figure 3, and Figure 4). Only 2 of the patients had unilateral sinonasal involvement. No evidence of bone erosion or extension of infection beyond the sinuses was identified in this control patient group. The differences between the 2 groups in the frequency of nasal cavity soft tissue thickening and of unilateral involvement were statistically significant (P<.001). The differences in the frequency of bone erosion or extrasinus extension did not reach statistical significance (P>.05).
Although many CT findings have been reported as suggestive of IFS, none are definitive. Silverman and Mancuso6 concluded that infiltration of the periantral fat planes in patients having both noncontrast and contrast-enhanced CT scans "may represent the earliest imaging evidence of invasive fungal disease." As early as 1993, Som and Curtin7 described soft tissue changes in a sinus with thickened, reactive bone inflammation, or with associated nasal cavity inflammation, as an early predictor of fungal disease. Nonspecific inflammatory soft tissue changes associated with bone destruction is also described by Gamba et al8 in patients with IFS.
Severe soft tissue thickening of the nasal cavity mucosa and soft tissues, including the turbinates, nasal walls, and septum, was present in 21 of the 23 patients in this study. The degree of nasal cavity inflammation was much higher than would be expected in routine rhinosinusitis in the absence of nasal polyposis (Figure 1, Figure 2, Figure 3, and Figure 4). Indeed, the degree of edema of the nasal mucosa is especially striking in patients who are neutropenic and, therefore, not expected to display a significant inflammatory response. This finding was rarely present in the control group of immunocompromised patients with sinusitis of nonfungal etiology. In that group, only 2 patients displayed nasal cavity soft tissue thickening and it was much less severe than in the IFS group (Figure 5 and Figure 6). In addition, 21 (91%) of the 23 patients had unilateral nasal mucosal thickening, which is not a typical finding in viral or bacterial rhinosinusitis. Only 2 patients (20%) in the control group displayed unilateral involvement. This severe unilateral nasal mucosal inflammatory process appears to be the most common radiologic characteristic of IFS in the early stages. Although this is a nonspecific finding, in our experience it is identified much more consistently than bone erosion or extrasinus extension, both of which are highly suggestive of IFS. The radiologic findings are consistent with the endoscopic and surgical findings of significant mucosal edema and inflammation surrounding areas of tissue ischemia. Frank tissue and bone necrosis was not commonly seen due to the early diagnosis of IFS in many of these patients.
Osseous erosion and extrasinus extension, obvious findings used for the radiologic diagnosis of IFS, are late-stage findings. The use of these 2 criteria alone would result in significant delay in diagnosis. In an immunocompromised patient with sinusitis symptoms, a high index of clinical suspicion, as well as radiographic detection of nasal cavity edema out of proportion to the patient's symptoms, can lead to early detection of IFS. We speculate that our high index of suspicion directly resulted in a cure rate of 83% (19/23) of the patients, a rate higher than that reported for IFS, with mortality reported in the range of 50% to 80%.9
To completely evaluate the sinuses and surrounding structures, CT scans should be performed at intervals no greater than 3 mm in both the axial and coronal planes. A "screening sinus technique," with coronal images only, is not sufficient. Both bone and soft tissue windows should be performed to better evaluate for bone erosion and extrasinus extension, respectively. Intravenous contrast is not usually necessary, although it may be helpful in delineating early orbital or dural enhancement.
Although no patient in this series had intracranial involvement, magnetic resonance imaging (MRI) should be obtained as well as neurosurgical evaluation in any patient suspected of having intracranial IFS. Howells and Ramadan10 propose the use of MRI as the initial radiologic procedure to diagnose IFS in patients at risk because of the nonspecific findings and the underestimation of the disease (compared with MRI and surgical findings) on CT scan. In their series of 4 patients, 2 had intracranial extension on MRI and intraoperatively, but it was missed on CT. Unlike our series, all of their patients had a more advanced disease, with extrasinus symptoms on presentation: 4 of 4 presented with facial swelling, 3 with orbital involvement, and 1 was obtunded. The mortality rate in their series was 100%, attributable to the advanced stage at presentation.10 In patients with neurologic and orbital symptoms at the time of presentation, the suspicion of IFS should be very high. These patients are better evaluated with MRI, although CT is also necessary for defining sinonasal anatomy. Since bone erosion is not a prerequisite for extrasinus extension, likely as a result of perivascular extension through penetrating vessels in the bone, MRI is more sensitive in disclosing intracranial and orbital invasion.8 In earlier stages of IFS, as demonstrated in our series, CT findings of significant unilateral nasal cavity soft tissue thickening appears to be a consistent finding.
At our institution, patients are closely monitored for IFS. Sinus CT scans are not routinely obtained prior to or after bone marrow transplantation. Sinus CT scans are ordered by the hematology or otolaryngology service to evaluate immunocompromised patients with sinonasal symptoms such as facial or nasal pain, headache, nasal obstruction, dental pain, facial swelling and/or paresthesias, or orbital complaints, or to evaluate unexplained fever. Patients with abnormalities on their sinus CT scan are evaluated by a rhinologist with rigid nasal endoscopy under topical anesthesia after decongestion. Biopsy specimens are obtained from areas of mucosal ischemia or necrosis and sent for frozen section analysis. If no mucosal abnormalities are seen but suspicion of IFS remains high, then directed biopsies are performed, particularly on edematous mucosa, either in the office or in the operating room. We recommend that intranasal biopsies be considered early in patients at risk even in the absence of specific nasal findings.10
Patients with confirmed IFS, and those with negative biopsy results but continued suspicion for IFS, are taken to the operating room for endoscopic examination, biopsy, and debridement. Prior to surgery, systemic antifungal therapy is initiated empirically in neutropenic patients with absolute neutrophil counts below 1000/µL. Granulocyte-monocyte colony-stimulating factor is administered to neutropenic patients when possible to restore normal immune function as early as possible, because return of the immune response is the single most important factor in recovering from IFS. This is reflected in the fact that all 4 patients who died in our series had absolute neutrophil counts of less than 500/µL. Patients are then closely followed up clinically, with frequent endoscopic examinations and periodic debridements as necessary, as well as serial noncontrast CT scans when necessary. As reflected in the 83% recovery rate for patients in our series, early detection and treatment is paramount to obtaining a good outcome.
The clinical challenge is identifying patients with invasive fungal disease as early as possible to permit initiation of medical and surgical intervention. However, the ability of the CT scan to detect the early stages of the disease is limited by high sensitivity and low specificity. Our review suggests that severe unilateral nasal cavity soft tissue inflammation is the most consistent early CT finding. Although a nonspecific finding, 21 of our 23 patients with IFS demonstrated nasal cavity edema much more severe than would be expected with routine rhinosinusitis, and 21 had unilateral disease, also unlike patients with viral or bacterial rhinosinusitis. Only 20% of immunocompromised, non-IFS sinusitis patients in the control group had nasal cavity soft tissue thickening and unilateral involvement, and the degree of thickening in these 2 patients was mild compared with the IFS group (Figure 1, Figure 2, Figure 3, Figure 4, and Figure 5). More extensive changes such as bone erosion, orbital involvement, retroantral fat pad thickening, and facial cellulitis, though highly suggestive of invasive fungal sinusitis, were seen much less frequently, as these are findings more common in more advanced IFS. These findings were absent in the control group. Since 11 of the 23 patients in our series had only coronal CT scans, we could have underestimated the frequency of cases with facial and retroantral soft tissue involvement. We recommend that CT scans be routinely obtained at 3-mm intervals in the axial and coronal planes, using both soft tissue and bone windows. If only the more classic CT findings are used to guide surgical intervention, then a high percentage of cases in the early stages may go undetected. The CT finding of excessive nasal mucosal edema can raise clinical suspicion but cannot definitively diagnose IFS. The diagnosis can only be reliably established with careful clinical evaluation and biopsy.
Invasive fungal sinusitis is one of the most challenging infections that otolaryngologists diagnose and treat. Favorable outcomes depend on early diagnosis and aggressive medical and surgical intervention. Unfortunately, there are no standard criteria for the radiographic diagnosis of IFS. Although bone erosion and extrasinus extension are the classic CT findings highly suggestive of IFS in immunocompromised patients, these are usually found late in the course of the disease. The most common early sign of IFS in our patient population is severe unilateral nasal cavity mucosal and soft tissue edema. Although this is a nonspecific finding, in a patient at risk the presence of significant mucosal edema in the nasal cavity should raise the suspicion of a possible invasive fungal infection. Physical examination with biopsy remains the gold standard for diagnosis. The clinician must maintain a high index of suspicion and be vigilant to establish the diagnosis of IFS.
Corresponding author: John M. DelGaudio, MD, 1365 Clifton Rd NE, Suite 2323A, Atlanta, GA 30322.
Accepted for publication June 27, 2002.