Abstract
Background and purpose
The cavernous sinus is a unique region owing to anatomical factors and the pathologies affecting it. The diagnosis of cavernous sinus syndrome (CSS) predominantly relies on clinicoradiological correlation. We studied the utility of computed tomographic (CT) scan versus magnetic resonance imaging (MRI) in the diagnosis of CSS.
Methods
A prospective observational study was conducted in a tertiary care center in north India. All patients presenting with a clinical syndrome of cavernous sinus involvement with radiologically confirmed lesions were enrolled in the study. MRI and CT scan with cavernous sinus cuts were done and reviewed by experienced neuroradiologists for cavernous sinus lesions and compared with the final diagnosis. Sensitivity and specificity were calculated.
Results
We included 48 patients in our study. A final diagnosis was achieved in 41 out of 48 (85.6%) patients. Fungal infections (16 (33.3%)) constituted the commonest cause of CSS, followed by neoplastic involvement (13 (27.1%)) and Tolosa–Hunt syndrome (12 (25%)). Vascular involvement was seen in three (6.3%) patients. Other rare causes were seen in four (8.3%) patients. CT scan had an overall sensitivity of 14.6% in achieving a final diagnosis, whereas MRI had an overall sensitivity of 70.7%, with a statistically significant difference (p < 0.001).
Conclusions
Although CT scan is a relatively cheap and accessible resource, its role in CSS diagnosis and management is limited because of poor yield. Hence, it is prudent to do an MRI as an initial investigation in cases of CSS.
Keywords: MRI, CT scan, cavernous sinus syndrome, Tolosa–Hunt syndrome, fungal sinusitis
Introduction
Ever since Jacobus Winslow1 coined the term ‘cavernous sinus’ in the year 1734, this region has attracted the attention of neurologists, neurosurgeons, neuro-ophthalmologists and neuro-otologists alike. The main reasons for this include: (a) its close proximity to the paranasal sinuses, orbit and several other important structures of brain; and (b) the important structures contained within its walls, namely ocular motor nerves, trigeminal nerve, ocular sympathetic nerves and, above all, internal carotid artery.2–4 These anatomic features makes the cavernous sinus (CS) region unique for several reasons. Firstly, CS is susceptible to several unique pathologies of brain such as infections (bacterial, fungal, etc.) which commonly spread from the paranasal sinuses, malignancies (e.g. nasopharyngeal carcinomas, pituitary tumors, metastases, etc.), vascular anomalies of the internal carotid artery or CS itself, and many others. Secondly, CS lesions are associated not only with considerable morbidity (due to effects on structures responsible for carrying out several vital functions essential for day to day life such as mastication and vision), but also with significant mortality (primarily due to the serious nature of underlying disease processes such as fungal infections and neoplasms). Thirdly, there is wide variation in the nature and seriousness of different disease processes affecting the CS, which range from relatively benign ones such as Tolosa–Hunt syndrome (THS) to extremely malignant ones such as nasopharyngeal carcinoma. Finally, the exact etiology of pathological processes affecting the CS is difficult to determine as this region is often not amenable to biopsy owing to the high risk of damage to important structures within it, mainly the internal carotid artery.1–6
Despite being associated with significant morbidity and mortality, there are still no universally accepted protocols for evaluation of cavernous sinus syndrome (CSS), a term which denotes any disease process affecting the CS. While there is general agreement that a gadolinium-enhanced MRI of the brain with CS cuts should be the imaging procedure of choice for evaluation of CSS, physicians continue to evaluate CSS with computed tomographic (CT) scans of brain.7,8 This is especially true in the developing world owing to financial constraints and lack of availability of MRI scans. As CT scans of brain are perceived to be less sensitive for detection of pathology in the CS region, their use for evaluation of suspected CSS may lead to diagnostic errors. However, to date, no study has compared the diagnostic utility of CT scan versus MRI scan of brain in suspected CSS. In the present study we compared the utility of CT scan versus MRI scan of brain in evaluation of suspected CSS.
Methods
The current prospective observational study was conducted on 48 patients with CSS attending a neurology outpatient department or who had been admitted to the neurology and emergency wards of the Postgraduate Institute of Medical Education and Research, Chandigarh, a tertiary care hospital and university teaching center in Northern India. Study duration was from January 2014 to July 2015. Written informed consent was obtained from all the patients before inclusion in the study, and the study was approved by the Institutional Ethics Committee. CSS was defined as involvement of two or more of the third, fourth, fifth (V1, V2) or sixth cranial nerves, or involvement of only one of them in combination with a neuroimaging-confirmed lesion in the CS. Once enrolled, all patients were subjected to a detailed neurological and general physical examination as per a predesigned pro forma and a meticulous history obtained. All the patients underwent routine investigations including detailed hemogram with erythrocyte sedimentation rate and C reactive protein, biochemistry profile, serum electrolytes and testing for human immunodeficiency virus as well as hepatitis B and C viruses. All patients underwent contrast-enhanced CT scan of brain with CS cuts on a multi-detector CT scan machine (GE- 64 slice). They also underwent a gadolinium-enhanced MRI of brain on a 1.5 Tesla Magnetom with thin-section cuts of the CS region, MR angiography and fat-suppressed MR images. All patients underwent a detailed etiological work up including biopsies, anti-nuclear and cytoplasmic anti-neutrophilic antibody testing, serum galactomannan, serum angiotensin-converting enzyme levels, contrast-enhanced CT scan of chest and abdomen, and cerebrospinal fluid examination, where indicated. In case at the end of diagnostic workup we are not able to find out a cause, the patient was labeled as CSS of undetermined etiology. Diagnosis of THS was made by ICHD-II criteria.9
Statistical analysis
Statistical analysis was done using SPSS version 22. Quantitative data were expressed as the mean or median. Qualitative/categorical data were expressed as a frequency/percentage. Chi-squared/Fisher’s exact test was applied to compare the discrete variables. A p-value of <0.05 was considered significant.
Results
Demographic profile and clinical features are shown in Table 1.
Table 1.
Clinical and demographic profile of patients with cavernous sinus syndrome.
| Variable | Value (n = 48) |
|---|---|
| Age in years | 43.25 ± 14.6 (range 11–70) |
| Men: women | 31:17 |
| Symptoms | |
| Acute presentation (duration of symptoms: <1 week) | 12 (25%) |
| Subacute presentation (duration of symptoms: 1–4 weeks) | 14 (29.2%) |
| Chronic presentation (duration of symptoms: >4 weeks) | 22 (45.8%) |
| Bilateral involvement | 10 (16.3%) |
| Headache (unilateral in 31 (67%); holocranial in 15 (33%)) | 46 (95.8%) |
| Diplopia | 44 (91.6%) |
| Ptosis (bilateral in 2 (6%)) | 33 (68.75%) |
| Proptosis (bilateral in 4 (26.6%)) | 15 (31.2%) |
| Facial numbness (unilateral in all) | 20 (41.7%) |
| Visual loss (bilateral in 1 (14.7%)) | 7 (14.5%) |
| Fever | 7 (14.6%) |
| Nasal blockage | 4 (8.3%) |
| Facial deviation (unilateral in all) | 5 (10.4%) |
| Hearing loss | 1 (2.1%) |
| Altered sensorium | 3 (6.3%) |
| Limb weakness | 5 (10.4%) |
| Seizures | 1 (2.3%) |
| Signs | |
| 6th cranial nerve | 40 (83.3%) |
| 3rd cranial nerve (pupils spared in 14 out of 40 (36.8%)) | 38 (79.2%) |
| 4th cranial nerve | 36 (75%) |
| Complete ophthalmoplegia | 29 (60.4%) |
| Trigeminal nerve | 22 (45.8%) |
| Ophthalmic division | 22/22 (100%) |
| Maxillary division | 15/22 (68.1%) |
| Mandibular division | 2/22 (9.9%) |
| Ophthalmic + maxillary divisions | 15/22 (68.1%) |
| All three divisions | 2/22 (9.9%) |
| 7th cranial nerve | 6 (12.5%) |
| Lower cranial nerves (9th to 12th ) | 3 (6.3%) |
| Optic nerve (bilateral in 2/9 (22.2%)) | 9 (18.8%) |
| Severe visual loss (visual acuity <3/60) | 4 (8.3%) |
| Horner’s syndrome (unilateral in all) | 3 (6.3%) |
Etiological profile of CSS (Table 2)
Table 2.
Common etiologies of cavernous sinus syndrome.
| Etiology | Number of patients (% age) (n = 48) |
|---|---|
| Fungal infections (aspergillosis in 7; mucormycosis in 3; probable fungal in 6) | 18 (33.3%) |
| Neoplastic involvement (metastases in 4; pituitary macroadenoma in 3; nasopharyngeal carcinoma in 2; meningiomas in 2; leukemia/lymphoma 1 each) | 13 (27.1%) |
| Tolosa–Hunt syndrome | 12 (25%) |
| Vascular causes | 3 (6.3%) |
| Others (hypertrophic pachymeningitis in 1; septic cavernous sinus thrombosis in 1; tuberculosis in 1; diabetic ophthalmoplegia in 1) | 4 (8.3%) |
In this study, we were able to achieve a definitive diagnosis in 41 out of 48 (85.6%) patients. In the remaining seven patients, six were diagnosed as probable fungal etiology. Two out of these six had pseudohyphae on nasal smears along with a clinical and radiological picture suggestive of fungal CSS. We were unable to do a biopsy in these two patients as one died and the other one left hospital against medical advice. In the remaining four patients a diagnosis of fungal CSS was made on the basis of radiological findings (T2 hypointense and T1 hyperintense lesion; microhemorrhages and severe paranasal sinus disease with extension of inflammation into cavernous sinus region). These four patients were given empirical antifungal therapy, two of whom improved, one died and one was lost to follow-up. In one patient diagnosis of diabetic ophthalmoplegia was made after detailed evaluation. He responded to steroid therapy and was asymptomatic at 6 months follow-up.
In our study, fungal infections (16 (33.3%)) constituted the commonest cause of CSS, closely followed by neoplastic involvement (13 (27.1%)) and THS (12 (25%)). Vascular etiologies (internal carotid artery aneurysms in one and carotido-cavernous fistulas in two) were seen in three (6.3%) patients. Other rare causes were seen in four (8.3%) patients. Amongst the neoplasms (n = 13), metastases accounted for four (30.8%), pituitary macroadenoma for three (23.1%), nasopharyngeal carcinoma for two (15.4%), meningiomas for two (15.4%) and lymphomas/leukemia accounted for one (7.7%) patient each. In all the patients with metastases, diagnosis was confirmed by biopsy from the primary site in four patients, namely adenocarcinoma of stomach, testicular seminoma, clear cell sarcoma and round cell tumor of femur. Diagnosis of nasopharyngeal carcinoma was made by biopsy from the involved site, while diagnosis of hematological malignancies was made by bone marrow examination. Pituitary adenomas were diagnosed on the basis of radiological findings, which was confirmed by histopathology after the excision. Diagnosis of meningiomas was made by MRI features and confirmed by biopsy. Amongst the 16 patients with fungal CSS, we could ascertain a definitive diagnosis (endoscopic sinus biopsy evidence of angioinvasion) in 10 (62.5%) patients (aspergillosis in 7; mucormycosis in 3). All 12 patients with THS fulfilled modified ICHD-2 criteria for diagnosis of THS. Of the four cases in the ‘others’ category, there was one patient each with idiopathic hypertrophic pachymeningitis (biopsy proven), septic CS thrombosis (clinicoradiological profile + blood and local tissue cultures), diabetic ophthalmoplegia (by exclusion) and tuberculosis (pulmonary tuberculosis + suggestive radiology).
MRI in CSS
In the present study, all the patients underwent gadolinium-enhanced MRI of brain, which was analyzed independently by a neuroradiologist as well as two neurologists with experience in the field of neuro-ophthalmology. Consensus was then obtained regarding MRI findings. MRI was abnormal in 45 (93.8%) patients with CSS. Regarding specific etiologies it was abnormal in 83.3% (10/12) of patients with THS, 100% of patients with fungal CSS, 100% of patients with neoplastic CSS and 100% of patients with vascular CSS. Out of three patients in which MRI brain was normal, two were diagnosed as THS and one had diabetes-related ophthalmoplegia. We further determined accuracy of MRI of brain in predicting the etiology of CSS. For this, we chose 41 patients in whom we were able to establish a definitive etiological diagnosis. The MRI-based etiological diagnosis was then compared with the final etiological diagnosis. We found that the etiological diagnosis considered on MRI of brain was the same as the final etiological diagnosis in 29 (70.7%) patients, while the final etiological diagnosis was considered as a differential diagnosis in another seven (17.1%) patients. The final etiological diagnosis was not considered on MRI of brain in five (12.2%) patients. Regarding specific etiologies, MRI was accurate in predicting exact etiology in 80% of fungal CSS, 66.7% of THS, 76.9% of neoplastic CSS and 100% of vascular CSS. These results are shown in Table 3.
Table 3.
CT and MRI brain findings in cavernous sinus syndrome (CSS).
| Frequency of MRI and CT abnormalities in CSS (n = 48) | ||||||
|---|---|---|---|---|---|---|
| Tolosa–Hunt syndrome (THS) (n = 12) | Fungal CSS (n = 16) | Neoplastic CSS (n = 13) | Vascular CSS (n = 3) | Other causes (n = 4) | Overall (n = 48) | |
| MRI brain abnormality present | 10 (83.3%) | 16 (100%) | 13 (100%) | 3 (100%) | 3 (75%) | 45 (93.8%) |
| CT brain abnormality present | 1 (8.3%) | 3 (18.7%) | 4 (30.7%) | 3 (100%) | 0 | 11 (22.9%) |
| Accuracy of MRI of brain in predicting final etiological diagnosis (n = 41) | ||||||
|
MRI diagnosis |
Tolosa–Hunt syndrome (n = 12) |
Fungal CSS (n = 10) |
Neoplastic CSS (n = 13) |
Vascular CSS (n = 3) |
Other causes of CSS (n = 3) |
Overall(n = 41) |
| Same as final etiological diagnosis | 8 (66.7%) | 8 (80%) | 10 (76.9%) | 3 (100%) | 0 | 29 (70.7%) |
| Final etiological diagnosis considered as differential | 2 (16.6%) | 2 (10%) | 0 | 3 (100%) | 7 (17.1%) | |
| Final etiological diagnosis not considered as differential | 2 (16.6%) | 0 | 3 (24.1%) | 0 | 0 | 5 (12.2%) |
| Accuracy of CT of brain in predicting final etiological diagnosis (n = 41) | ||||||
| Same as final etiological diagnosis | 0 | 2 (20%) | 1 (7.7%) | 3 (100%) | 0 | 6 (14.6%) |
| Final etiological diagnosis considered as differential | 1 (8.3%) | 1 (10%) | 3 (23.1%) | 0 | 0 | 5 (12.2%) |
| Final etiological diagnosis not considered as differential | 11 (91.7%) | 7 (70%) | 9 (69.2%) | 0 | 3 (100%) | 30 (73.2%) |
| Comparison of CT scan of brain versus MRI of brain | ||||||
MRI scan of brain was significantly more sensitive (p < 0.00001) and accurate (p = 0.0001) than CT scan of brain in detecting abnormality in cavernous sinus region in patients with cavernous sinus syndrome. Regarding specific etiologies, MRI of brain is significantly more sensitive (p = 0.0001; p < 0.01; p < 0.01) and significantly more accurate (p = 0.01; < 0.01 < 0.01) than CT of brain in THS, fungal CSS and neoplastic CSS. However, for vascular CSS, this difference was statistically insignificant.
CT of brain in CSS (Table 3)
All patients in the present study underwent contrast-enhanced (CE) CT scan of brain which was analyzed in same way as MRI scan of brain. Overall CECT of brain was reported abnormal in 11 (22.9%) patients with CSS. It was reported abnormal in all patients (n = 3) with vascular CSS, four (30.8%) patients with neoplastic CSS, three (18.7%) patients with fungal CSS and one (8.3%) patient with THS. Overall CECT scan of brain predicted accurate etiology in 14.6% of patients, while etiological diagnosis was considered as differential in 12.2% of patients. In 73.2% of patients final etiological diagnosis was not considered as differential on CECT scan of brain.
CECT scan of brain versus gadolinium-enhanced MRI scan (GdMRI) of brain in CSS
In the current study we analyzed the sensitivity as well as accuracy of GdMRI of brain versus CECT scan of brain in depicting the reason for CS abnormality. When analyzed, it was found that GdMRI scan of brain was not only much more sensitive for detecting abnormalities in the CS region but also much more accurate in determining the exact etiological diagnosis both for the CSS group as a whole as well as for specific subgroups, namely THS, fungal CSS (Figure 1(a)–(d)) and neoplastic CSS. The results are shown in Table 3.
Figure 1.
(a) T2-weighted images of the cavernous sinus showing T2 hypointense lesion in the right cavernous sinus with absent flow void in the right internal carotid artery, suggestive of thrombosis. (b) Contrast MRI of the right cavernous sinus showing intense homogeneous enhancement compared with contrast CT of the orbits, which appeared normal (c and d).
Discussion
In the western world, GdMRI scan of brain is usually the first investigation of choice in evaluation of suspected CSS, but the same in not true in developing countries where CT scan of brain (with or without contrast) continues to be used in evaluation of CSS due mainly to financial constraints and lack of availability of MRI. In the present study we evaluated the role of CECT of brain versus GdMRI of brain in evaluation of patients with CSS. To the best of our knowledge this is the first study which has addressed this important question.
The demographic profile of patients in the current study was a bit different from previous studies. As compared to the study by Fernadez et al.,4 whose cohort consisted of more females than males, our study group has predominantly males. Mean age in the current cohort was lower. Compared to a previous study,4 the frequency of headache, diplopia and facial numbness as well as 4th and 6th cranial nerve involvement was greater in our study group, while frequency of involvement of optic and 3rd cranial nerves was almost the same. As compared to the study by Keane et al.,3 frequency of involvement of 6th cranial nerve and optic nerve was lower, while frequency of involvement of 5th, 3rd and 4th cranial nerves was slightly higher in our cohort. The main reason for these differences may be related to differences in etiological profile of CSS in different series. While tumors were the commonest causes in other series,3,4 fungal infections accounted for most cases of CSS in the current cohort. This is expected in the developing world where infections are likely to dominate the etiologies of CSS.
GdMRI scan versus CECT scan of brain in evaluation of CSS
In the current study we determined relative sensitivity and accuracy of CT of brain versus MRI of brain in evaluation of CSS. MRI of brain was found to much more sensitive as well as accurate both for detection of abnormalities and for prediction of final etiological diagnosis in CSS. The findings of this study have several important implications. Firstly, we insist that GdMRI brain should be used as the initial imaging modality of choice in patients with suspected CSS. The urgent need for spreading this knowledge is even more important in the developing world where patients with CSS are not always seen by a neurologist or neurosurgeon. More often than not, these patients are seen by ophthalmologists and otologists as well as by general physicians, where use of a CT-based diagnostic protocol may lead to diagnostic inaccuracies and inadvertent delay in treatment. Secondly, in the current study, we found that gadolinium-enhanced MRI of brain predicted accurate etiological diagnosis of CSS in 70.7% of patients. Specifically for the fungal subgroup, MRI diagnosis was accurate in 80% of cases. Based on these results, we recommend starting empirical antifungal treatment based on suggestive MRI findings while awaiting the results of more definitive tests. This is especially important as a delay in starting antifungal treatment substantially increases the mortality in fungal CSS.10–12
A word of caution needs to be mentioned while interpreting the results of the present study. It needs to be acknowledged that CT of the brain has a unique role in evaluation of patients with CSS. It helps to better delineate the bony structures at the base of the skull as well the bony wall of the paranasal sinuses. Even in this study, we did a CT scan of the base of the skull to better delineate the bony anomalies where MRI STIR sequence showed bony abnormality. A brief algorithm to approach a case of suspected CSS is given in Figure 2.
Figure 2.
An algorithm for the approach for a patient suspected of cavernous sinus syndrome. CS: cavernous sinus; CSS: cavernous sinus syndrome; ACE: angiotensin-converting enzyme; IgG4: immunoglobulin subunit 4; GdMRI: gadolinium-enhanced magnetic resonance imaging; cANCA: cytoplasmic anti-neutrophil cytoplasmic antibodies; THS: Tolosa–Hunt syndrome; NA: not ascertained; FDG-PET: fluorodeoxyglucose positron emission tomography; NPC: nasopharyngeal carcinoma.
To conclude, our study proves better sensitivity and accuracy of GdMRI in evaluation of patients with suspected CSS when compared to CECT. Thus, it should be used as an initial imaging modality of choice in evaluation of suspected CSS. However, CT scan of brain may be used to delineate structures at the base of the skull where needed.
Footnotes
Conflict of interest: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data source: Data can be made available after reasonable request to the authors.
Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.
ORCID iDs
Karthik Vinay Mahesh https://orcid.org/0000-0001-7163-1190
Manoj K Goyal https://orcid.org/0000-0001-7375-4215
Ritu Shree https://orcid.org/0000-0002-2437-5507
References
- 1.Parkinson D. Lateral sellar compartment O.T. (cavernous sinus): history, anatomy, terminology. Anat Rec 1998; 251: 486–490. [DOI] [PubMed] [Google Scholar]
- 2.van Overbeeke JJ, Jansen JJ, Tulleken CA. The cavernous sinus syndrome. An anatomical and clinical study. Clin Neurol Neurosurg 1988; 90: 311–319. [DOI] [PubMed] [Google Scholar]
- 3.Keane JR. Cavernous sinus syndrome. Analysis of 151 cases. Arch Neurol 1996; 53: 967–971. [DOI] [PubMed] [Google Scholar]
- 4.Fernandez S, Godino O, Martinez-Yelamos S, et al. Cavernous sinus syndrome: a series of 126 patients. Medicine (Baltimore) 2007; 86: 278–281. [DOI] [PubMed] [Google Scholar]
- 5.Yoshihara M, Saito N, Kashima Y, et al. The Ishikawa classification of cavernous sinus lesions by clinico-anatomical findings. Jpn J Ophthalmol 2001; 45: 420–424. [DOI] [PubMed] [Google Scholar]
- 6.Smith JL, Taxdal DS. Painful ophthalmoplegia. The Tolosa-Hunt syndrome. Am J Ophthalmol 1966; 61: 1466–1472. [PubMed] [Google Scholar]
- 7.Castillo M. Imaging of the upper cranial nerves I, III-VIII, and the cavernous sinuses. Magn Reson Imaging Clin N Am 2002; 10: 415–431, v. [DOI] [PubMed] [Google Scholar]
- 8.Lee JH, Lee HK, Park JK, et al. Cavernous sinus syndrome: clinical features and differential diagnosis with MR imaging. AJR Am J Roentgenol 2003; 181: 583–590. [DOI] [PubMed] [Google Scholar]
- 9.Colnaghi S, Versino M, Marchioni E, et al. ICHD-II diagnostic criteria for Tolosa-Hunt syndrome in idiopathic inflammatory syndromes of the orbit and/or the cavernous sinus. Cephalalgia 2008; 28: 577–584. [DOI] [PubMed] [Google Scholar]
- 10.Reed C, Bryant R, Ibrahim AS, et al. Combination polyene-caspofungin treatment of rhino-orbital-cerebral mucormycosis. Clin Infect Dis 2008; 47: 364–371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gleissner B, Schilling A, Anagnostopolous I, et al. Improved outcome of zygomycosis in patients with hematological diseases? Leuk Lymphoma 2004; 45: 1351–1360. [DOI] [PubMed] [Google Scholar]
- 12.Chamilos G, Lewis RE, Kontoyiannis DP. Delaying amphotericin B-based frontline therapy significantly increases mortality among patients with hematologic malignancy who have zygomycosis. Clin Infect Dis 2008; 47: 503–509. [DOI] [PubMed] [Google Scholar]