Intracranial Hypotension: Improved MRI Detection With Diagnostic Intracranial Angles

Transcription

1 Neuroradiology/Head and Neck Imaging Original Research Shah et al. MRI Diagnosis of Intracranial Hypotension Neuroradiology/Head and Neck Imaging Original Research Lubdha M. Shah 1 Logan. McLean Marta E. Heilbrun Karen L. Salzman Shah LM, McLean L, Heilbrun ME, Salzman KL Keywords: intracranial hypotension, MRI DOI: /JR Received January 23, 2012; accepted after revision June 7, ll authors: Department of Radiology, University of Utah Health Sciences Center, 30N 1900 E, Rm 171, Salt Lake City, UT ddress correspondence to L. M. Shah (lubdha.shah@hsc.utah.edu). JR 2013; 200: X/13/ merican Roentgen Ray Society Intracranial Hypotension: Improved MRI Detection With Diagnostic Intracranial ngles OJECTIVE. Intracranial hypotension is an uncommon cause of headaches that is often misdiagnosed. The classic MRI features of intracranial hypotension can be variable and subjective. The purpose of this study was to provide objective criteria in the MRI evaluation of intracranial hypotension by quantifying normal values for the pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle. MTERILS ND METHODS. retrospective review of patients with the clinical diagnosis of intracranial hypotension and a control group was performed with measurements of the pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle. Qualitative evaluation of other MRI findings included dural enhancement, venous engorgement, subdural collections, brainstem slumping, and tonsillar herniation. RESULTS. In 29 patients with intracranial hypotension, the mean pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle were 41.2 (SD, ± 17.4 ), 4.4 mm (SD, ± 1.8), and (SD, ± 9.8 ), respectively. In the control group, the mean pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle were 65 (SD, ± 9.9 ), 7.0 mm (SD, ± 1.3), and (SD, ± 5.7 ), respectively. The differences in the pontomesencephalic angle and mamillopontine distance values for the intracranial hypotension group versus the control group were statistically significant (p < 0.01). The difference in the lateral ventricular angle measurements was not statistically significant (p = 0.37). Cutoff points of a 5.5-mm mamillopontine distance and 50 pontomesencephalic angle were estimated using receiver operating characteristic curves. CONCLUSION. In patients with the clinical suspicion of intracranial hypotension, we found that cutoff values of 5.5 mm or less for the mamillopontine distance and 50 or less for the pontomesencephalic angle were sensitive and specific in strengthening the qualitative MRI findings. Therefore, quantitative assessments may provide a more accurate diagnosis. I ntracranial hypotension may be spontaneous or related to prior injury, such as brain or spine trauma. The injury may be iatrogenic such as in cases of intracranial hypotension that develop as a sequela of lumbar puncture or surgery along the neuroaxis. lthough the site of CSF leak in spontaneous intracranial hypotension is occult in many cases [1], some cases might be due to a leak along a spinal nerve root sleeve or a dural defect in the skull base. There are no population-based epidemiologic studies regarding the prevalence of spontaneous intracranial hypotension to our knowledge. However, an emergency department based study estimated the annual incidence of spontaneous intracranial hypotension as 5 per 100,000 persons [2]. Patients with intracranial hypotension may present with the classic postural headache and a broad spectrum of symptoms in- cluding nausea, vomiting, neck pain, visual and hearing disturbances, and vertigo. These various clinical presentations may lead to misdiagnosis of intracranial hypotension, thus delaying effective treatment. In some cases, patients may undergo treatments for disorders mimicking intracranial hypotension that have their own associated risks [3]. MRI is vital in the diagnosis of intracranial hypotension. number of imaging findings have been described including dural enhancement, reversible pituitary enlargement [4, 5], subdural collections, brainstem slumping, and caudal tonsillar displacement [4, 6 8]. These MRI signs are variable and may not always be present in patients with convincing clinical signs. Diffuse pachymeningeal enhancement has been reportedly absent in some cases [9]. Messori et al. [10] found subdural collections in fewer than half of 400 JR:200, February 2013

2 MRI Diagnosis of Intracranial Hypotension their patients with intracranial hypotension. Qualitative MRI assessment of the brainstem may be performed in symptomatic patients to assess for slumping; however, there can be considerable subjectivity in this approach. The aim of this study was to provide objective criteria in the MRI evaluation of intracranial hypotension by quantifying normal values for the pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle. We hypothesized that these objective criteria will improve the accurate diagnosis of this commonly misdiagnosed syndrome. TLE 1: Demographic Characteristics of Patients With Intracranial Hypotension Subject No. ge (y) Sex Clinical History 1 54 F History of lumbar puncture; positional headache 2 76 F History of shunt placement 3 51 M History of shunt placement 4 24 M History of pseudomeningocele repair 5 76 F History of shunt placement 6 29 F Trauma 7 22 F Positional headache 8 23 F History of shunt placement 9 24 F Positional headache; CSF pressure of 5 cm H M Positional headache F Positional headache; tinnitus M History of shunt placement M Vertigo; increased falls F History of lumbar puncture; positional headache M Positional headache; neck stiffness F Intermittent headaches; vision loss F Positional headache F Pulsatile headache; suboccipital craniectomy without improvement F History of shunt placement F History of temporal surgery, suboccipital craniectomy, right temporal lobe surgery F Positional headache; nasal CSF leak F Positional headache F Positional headache; tinnitus M Positional headache; suboccipital craniectomy without improvement M Positional headache F Positional headache M Positional headache M Positional headache M Positional headache Materials and Methods fter institutional review board approval was obtained, a retrospective review of the PCS and electronic medical records was performed of patients who underwent imaging from December 2004 through July Patients were first identified in the imaging database by searching for the keywords intracranial hypotension. The criteria for inclusion in the intracranial hypotension group were clinical symptoms of intracranial hypotension and MRI of the brain with sagittal and coronal planes. Control subjects were selected from the same imaging time period and were matched for age and sex. The control subjects underwent contrast-enhanced MRI for clinical symptoms of headache or migraine, but their symptoms were unrelated to intracranial hypotension. Patients with tumor, hydrocephalus, ischemia, and hydrocephalus were excluded (Table 1). Two certificate of added qualification certified neuroradiologists with combined 20 years experience reviewed the MRI studies. The following parameters were assessed for both groups: pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle. The midline sagittal image was selected as defined by visualization of the optic chiasm and corpus callosum to measure the pontomesencephalic angle and mamillopontine distance. The pontomesencephalic angle is defined as the angle between a line tangential to the anterior margin of the midbrain and the line tangential to the superior margin of the pons (Fig. 1). The mamillopontine distance is defined as the distance between the inferior aspect of the mamillary bodies and the superior aspect of the pons. This measurement approximates the interpeduncular cistern (Fig. 1). The lateral ventricular angle is delineated as the angle between the medial margins of the right and left lateral ventricles (Fig. 1C). This angle was measured on coronal images at the level of the fornices, third ventricle, and pituitary infundibulum. inary evaluation of other MRI features included dural enhancement, venous engorgement, subdural collections, brainstem slumping, and tonsillar herniation. Venous engorgement is defined as prominent dural venous sinus enhancement with enlarged and rounded sinuses. prominent epidural venous plexus at the craniocervical junction and an enlarged pituitary are also supportive features of venous engorgement, and these features were evaluated in the assessment of venous engorgement. Cerebellar tonsillar ectopia of more than 5 mm was used as the criterion for tonsillar herniation [11]. rainstem slumping is defined as a low-lying third ventricle at or below the level of the floor of the sella turcica, horizontal configuration of the infundibulum, and red nuclei below the level of the tentorium [12]. Statistic nalysis For group comparisons, categoric variables were compared using a chi-square test or Fisher exact test, as appropriate. For continuous variables, the Student t test was used. ll data were analyzed using twotailed tests and a p value of < 0.05 was considered significant. Interobserver agreement between the readers was assessed using weighted Cohen kappa statistics. kappa value of 0 indicated poor agreement; , slight agreement; , fair agreement; , moderate agreement; , good agreement; and , excellent agreement. cut point for the discriminatory variables was estimated using receiver operating characteristic (ROC) curves. ll statistical analysis was performed using statistics software (Stata11, version 11.2, Stata- Corp) for Macintosh (pple). Results There were 29 patients with the clinical diagnosis of intracranial hypotension and 29 JR:200, February

3 Shah et al. control subjects. Of the 29 patients with clinical intracranial hypotension, 11 were men (37.9%) and 18 were women (62.1%); they ranged in age from 22 to 76 years old (mean age, 45.3 years). Imaging studies for the intracranial hypotension subgroup were obtained from December 2004 through July In the 29 patients with intracranial hypotension and variable MR findings of intracranial hypotension, the mean pontomesencephalic angle was 41.2 (SD, ± 17.4 ; CI, ), the mean mamillopontine distance was 4.4 mm Fig. 1 Healthy 41-year-old woman (control subject)., Pontomesencephalic angle is defined as angle between line drawn along anterior margin of midbrain and anterior superior margin of pons (lines). Mean value in patients with intracranial hypotension was 41.2 (SD, ± 17.4 ), Mamillopontine distance is defined as distance between inferior aspect of mamillary bodies to superior aspect of pons (line). Mean value in patients with intracranial hypotension was 4.4 mm (SD, ± 1.8). C, Lateral ventricular angle is defined as angle between medial margins of right and left lateral ventricles (lines). This angle was measured on coronal imaging at level of fornices (straight arrow), third ventricle (star), and pituitary infundibulum (curved arrow). Mean value in patients with intracranial hypotension was (SD, ± 9.8 ). Fig. 2 Diagnostic intracranial angles in patient with intracranial hypotension., 31-year-old man with 3-month history of positional headaches and neck stiffness. Pontomesencephalic angle is narrowed to 25 (dashed lines) as measured on PCS using standard angle-measuring tool. Note also low-lying cerebellar tonsils and brainstem slumping., 31-year-old man with 3-month history of positional headaches and neck stiffness. Mamillopontine distance () is narrowed to 3.2 mm as measured on PCS using standard ruler tool. Note also brainstem slumping and downward retraction of pituitary infundibulum. There is also mild cerebellar tonsillar displacement. C, 39-year-old woman with history of suboccipital craniectomy for Chiari decompression who presented with complaints of new onset headaches. Lateral ventricles are mildly effaced, and lateral ventricular angle (lines) is narrowed to 106. (SD, ± 1.8; CI, ), and the mean lateral ventricular angle was (SD, ± 9.8 ; CI, ) (Figs. 2 2C). In the control group of 29 subjects, there were 11 men (37.9%) and 18 women (62.1%); the subjects ranged in age from 17 to 86 years (mean age, 45.3 years). Imaging studies of the control group were obtained from December 2008 through December In the control group, the mean pontomesencephalic angle was 65 (SD, ± 9.9 ; CI, ), the mean mamillopontine distance was 7 mm (SD, ± 1.3; CI, ), and the mean lateral ventricular angle was (SD, ± 5.7 ; CI, ). Of the qualitatively evaluated variables (dural enhancement, venous engorgement, subdural collection, and tonsillar herniation), there was excellent interobserver agreement. Dural enhancement was observed in 17 of 29 (58.6%) intracranial hypotension patients with excellent interobserver agreement. Similarly, 14 of 29 intracranial hypotension patients (48.3%) showed MR features of venous engorgement with high C C 402 JR:200, February 2013

4 MRI Diagnosis of Intracranial Hypotension TLE 2: Qualitative MRI Findings in Patients With Intracranial Hypotension Subject No. Dural Enhancement Venous Engorgement Subdural Collection Tonsillar Herniation 1 N N Y N 2 Y Y N N 3 Y N Y N 4 Y N N N 5 Y Y Y N 6 Y N N N 7 N N N Y 8 N N N Y 9 N N N Y 10 N N Y Y 11 Y Y N Y 12 Y N N N 13 Y Y N N 14 N N N N 15 Y Y N Y 16 Y Y N N 17 N N N Y 18 N N N Y 19 Y Y Y N 20 N N N Y 21 Y Y N Y 22 Y Y N N 23 Y Y Y N 24 Y N Y Y 25 Y Y Y N 26 N Y N Y 27 N Y Y N 28 N N N Y 29 Y Y N Y Note Y = Yes, MRI finding was present; N = No, MRI finding was not present. concordance between readers. Thirty-one percent (9/29) of the intracranial hypotension cohort had subdural collections on MRI, which had excellent interreader agreement. Fourteen of 29 (44.3%) intracranial hypotension patients exhibited tonsillar herniation. There was very good interreader agreement for this observation. rainstem slumping showed only slight agreement between readers (κ coefficient = 0.39). One patient had no MRI qualitative findings, whereas 11 patients had one qualitative variable, nine had two, and eight had three. None of the patients had all four qualitative MRI findings (Table 2). The difference in the pontomesencephalic angle values for the intracranial hypotension group versus the control group was statistically significant (p < 0.01). Similarly, the difference in the mamillopontine distances between the two groups was also statistically significant (p < 0.01) (Figs. 3 and 3). The difference in lateral ventricular angle measurements for the patient group and control group was found to not be statistically significant (p = ). Good agreement was seen between observers for the pontomesencephalic angle and mamillopontine distance (κ coefficient = 0.62 and 0.71, respectively). Moderate agreement was noted between observers for the lateral ventricular angle (κ coefficient = 0.54). There was no statistically significant difference in the mamillopontine distance and lateral ventricular angle measurements between observers; however, the interobserver difference in the pontomesencephalic angle measurements was statistically significant. Discussion Intracranial hypotension can be a challenging diagnosis because of the varied spectrum of clinical symptoms and is sometimes misdiagnosed as migraine headaches, meningitis, or a psychogenic disorder [3], which can delay effective treatment. In some instances the cause of intracranial hypotension may be identifiable, such as a sequela of trauma from skull base or facial fracture, craniospinal surgery, or lumbar puncture. The latter has a 10 30% incidence of postural headache [5, 13]. In cases of spontaneous intracranial hypotension, the cause is often unknown, but spinal meningeal weakness is suspected [7]. n underlying generalized connective disorder has been reported in up to two thirds of patients with spontaneous intracranial hypotension [14], whereas a mechanical role is suggested in cases in which there is a history of only minor trauma preceding symptoms. history of trauma is elicited in approximately one third of the cases [15, 16]. Osseous abnormalities have also been implicated in the pathogenesis of spontaneous intracranial hypotension due to either degenerative osteophytes [17] or congenital spurs [18] piercing the dura. In clinically suspected intracranial hypotension cases without a definite inciting event, MRI of the brain is an important component of the current diagnostic criteria. The purpose of this study was to provide objective criteria in the MRI evaluation of intracranial hypotension by quantifying normal and abnormal values for the pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle in an effort to more accurately diagnose intracranial hypotension. The main diagnostic criteria for intracranial hypotension are listed in the International Classification of Headache Disorders (ICD-2) [19], which was revised in These diagnostic criteria include the presence of orthostatic headache with at least one secondary clinical finding including neck stiffness, hyperacusia, photophobia, and nausea. For the diagnosis of intracranial hypotension, the criteria also require either a CSF pressure of less than 6 cm H 2 O or MRI features of intracranial hypotension [19]. However, because of these varied and sometimes perplexing clinical features of intracranial hypotension, other investigators have proposed additional diagnostic criteria based on brain and spine im- JR:200, February

5 Shah et al. Pontomesencephalic ngle ( ) Control Subjects aging findings, clinical manifestations, lumbar puncture results, and response to epidural blood patch [20]. Schievink et al. [20] confirmed the diagnosis of intracranial hypotension in 94 of 107 patients using these criteria. In our study, 10 of 29 patients who were treated with one or more epidural blood patches experienced relief of symptoms. It has been reported that most patients with intracranial hypotension require two or more epidural blood patches [20]. Resolution of symptoms 72 hours after an epidural blood patch is not typical, but it is an obligatory criterion in the ICD-2 criteria [19]. Shunt adjustment abated symptoms in six of 28 cases in which overshunting was the cause for intracranial hypotension symptoms. Patients With Intracranial Hypotension Mamillopontine Distance (mm) Control Subjects The clinical hallmark of intracranial hypotension is orthostatic headache [3, 5]. Sixteen patients in our series presented with positional headache. It is important to note that, with time, the posture-related component of the headache often abates or may even disappear if a CSF leak is left untreated [3]. One patient in our study had intermittent headaches and another complained of pulsatile headaches. Myriad other symptoms may be associated with intracranial hypotension: nausea, vomiting, photophobia, and posterior neck stiffness. Some of these symptoms may become more prominent than postural headache and may erroneously lead one to consider subarachnoid hemorrhage or infectious meningitis. These symptoms may be caused by downward displacement of the brainstem. Stretching of the oculomotor nerves is thought to explain diplopia experienced by some patients, and stretching of the optic apparatus over the sella turcica may be the cause of visual field effects reported by others [21]. One patient with intracranial hypotension in this study experienced slow deterioration in her vision thought to be caused by stretching of the optic chiasm over an enlarged pituitary gland. Shunt revision stopped further vision decline, and subsequent MRI findings revealed decreased pituitary enlargement and resolution of dural enhancement. Patients with intracranial hypotension may experience hearing loss (cochlear nerve), vertigo (vestibular nerve), facial pain (trigeminal nerve), facial spasm (facial nerve), or dysgeusia (chorda tympani) due to similar mechanisms [6, 7]. CSF pressure changes may be Patients With Intracranial Hypotension Fig. 3 ox plots., ox plot shows statistically significant difference in average pontomesencephalic angles between patients with intracranial hypotension (mean ± SD, 41.2 ± 17.4 ) and control subjects (65 ± 9.9 )., ox plot shows statistically significant difference in average mamillopontine distance between patients with intracranial hypotension (mean ± SD, 4.4 ± 1.8 mm) and control subjects (7.0 ± 1.3 mm). Fig year-old woman with positional headaches and tinnitus (patient 11 in Tables 1 and 2). rrows indicate thin, diffuse pachymeningeal (dural) enhancement on contrast-enhanced coronal T1- weighted MR image. an alternative explanation for the disturbances in hearing and vertigo [22]. Tinnitus was observed in two patients and vertigo with falls was reported in one patient in our series. Upper extremity radiculopathy due to stretching of cervical nerve roots has also been reported in some patients [3]. Severe cases of intracranial hypotension may result in obtundation, even coma, due to downward brainstem displacement [23]. Rare associations of dementia [24] and parkinsonism have also been hypothesized [25]. Imaging is a critical part of identifying the diagnostic criteria for intracranial hypotension. There are five typical but variably present imaging characteristics: dural (pachymeningeal) enhancement, venous engorgement, pituitary hyperemia, subdural collections, and brainstem slumping [2]. The classic MRI feature of diffuse dural enhancement can be seen in 56 80% of patients with intracranial hypotension [6, 20]. Dural enhancement can be supratentorial and infratentorial, is typically thin and diffuse, and does not involve the leptomeninges [8] (Fig. 4). Seventeen of the 29 intracranial hypotension patients in our study showed this type of diffuse dural enhancement. This diagnostic criterion for intracranial hypotension is a reliable MR finding because interobserver agreement was excellent in our study. potential confounder for dural enhancement may be postoperative pachymeningeal irritation. Eleven of 29 patients had either a prior lumbar puncture or prior intracranial surgery, which the readers concordantly observed. Seven of these postoperative cases 404 JR:200, February 2013

6 MRI Diagnosis of Intracranial Hypotension with a clinical diagnosis of intracranial hypotension had dural enhancement. lthough infectious meningitis may cause diffuse enhancement, infectious meningitis typically shows leptomeningeal involvement with curvilinear enhancement interdigitating along the sulci rather than pachymeningeal enhancement. None of the patients in our patient group had clinical signs or symptoms of infection. nother characteristic imaging finding of intracranial hypotension is the enlargement and rounded appearance of the dural venous sinuses on MRI [5] (Fig. 5). On sagittal T1- weighted MRI, the convex bulging of the inferior border of the dominant transverse sinus seen in intracranial hypotension patients has been described as the venous distention sign [26] (Fig. 5). Cerebral angiography may reveal engorgement of the dural venous sinuses or large cortical veins [27]. In this study, 48.3% of the intracranial hypotension cases revealed MR features of venous engorgement. Enlargement of the intracranial venous structures is most easily seen when pre- and posttreatment scans are compared. Hyperemia of the dural and epidural venous sinuses may be associated with reactive hyperemia of the pituitary gland, which may be mistaken for a tumor or hyperplasia [28, 29]. Increased height of the pituitary gland (mean ± SD, 6.9 ± 2.3 mm) has been reported to have a sensitivity of 63% and specificity of 97% for the diagnosis of intracranial hypotension[30]. This pituitary enlargement is reversible [4, 5], which can be challenging to recognize without presymptomatic scans and posttreatment scans for direct comparison. Fig. 5 Venous sinus findings in patient with intracranial hypotension., 31-year-old man with positional headaches (patient 15 in Tables 1 and 2). Coronal T1-weighted MR image shows venous engorgement with round contour of superior sagittal sinus (arrow) [26]. Note also diffuse pachymeningeal (dural) enhancement., 53-year-old man with vertigo and increasing falls (patient 13 in Tables 1 and 2). Sagittal T1-weighted MR image shows venous distention sign with convexity of inferior contour of dominant transverse sinus (arrow) [26]. In addition to changes in vascular volumes in cases of intracranial hypotension, there may be alterations in the extraaxial compartment. In 36 50% of patients with intracranial hypotension, subdural collections may be caused by transudation of fluid [20, 31] (Fig. 6). Subdural collections were noted in 31.0% of the intracranial hypotension cases in this study. Most of these subdural collections overlie bilateral cerebral convexities and are thin without significant mass effect. They may also be present in the posterior fossa overlying the cerebellar hemispheres and in the retroclival region [2]. Occasionally, subdural hematomas may also occur and have varying degrees of mass effect [31]. rainstem slumping is an MRI sign that is specific for intracranial hypotension [4, 6 8, 15] and is seen in approximately 51% of cases [20]. Several imaging features of brainstem slumping have been identified and described in the literature including ventricular effacement [8], effacement of the suprasellar and prepontine cisterns, bowing of the optic chiasm over the sella turcica, flattening of the ventral pons [12], and caudal displacement of the cerebellar tonsils [2, 3, 5] (Figs. 7 and 7). In addition to the position of the cerebellar tonsils, the position of the fourth ventricle, the position of the infundibular recess, and the angle between the bicommissural line and a line tangential to the floor of the fourth ventricle have been evaluated in cases of intracranial hypotension [10]. manifestation of brainstem slumping is narrowing of the angle between the vein of Galen and the straight sinus that, in turn, may impair venous drainage of the thalami and basal ganglia [12]. Despite, or perhaps because of, the various descriptive features of brainstem slumping, brainstem slumping can be difficult to identify on imaging: Interobserver agreement for brainstem slumping was only fair between the readers in this study. rainstem slumping is an important MRI characteristic of intracranial hypotension but MR identification of brainstem slumping can be subjective. Therefore, objective measures such as the pontomesencephalic angle and mamillopontine distance can be valuable in making the diagnosis of intracranial hypotension. We evaluated the pontomesencephalic angle, mamillopontine distance, and lateral ventricular angle in subjects with intracranial hypotension and in control subjects. We quantified the mamillopontine distance: a mean of 4.4 mm in patients with intracranial hypotension as compared with 7.0 mm in control subjects. This decreased parameter has been one of the qualitative descriptors of brainstem slumping [5]. dditionally, the pontomesencephalic angle was decreased in patients with intracranial hypotension, with a mean of 41.2 as compared with a mean of 65 seen in control subjects. Differences between patients and control subjects in the mamillopontine distance and pontomesencephalic angle measurements were statistically significant. Furthermore, interobserver agreement for the mamillopontine distance was good. There was a statistically significant difference between Fig year-old man with positional headaches (patient 10 in Tables 1 and 2). xial FLIR MR image displays hyperintense subdural collections (arrows) overlying bilateral cerebral convexities with mild mass effect on subjacent parenchyma. JR:200, February

7 Shah et al. observers for the pontomesencephalic angle measurements, which may be because of difficulty in measuring the angle when there is anatomic distortion. However, the overall difference among patients with intracranial hypotension and control subjects for the pontomesencephalic angle was statistically significant. Moreover, we found that a 5.5-mm mamillopontine distance and 50 pontomesencephalic angle are optimally sensitive and specific cutoff values. ecause the lateral ventricles can be effaced in intracranial hypotension, we would expect that the lateral ventricular angle would be affected. The lateral ventricular angle, also referred to as the corpus callosal angle, has been used to differentiate between pressure hydrocephalus, normal pressure hydrocephalus, and hydrocephalus ex vacuo [32]. Ishii et al. [33] used this measurement to differentiate patients with normal pressure hydrocephalus from patients with lzheimer disease and from healthy subjects. However, although the lateral ventricular angle may be altered by changes in CSF pressure or volume, our study found no statistically significant difference in the lateral ventricular angle measurements between control subjects and patients with intracranial hypotension. Of the three objective criteria evaluated in this study, the mamillopontine distance and pontomesencephalic angle are useful when the qualitative MRI findings are variable. None of the 29 patients displayed all four qualitative MRI features. The patient with no qualitative Fig. 7 Qualitative MR findings in patient with intracranial hypotension., 47-year-old woman with positional headaches and tinnitus (patient 11 in Tables 1 and 2). Sagittal T1-weighted MR image illustrates findings of brainstem slumping with effacement of prepontine cistern (arrowhead) and tonsillar herniation (arrow). Note also small mamillopontine distance and small pontomesencephalic angle., 36-year-old man with positional headaches (patient 28 in Tables 1 and 2) requiring multiple epidural blood patches. Sagittal T1-weighted MR image shows numerous imaging findings of intracranial hypotension including brainstem slumping, effacement of suprasellar cistern (arrowhead), flattening of central pons (star), and draping of optic apparatus over sella turcica (arrow). Note also cerebellar tonsillar herniation. findings had a decreased mamillopontine distance of 5.15 mm and a marginally reduced pontomesencephalic angle of 56.5, both less than control values and close to the optimal cutoff values. Given the high clinical suspicion in that case the patient had developed postural headaches after a lumbar puncture the quantitative findings supported the clinical diagnosis of intracranial hypotension. Further investigation revealed the site of the CSF leak. Of the 28 patients with one to three qualitative MRI findings, the mamillopontine distance, pontomesencephalic angle, or both were less than the cutoff values in most patients (n = 24). The remaining four patients had compelling clinical symptoms of intracranial hypotension and were treated. supportive feature of brainstem slumping is cerebellar tonsillar herniation, which was seen in 14 of 29 intracranial hypotension patients. lthough three patients had occipital decompression, the tonsillar morphology was peglike and elongated and indicative of ectopia. However, cerebellar tonsillar ectopia can be mistaken for a Chiari I malformation [2] and patients may undergo unnecessary decompressive posterior fossa surgery [7]. Two patients in this study underwent suboccipital craniectomy for tonsillar ectopia and headache that did not improve their symptoms. congenital Chiari I malformation is defined radiologically as 5-mm displacement of cerebellar tonsils below the foramen magnum [11]. However, arkovich et al. [34] showed that tonsils may lie below the foramen magnum in 14% of healthy control subjects and that cerebellar tonsils may lie 5 mm or more below the foramen magnum in 0.005% of healthy subjects. Chiari I malformation is a disorder of the paraxial mesoderm resulting in underdevelopment of the posterior cranial fossa, crowding of the hindbrain, and CSF flow abnormalities. In approximately 52% of patients with a Chiari I malformation and no associated etiologic cofactor, Milhorat et al. [35] found a small posterior cranial fossa with constriction inferior to the line between the tuberculum sella and internal occipital protuberance (Twining line) and narrowing of the inferior and superior outlet areas. They concluded this classic Chiari malformation to be secondary to the premature stenosis of the basiexoccipital and exosupraoccipital synchondroses [35]. The imaging findings of intracranial hypotension can be explained by the pathophysiology, which is hypothesized by the Monro- Kellie doctrine [36]. ccording to this theory, the volume of the intracranial and intraspinal spaces is a fixed constant, and these spaces are composed of the CSF, brain and spinal cord parenchyma, and arteries and veins. If there is a change in one of the compartments, particularly the CSF and blood volume compartments, there will be a compensatory change in the others. There is an increase in the volumes of the meninges and vessels because the blood-brain barrier restricts extracellular fluid expansion, thereby generating interstitial edema [26]. decrease in CSF volume results in an increase in vascular volume in particular, in the volume of venous blood because of less resistance relative to the arterial system [12]. Subdural collections may develop if vascular volume enlargement does not match CSF volume loss. With age, there is parenchymal volume loss, which may allow increased volume in the intracranial vascular compartment. We hypothesize this scenario to explain the positive correlation between age and dural enhancement and venous sinus enlargement. Consequently, observations of vascular engorgement should be correlated with clinical context in the setting of senescent changes. The negative correlation between tonsillar herniation and age may also be related to parenchymal volume. Younger patients are expected to have lower-lying cerebellar tonsils than older patients because of their relatively full parenchymal volume. Tonsillar morphology coupled with additional features of brainstem slumping including quantitative measurements such the pontomesencephalic angle 406 JR:200, February 2013

8 MRI Diagnosis of Intracranial Hypotension and mamillopontine distance will help confirm the diagnosis of intracranial hypotension. One limitation of this study is its retrospective nature. The readers were not blinded to which cases were patients with intracranial hypotension and which were control subjects. The intracranial hypotension group was clinically diagnosed, which can be difficult as described. In some clinically suspicious cases of intracranial hypotension, the brain MRI findings were used to support the clinical diagnosis. The site of leak in the spontaneous intracranial hypotension cases was not always determined. There is risk of a high false-positive rate using the mamillopontine distance and pontomesencephalic angle cutoff values given the small sample size used for ROC analysis. However, if used in patients with a clinical suspicion of intracranial hypotension and if supported by qualitative MRI findings, the mamillopontine distance and pontomesencephalic angle are valuable for more definitive diagnosis of intracranial hypotension. Conclusion In patients with clinical findings suggestive of intracranial hypotension, qualitative MRI findings can sometimes be equivocal or can even be absent. We found cutoff values of 5.5 mm or less for the mamillopontine distance and 50 or less for the pontomesencephalic angle to be both sensitive and specific in strengthening qualitative MRI findings suggestive of intracranial hypotension. Therefore, quantitative assessments may provide a more accurate diagnosis of this commonly misdiagnosed syndrome. References 1. Mokri. Headaches caused by decreased intracranial pressure: diagnosis and management. Curr Opin Neurol 2003; 16: Schievink WI. Spontaneous spinal cerebrospinal fluid leaks and intracranial hypotension. JM 2006; 295: Schievink WI. Misdiagnosis of spontaneous intracranial hypotension. rch Neurol 2003; 60: Dillon WP, Fishman R. Some lessons learned about the diagnosis and treatment of spontaneous intracranial hypotension. JNR 1998; 19: Yuh EL, Dillon WP. Intracranial hypotension and intracranial hypertension. Neuroimaging Clin N m 2010; 20: Mokri, Piepgras DG, Miller GM. Syndrome of orthostatic headaches and diffuse pachymeningeal gadolinium enhancement. Mayo Clin Proc 1997; 72: Schievink WI, Meyer F, tkinson JL, Mokri. Spontaneous spinal cerebrospinal fluid leaks and intracranial hypotension. J Neurosurg 1996; 84: Fishman R, Dillon WP. Dural enhancement and cerebral displacement secondary to intracranial hypotension. Neurology 1993; 43: Mokri, tkinson JL, Dodick DW, Miller GM, Piepgras DG. bsent pachymeningeal gadolinium enhancement on cranial MRI despite symptomatic CSF leak. Neurology 1999; 53: Messori, Simonetti F, Regnicolo L, Di ella P, Logullo F, Salvolini U. Spontaneous intracranial hypotension: the value of brain measurements in diagnosis by MRI. Neuroradiology 2001; 43: Elster D, Chen MY. Chiari I malformations: clinical and radiologic reappraisal. Radiology 1992; 183: Savoiardo M, Minati L, Farina L, et al. Spontaneous intracranial hypotension with deep brain swelling. rain 2007; 130: Raskin NH. Lumbar puncture headache: a review. Headache 1990; 30: Schievink WI, Gordon OK, Tourje J. Connective tissue disorders with spontaneous spinal cerebrospinal fluid leaks and intracranial hypotension: a prospective study. Neurosurgery 2004; 54:65 70; discussion, Schievink WI. Spontaneous spinal cerebrospinal fluid leaks: a review. Neurosurg Focus 2000; 9:e8 16. Schievink WI, Ebersold MJ, tkinson JL. Rollercoaster headache due to spinal cerebrospinal fluid leak. Lancet 1996; 347: inder DK, Sarkissian V, Dillon WP, Weinstein PR. Spontaneous intracranial hypotension associated with transdural thoracic osteophyte reversed by primary: dural repair case report. J Neurosurg Spine 2005; 2: Vishteh G, Schievink WI, askin JJ, Sonntag VK. Cervical bone spur presenting with spontaneous intracranial hypotension: case report. J Neurosurg 1998; 89: Headache Classification Subcommittee of the International Headache Society. The international classification of headache disorders, 2nd ed. Cephalalgia 2004; 24(suppl 1): Schievink WI, Maya MM, Louy C, Moser FG, Tourje J. Diagnostic criteria for spontaneous spinal CSF leaks and intracranial hypotension. JNR 2008; 29: Horton JC, Fishman R. Neurovisual findings in the syndrome of spontaneous intracranial hypotension from dural cerebrospinal fluid leak. Ophthalmology 1994; 101: Portier F, de Minteguiaga C, Racy E, Huy PT, Herman P. Spontaneous intracranial hypotension: a rare cause of labyrinthine hydrops. nn Otol Rhinol Laryngol 2002; 111: inder DK, Dillon WP, Fishman R, Schmidt MH. Intrathecal saline infusion in the treatment of obtundation associated with spontaneous intracranial hypotension: technical case report. Neurosurgery 2002; 51: ; discussion, Hong M, Shah GV, dams KM, Turner RS, Foster NL. Spontaneous intracranial hypotension causing reversible frontotemporal dementia. Neurology 2002; 58: Pakiam S, Lee C, Lang E. Intracranial hypotension with parkinsonism, ataxia, and bulbar weakness. rch Neurol 1999; 56: Farb RI, Forghani R, Lee SK, Mikulis DJ, gid R. The venous distension sign: a diagnostic sign of intracranial hypotension at MR imaging of the brain. JNR 2007; 28: Koss S, Ulmer JL, Hacein-ey L. ngiographic features of spontaneous intracranial hypotension. JNR 2003; 24: lvarez-linera J, Escribano J, enito-leon J, Porta-Etessam J, Rovira. Pituitary enlargement in patients with intracranial hypotension syndrome. Neurology 2000; 55: Mokri, tkinson JL. False pituitary tumor in CSF leaks. Neurology 2000; 55: Rohr, Jensen U, Riedel C, et al. MR imaging of the optic nerve sheath in patients with craniospinal hypotension. JNR 2010; 31: Schievink WI, Maya MM, Moser FG, Tourje J. Spectrum of subdural fluid collections in spontaneous intracranial hypotension. J Neurosurg 2005; 103: Sjaastad O, Nordvik. The corpus callosal angle in the diagnosis of cerebral ventricular enlargement. cta Neurol Scand 1973; 49: Ishii K, Kanda T, Harada, et al. Clinical impact of the callosal angle in the diagnosis of idiopathic normal pressure hydrocephalus. Eur Radiol 2008; 18: arkovich J, Wippold FJ, Sherman JL, Citrin CM. Significance of cerebellar tonsillar position on MR. JNR 1986; 7: Milhorat TH, Nishikawa M, Kula RW, Dlugacz YD. Mechanisms of cerebellar tonsil herniation in patients with Chiari malformations as guide to clinical management. cta Neurochir (Wien) 2010; 152: Mokri. The Monro-Kellie hypothesis: applications in CSF volume depletion. Neurology 2001; 56: JR:200, February