Sensitivity and specificity of neuroimaging signs in patients with idiopathic intracranial hypertension

Abstract

Background

The primary role of neuroimaging in idiopathic intracranial hypertension (IIH) is to exclude secondary causes of raised intracranial pressure. Recently, a few imaging markers have been described which may suggest diagnosis of IIH in atypical cases. We carried out this study to assess the prevalence and accuracy of these neuroimaging signs in predicting the diagnosis of IIH.

Methods

Eighty treatment-naive patients with IIH and 30 controls were recruited as per a predefined criterion. Magnetic resonance imaging (MRI) brain with detailed sella imaging was done in all patients.

Results

The most common abnormality noted was optic nerve tortuosity in 82.5% of patients, followed by posterior scleral flattening in 80%, perioptic subarachnoid space (SAS) dilatation in 73.8% and partial empty sella in 68.8% of patients. The presence of optic nerve tortuosity was the most sensitive sign on neuroimaging, though the highest specificity was seen for posterior scleral flattening and perioptic SAS dilatation. The presence of more than three neuroimaging features correlated with severity of vision loss.

Conclusion

In suggestive clinical scenarios, posterior scleral flattening, perioptic SAS dilatation and optic nerve tortuosity are highly sensitive and specific signs in IIH. This study also highlights the utility of MRI as a valuable tool for prognosis of visual outcome in patients with IIH.

Introduction

Idiopathic intracranial hypertension (IIH) was initially described by Quincke in 1893, who labeled it as meningitis serosa. Since then, the entity has gone numerous changes in nosology. In 2013, Freidman et al. ¹ proposed that this condition should best be described by using the umbrella term of ‘pseudotumor cerebri syndrome (PTCS)’, in which IIH forms the primary PTC category and identifiable causes like venous sinus thrombosis form the secondary PTC group. IIH has been characterized by presence of raised intracranial pressure (ICP), a normal cerebrospinal fluid (CSF) examination (other than raised opening pressure), a normal brain imaging, with either normal or slit-like ventricles on neuroimaging, and a normal neurological examination.

Historically, normal neuroimaging formed the basis of diagnosis, helping to exclude the secondary causes of raised ICP. Various imaging patterns suggestive of increased ICP have been recognized, which may aid to establish diagnosis in doubtful cases. While these imaging findings may suggest or support diagnosis of IIH, they cannot be solely considered diagnostic. The neuroimaging signs that may suggest a diagnosis of IIH include: empty sella turcica, slit-like ventricles, posterior scleral flattening, distension of optic nerve sheath, enhancement of the optic nerve head, tortuosity of the optic nerves, prominence of Meckel’s cave and transverse sinus stenosis. ² In patients with diagnostic dilemmas like the ones with IIH without papilledema, or with asymmetrical papilledema, diagnosis may strongly rely on neuroimaging findings. ³ The latest diagnostic criteria for pseudotumor cerebri incorporate neuroimaging findings as supportive criteria. Conversely, some of them such as partial empty sella have also been reported as incidental findings, likewise posterior scleral flattening has been described in patients with ocular hypotony. ⁴ These facts underscore the importance of assessing the sensitivity and specificity of the neuroimaging findings in IIH. Even though neuroimaging findings have been studied previously in IIH, there is still insufficient data on correlation of these findings with clinical features.

The purpose of our study was to assess the prevalence of these neuroimaging signs in IIH and evaluate their role as noninvasive biomarkers in the diagnosis of IIH.

Patient and methods

This prospective observational study was carried out in the Department of Neurology of a tertiary care institute in Northern India over a period of 1.5 years. This study was approved by the institutional ethics committee and written consent was obtained from all participants before inclusion in the study.

Patients

Eighty patients with IIH were enrolled over a period of 1.5 years in the study as per predefined inclusion and exclusion criteria. All patients with any secondary cause of increased ICP were excluded from the study (see supplementary material). Detailed history and examination findings were noted in all patients as per a predesigned proforma. Lumbar puncture was performed under aseptic conditions in all patients after informed consent with a 20-G lumbar puncture needle. CSF pressure was measured using a disposable open manometer after proper positioning of the patient. Diagnosis of IIH was based on Friedman Revised diagnostic criteria. ¹ Visual acuity was checked using Snellen’s chart and papilledema graded as per the Modified Frisen Scale (see supplementary material). ⁵ Detailed hemogram and biochemistry investigations were done.

Controls

Thirty controls were included in the study. All controls were recruited from the population of patients presenting to the Department of Neurology for ailments other than raised ICP. The diagnosis of all 30 controls included in this study is available in supplementary material (Annexure 3). Control subjects had no overt signs and symptoms suggestive of IIH.

Neuroimaging

All patients underwent gadolinium enhanced magnetic resonance imaging (MRI) of brain and orbits along with MR venography using standard sequences for brain, namely 3DT1 weighted magnetization prepared–rapid gradient echo (MPRAGE), axial T2 and T2 FLAIR weighted, susceptibility weighted and diffusion weighted imaging, gadolinium enhanced 3DT1 weighted (MPRAGE) images and 2D TOF MR venography along with axial and coronal T2 weighted images for the orbits. The MRI findings were assessed by a neuroradiologist who was blinded to the presenting signs and symptoms.

Special considerations were given to the following findings:

1. Empty sella

2. Optic nerve tortuosity

3. Posterior scleral flattening

4. Perioptic subarachnoid space dilatation

5. Pituitary tissue height on T1 weighted images

Treatment

All patients were treated with medical therapy in the form of acetazolamide (maximum dose 2–4 g), topiramate (maximum dose 200 mg) and furosemide (maximum dose 40 mg) as per standard treatment protocols used in our institute. In case of worsening of the visual symptoms on medical therapy or a very poor visual acuity at presentation, patients were offered surgical treatment/CSF diversion procedure. ⁶

Statistical analysis

The statistical analysis was carried out using Statistical Package for Social Sciences 23.0 for Windows. Mean and standard deviation was calculated for all continuous variables namely age, body mass index (BMI) and CSF pressure, except for duration of disease where median and interquartile range (IQR) was calculated (in weeks). Similarly, frequencies and proportions were calculated for all categorical variables. Chi-square test was used to compare the presence of various neuroimaging signs between the IIH and control group. A p-value of ≤0.05 was considered as statistically significant. For each proposed MRI sign in IIH, the sensitivity and specificity were calculated to measure the diagnostic strength of the sign. The 95% confidence interval (CI) was calculated to provide a range in which there is 95% confidence that the true values lay.

Results

Mean age of patients (SD) was 30.91 (8.84) years. The study group included 73 women (91.2%) and 7 men (8.8%); 37 patients (46.3%) were between 21 and 30 years of age.

The control group included 30 patients: 23 females (76.7%) and 7 males (23.3%). Mean age of the control group (SD) was 35.16 (6.57) years.

Clinical profile

The median (IQR) duration between onset of symptoms to first presentation to our hospital was 12.0 (4.5, 33.5) weeks, and the majority of patients (n=42, 52.5%) presented between 1 and 6 months after the onset of symptoms. The most common symptom at presentation was headache (n=80, 100%), which was migrainous in character in 56 (70%). The second most common complaint was transient visual obscurations (TVO), which were present in 63 (78.8%) patients, followed by pulsatile tinnitus in 45 (56.3%). Sixteen (20%) patients complained of vision loss at presentation. Diplopia was observed in 15 (18.8%) patients.

Mean BMI (SD) of patients in our cohort was 25.98 (3.74) kg/m² (range: 19.5 to 38.05). A total of 60 (75%) patients were overweight and obese as per Asia Pacific Task Force recommendation (2000). ⁷ Recent weight gain was complained of by 21 (26.3%) patients within 12 months of presentation, with average weight gain (SD) of 5.33 (1.74) kg. The baseline data of patients at presentation is detailed in Table 1.

Table 1.

Baseline data of IIH patients.

Characteristics	n	Percentage
Gender
Female	73	(91.2)
Male	7	(8.8)
Mean (SD) age in years	30.9	(8.84)
Age group (in years)
<20	7	(8.8)
21–30	37	(46.3)
31–40	22	(27.5)
41–50	13	(16.3)
>50	1	(1.3)
Median (IQR) duration of disease in weeks	12	(4.5, 33.5)
Mean (SD) BMI in kg/m²	25.98	(3.74)
Mean (SD) CSF pressure in cm of water	29.54	(4.40)
Clinical features
Headache	80	(100)
Papilledema (Modified Frisen grading)	76	(95.0)
TVOs	63	(78.8)
Visual field defects	48	(60.0)
Pulsatile tinnitus	45	(56.3)
Vision loss	23	(28.7)
• Mild vision loss (6/9–6/12)	10	(12.5)
• Moderate vision loss (6/18–6/60)	6	(7.5)
• Severe vision loss (<6/60)	7	(8.75)
Menstrual irregularities	22	(27.5)
Diplopia	15	(18.8)

BMI: body mass index; CSF: cerebrospinal fluid; TVOs: transient visual obscurations.

Clinical examination

While only 20% of patients had vision loss as presenting complaint, visual acuity was found to be abnormal in 28.7% of patients (n=23), and two patients (2.5%) presented with fulminant IIH. For convenience of analysis, we took the eye with the worst vision into consideration. As per classification by Wall et al., ⁸ ten patients (12.5%) had mild vision loss, six (7.5%) had moderate vision loss and seven (8.75%) had severe vision loss (Table 1). On automated perimetry, 48 patients (60%) had abnormal visual fields at presentation with most common finding being enlargement of the blind spot in 38 patients (80%), followed by localized nerve fiber bundle abnormalities in 18 patients (37.5%). In total, 76 (95%) patients were found to have papilledema. Four patients (5%) had no papilledema; however, they fulfilled the criteria for diagnosis of IIH without papilledema ⁹ (Table 1).

Cerebrospinal fluid (CSF) opening pressure was noted in all patients. Mean CSF pressure (SD) was 29.54 (4.40) cm of water. The majority of patients (n=56, 70%) had a CSF pressure of between 25 and 30 cm of water. All patients had negative workup for secondary causes including infections (Table 1).

Neuroimaging

Neuroimaging was abnormal at baseline in 74 (92.5%) patients. The most common abnormality noted on MRI was optic nerve tortuosity, seen in 66 (82.5%) patients, followed by posterior scleral flattening in 64 (80%) patients, perioptic subarachnoid space (SAS) dilatation in 59 (73.8%) and optic nerve head hyperintensity in 16 (20%). Transverse sinus stenosis on MR venography was found in 50 (62.5%) patients. Partial empty sella was found in 55 (68.8%) patients, and none of the patients had complete empty sella (Table 2). The mean pituitary height (SD) in patients with partial empty sella was 0.35 cm (0.08), and in patients with normal sella, it was 0.632 cm (0.05). Seventy-one (88.8%) patients had presence of more than one MRI sign, and 49 patients (61.3%) had more than three MRI signs of IIH.

Table 2.

Neuroimaging signs in idiopathic intracranial hypertension (IIH) (frequencies, sensitivity and specificity).

Sign	IIH (n/%)	Controls (n/%)	Sensitivity (95% CI)	Specificity (95% CI)
Optic nerve tortuosity*	66 (82.5%)	2 (6.7%)	82.5% (72.38–90.09%)	93.3% (77.93–99.18%)
Posterior scleral flattening*	64 (80%)	1 (3.3%)	80% (69.56–88.11%)	96.67% (82.78–99.92%)
Perioptic SAS dilatation*	59 (73.8%)	1 (3.3%)	73.75% (62.71–82.96%)	96.67% (82.78–99.92%)
Partial empty sella (PES)*	55 (68.8%)	7 (23.3%)	68.75% (57.41–78.65%)	76.67% (57.72–90.07%)
Transverse sinus stenosis*	50 (62.5%)	2 (6.7%)	62.5% (50.96–73.08%)	93.3% (77.93–99.18%)

*p-value was <0.001 (chi-square test) in all cases.

SAS: subarachnoid space.

Neuroimaging abnormalities were seen in 12 controls (40%). Partial empty sella was seen in seven controls (23.3%), while tortuous optic nerves and transverse sinus stenosis were seen individually in two controls (6.7%). The presence of posterior scleral flattening and perioptic SAS dilatation was seen in one control each (3.3%). One patient in the control group had both optic nerve tortuosity and partial empty sella (PES); hence, there were a total of 13 neuroimaging signs in the control group. None of the patients in the control group had three or more neuroimaging signs suggestive of IIH (Table 2).

On statistical analysis, all signs were found to have a significant association with IIH (p<0.001). The presence of optic nerve tortuosity was the most sensitive sign on neuroimaging (82.5%) followed by posterior scleral flattening (80%). The least sensitive sign was transverse sinus stenosis. The highest specificity (96.67%) was seen for posterior scleral flattening and perioptic SAS dilatation. Partial empty sella was found to have a sensitivity of 68.75% and specificity of 76.67% (Table 2).

Correlation of neuroimaging abnormalities with various clinical parameters

The presence of tortuous optic nerves was found to have a significant correlation with presence of diplopia, tinnitus, vision loss and papilledema (p<0.05). Posterior scleral flattening had a significant correlation (p<0.05) with tinnitus and vision loss. Perioptic SAS dilatation and transverse sinus stenosis correlated significantly with diplopia, while PES was found more in patients with vision loss (p<0.05). We tried to assess whether the number of signs on neuroimaging had any bearing on severity of the disease. The presence of more than three neuroimaging features of IIH had significant correlation with severity of vision loss (p<0.05). Since vision loss was observed in 23 patients, the visual outcome in these patients was assessed by comparing the visual acuity in the worst eye at presentation and at 6 months follow-up. A varied degree of visual improvement was seen in these patients. While all patients with vision loss and less than three neuroimaging signs showed improvement in visual acuity, four patients with more than three neuroimaging signs had worsening of vision on medical therapy and required CSF diversion procedures. The patients with less than three neuroimaging signs had a better visual outcome than patients with more than three imaging features of IIH (p<0.05).

Discussion

Although over a century has passed since the earliest description of IIH, it still remains underexplored, especially as far as the pathophysiology is concerned. Several neuroimaging findings have been put forward as signs supporting IIH; however, the estimates of their occurrence, sensitivity, specificity and relevance vary widely vary in the published literature. ¹⁰ Historically, the primary role of imaging in the diagnosis of IIH was to exclude secondary causes of raised ICP and papilledema. With time, various MRI signs, including empty sella turcica, reduced pituitary gland height, slit-like ventricles, flattening of posterior sclera, distension of optic nerve sheath, tortuosity of optic nerves and transverse sinus stenosis, emerged as neuroimaging markers of IIH. ¹¹

In our current study, neuroimaging was abnormal at baseline in 92.5% of patients, with most common finding being optic nerve tortuosity (82.5% patients), which along with frequency of other neuroimaging features was higher when compared to other studies.^12–14 High threshold of suspicion may lead on to early diagnosis. It is important to screen for the MRI signs of raised ICP with dedicated imaging modalities. We performed a dedicated imaging of brain and orbits along with MR venography to detect even subtle signs of raised ICP. This may be the reason for higher frequency of detection of neuroimaging signs in our study. Long-standing increased ICP due to delay in presentation may be associated with prominent imaging signs in a subset of patients.

Optic nerve tortuosity (Figure 1(a)) was found to be the most sensitive (82.5%) marker of intracranial hypertension in our study with a high specificity of 93.3%. The distal and the proximal points of fixation of the optic nerve allow it to kink easily in its path to the globe on protrusion of intracranial contents under pressure. The presence of optic nerve tortuosity in the absence of other signs of raised ICP has been reported previously in patients with neurofibromatosis and optic nerve glioma. Viglianesi et al. recently hypothesized that isolated optic nerve tortuosity can arise from anomalies in formation of optic stalk or migration of nerve fibers from optic cup to the forebrain. ¹⁵ Posterior scleral flattening (Figure 1(b)) was observed to be the second most sensitive sign (80%) with a specificity of 96.67%. This finding can be explained by transmission of elevated CSF pressure (ICP) through the subarachnoid space, followed by optic nerve sheath and hence to the posterior globe. Perioptic SAS dilatation (Figure 1(c)), though a less sensitive marker, was found to have a specificity of 96.67% for IIH in the current study. Perioptic nerve sheath enlargement is an outcome of widened CSF signal surrounding the optic nerves and is also attributed to the transmission of raised ICP. In settings where invasive ICP monitoring is not feasible, the estimation of optic nerve sheath diameter by B scan ultrasonography has been described as an indirect screening tool to detect raised ICP, and subsequently can be used to monitor response to therapy in patients with IIH. ¹⁶ , ¹⁷ Partial empty sella (Figure 2(b)) is a common neuroimaging finding in IIH and is attributed to raised ICP leading to herniation of subarachnoid space into the anterior portion of sella turcica, which fills with CSF. Though there are various studies on endocrinal abnormalities in empty sella, PES has also been reported in 8–35% of the general population. ¹⁸ In our study the prevalence of PES was 68.8% in IIH and 23.3% in the control population, with a sensitivity of 68.75% and specificity of 76.67%. Transverse sinus stenosis (Figure 2(c)) was found to be the least sensitive finding (62.5%) with a specificity of 93.3%. Narrowing of the transverse sinuses is best visualized on MR venographv; the compressible nature of this sinus makes it vulnerable to tapering in presence of raised ICP.

Figure 1.

Neuroimaging features in IIH. (a) Tortuous optic nerves; (b) posterior sclera flattening; (c) perioptic subarachnoid space dilatation.

Figure 2.

(a) Normal sella turcica; (b) partial empty sella; (c) transverse sinus stenosis.

According to the available literature, posterior scleral flattening is the most common neuroimaging sign in IIH, followed by optic nerve enhancement and vertical tortuosity of optic nerves. ¹⁹ In a retrospective case series, Brodsky and Vaphiades noted that a single examiner could predict the presence of intracranial hypertension in 90% of cases and the absence of raised ICP in all controls. ¹⁴ This was based on presence of the following MR signs: empty sella, enhancement and vertical tortuosity of the optic nerve. In another study, Agid et al. ¹³ reported that signs including empty sella turcica, posterior globe flattening, optic nerve sheath distension and optic nerve tortuosity are significantly associated with IIH, with posterior globe flattening being 100% specific. A recent study used contrast enhanced magnetic resonance venography to demonstrate the significant association of distal narrowing of the transverse sinuses with IIH, and this narrowing was found to have a sensitivity and specificity of 93% for identifying raised ICP. ²⁰ In a recent systematic review, a pooled specificity of MRI signs including empty sella, posterior globe flattening, perioptic nerve sheath distension, optic nerve tortuosity and slit-like ventricles was found to be high, though with a less pooled sensitivity. Transverse sinus stenosis was the only sign with a high pooled sensitivity and specificity. ²¹ Ibrahim et al. described venous attenuation sign on non-contrast enhanced computed tomography (CT) scan with a high sensitivity (96%), specificity (83.3%) and positive predictive value (85.7%) for detection of IIH. ²²

There are limited studies assessing the correlation of specific symptoms and signs of IIH with neuroimaging features with varied results. Hence, we attempted to assess any correlation between clinical features and neuroimaging findings in IIH. The presence of tortuous optic nerves was significantly associated with signs and symptoms of raised ICP including diplopia, tinnitus, vision loss and papilledema (p<0.05) suggesting that its presence may be an important diagnostic clue toward the diagnosis of IIH.

The number of MRI findings correlated with the visual outcome in these patients. Unlike a previous study by Saindaine et al., ²³ the severity of vision loss was found to have a significant association with presence of more than three neuroimaging signs of IIH in our study. Wong et al. in their retrospective chart review of 88 patients, reported a weak negative correlation between color vision and increased perineural fluid on MRI brain. They also reported a positive correlation between intraocular optic nerve protrusion and venous sinus stenosis with severity of disk edema. ²⁴ It is prudent to note that the assessment of disk edema by fundoscopic examination could lead to a bias, as in patients with IIH, the optic disk can appear edematous even during the early phase of raised ICP or when the ICP is normalized. Hence, recent studies have advocated use of A-scan sonography for detecting papilledema in these patients as it is more sensitive and can even differentiate papilledema from pseudopapilledema. ²⁵ In addition, the interpretation of visual outcome using disk edema may be faulty as it does not account for underlying secondary optic atrophy in a subset of patients.

The above findings reinforce the importance of early recognition of neuroimaging findings including optic nerve tortuosity, posterior sclera flattening and perioptic SAS dilatation, especially in asymptomatic patients or in patients with doubtful diagnosis. This may allow early diagnosis and timely management of IIH.

Conclusion

This study underscores the importance of neuroimaging signs in IIH, especially in asymptomatic patients or those with subtle ocular findings on examination. Though a variety of neuroimaging signs have been reported, in suggestive clinical scenarios posterior scleral flattening, perioptic SAS dilatation and optic nerve tortuosity are highly sensitive and specific signs in IIH.

Although PES is not a very specific or sensitive sign on neuroimaging, its presence must alert a clinician to look for other signs of increased ICP before considering the diagnosis of IIH.

The current study also highlights the utility of MRI as a valuable tool for prognosis of visual outcome in patients with IIH.