Brain Herniation Into Arachnoid Granulations: Clinical and Neuroimaging Features

Brain Herniation into Arachnoid Granulations: Clinical and Neuroimaging Features

Greta B. Liebo, John (Jack) I. Lane, Jamie J. Van Gompel, Laurence J. Eckel, Kara M. Schwartz, Vance T. Lehman From the Department of Radiology (GBL, J(J)IL, LJE, KMS, VTL); and Department of Neurologic Surgery and Otolaryngology (JJVG), Mayo Clinic-Rochester, Rochester, MN.

ABSTRACT

BACKGROUND/PURPOSE: Brain herniation into presumed arachnoid granulations (BHAG) is an increasingly recognized entity. Though it has previously been described as isointense to brain matter on magnetic resonance imaging (MRI), we have encountered 21 cases in our clinical practice, many of which have signal abnormality in the herniated or adjacent parenchyma. We performed a systematic analysis on these cases to further understand the clinical significance. METHODS: This is a retrospective analysis of cases collected from our clinical MRI practice. Each case was evaluated for location, size, lobe of origin, associated signal abnormality, symptoms, and the presence or absence of intracranial hypertension. RESULTS: Twenty-one cases of BHAG identified in 16 patients. Seven cases (7/21, 33%) demonstrated signal abnormality in the herniated or underlying parenchyma. On serial imaging, the size of herniation either fluctuated (2/10, 20%) or stayed the same (8/10, 80%). The associated signal abnormalities increased (1/5, 20%), decreased (1/5, 20%), or remained stable (3/5, 60%). Four patients (4/16, 25%) had signs/symptoms that could be associated with the BHAG. Ten patients (10/16, 63%) had either imaging findings or clinical signs that could be associated with intracranial hypertension. CONCLUSION: In contrast to previously reported cases, our study demonstrates that BHAG can be associated with MRI signal abnormalities of the herniated and adjacent brain. Symptoms are typically absent, though may be present in a small subset of patients, regardless of the presence of signal change or encephalomalacia of the involved brain. Indirect evidence of intracranial hypertension is present in a majority of patients but completely lacking in others.

Keywords: Brain herniation, arachnoid granulation, encephalocele.

Acceptance: Received April 18, 2016. Accepted for publication April 28, 2016. Correspondence: Address correspondence to Greta B. Liebo, Department of Radiology, Mayo Clinic-Rochester, 200 First St SW, Rochester, MN 55905. E-mail: [email protected].

Presentation at Meetings: Portions presented in electronic poster exhibit at ASNR, 2015 Annual Meeting, Chicago, IL, April 25–30. “Brain Herniation into Arachnoid Granulations: an Uncommon Entity with Variable Clinical and Radiographic Manifestations.” Portions presented in electronic poster exhibit at ASHNR, 2015 Annual Meeting, Naples, Florida; September 9–13. “Arachnoid Granulations Involving the Underlying Brain: Imaging and Clinical Features.”

Acknowledgements and Disclosure: None.

J Neuroimaging 2016;00:1-7. DOI: 10.1111/jon.12366

Introduction condition.9–11,14,15 Symptoms attributed to these abnormalities Arachnoid granulations (AGs) are invaginations of the arach- are variable and largely anecdotal.2–4,9–11,13,15 noid membrane that protrude through gaps in the dura into In all previously reported cases, the MRI signal in the her- the dural venous sinuses and occasionally the calvarium. They niated brain parenchyma and surrounding CSF has been de- are functionally and histologically related to arachnoid villi, dif- scribed as isointense to the underlying brain and subarachnoid fering only in size and complexity, providing a route for cere- space. In our experience, these herniations are sometimes asso- bral spinal fluid (CSF) resorption into the bloodstream.1–4 Fre- ciated with parenchymal signal abnormality or focal encephalo- quently seen on imaging, AGs are considered normal anatomic malacia in the underlying brain. To further understand the structures, though rare cases have been reported that indirectly imaging and clinical significance of this increasingly recognized implicate AGs with symptoms such as headache, tinnitus, CSF entity, we performed a systematic analysis of all 21 examples of otorrhea, and intracranial venous hypertension.5–14 brain herniation into presumed AGs (BHAG) encountered in Recently, herniation of brain parenchyma into the dural our clinical practice, specifically correlating imaging and clini- venous sinuses and calvarium have been reported, sometimes cal findings to the presence or absence of parenchymal signal referred to as occult encephaloceles, as the herniation does abnormality in the herniation or underlying brain. not extend through the outer cortex of the calvarium or skull base.2–4,9–12,15–17 While it remains to be proven, the association Methods of the herniation with presumed preexisting AGs has been sug- Case Identification gested in prior reports. Some authors have suggested that the abnormality occurs spontaneously,2,10,11 while others have pos- Institution review board approval was obtained for this Health tulated that intracranial hypertension could be a predisposing Insurance Portability and Accountability Act compliant study.

Copyright ◦C 2016 by the American Society of Neuroimaging 1 Cases were identified by the authors during routine clinical scanned, though all had a slice thickness of <1.5 mm. A tabu- practice between 2012 and 2015. The diagnosis was based on lated overview of the cases can be found in Table 1. criteria established from previously reported cases of AGs on magnetic resonance (MR) imaging3,8,18 in conjunction with (i) Imaging Analysis direct visualization of brain herniation into the AG or (ii) teth- The majority of the herniations occurred within presumed AGs ering of brain to the AG with underlying parenchymal signal along inner table of the calvaria or skull base (16/21, 76%). abnormality and architectural distortion. All cases meeting di- Fourteen of 21 (67%) involved the occipital crests and 2/21 agnostic imaging criteria by group consensus were included. (10%) involved the posterior petrous bones. The remaining five Two cases were confirmed pathologically. (24%) herniations occurred into presumed AGs protruding into the dural venous sinuses, including four (19%) in the transverse Imaging Analysis sinus and one (5%) in the torcula herophili. Four of 16 (25%) pa- All images were reviewed separately by two neuroradiologists tients had multiple herniations: three cases with two herniations (GBL and VTL), the latter of whom holds a certificate of added each and one case with three separate herniations. All hernia- qualification in neuroradiology and 3 years’ experience. Al- tions in these 4 patients involved herniation of the cerebellar though only one MRI was requisite for inclusion, all available hemisphere into the occipital crests. cross sectional imaging studies of the brain were analyzed in All herniations arose either from the posterior fossa (cerebel- each patient. The utilization of high-resolution 3-dimensional lum) or the cerebrum immediately adjacent to it. Herniations of gradient echo or spinecho imaging was also recorded, defined the cerebellar hemispheres (n = 19, 90%) exceeded that of the by a maximum slice thickness of 1.5 mm. cerebral hemispheres (n = 2, 10%). Cerebellar hemisphere her- Each case was evaluated carefully to assess (1) AG loca- niations were on the right in 9 cases and on the left in 10 cases; tion; (2) mean AG diameter; (3) lobe of origin; (4) presence no cases of herniation of the cerebellar vermis were identified. or absence of signal abnormality in the herniated brain; or (5) Herniations arising from the cerebral hemispheres included one encephalomalacia in the underlying brain parenchyma. When case (4.5%) originating from the left inferior temporal gyrus and serial examinations were available, the stability or evolution of one case (4.5%) originating from the lateral left inferior occipital these features was documented. lobe. For patients with multiple presumed AGs containing herni- Overall, MR signal abnormality in the herniated and/or un- ated brain parenchyma, each area was analyzed separately. All derlying brain was identified in 7/21 (33%) cases of BHAG, present and prior examinations were then analyzed for recog- involving 7/16 (44%) of patients. This subset of patients had nized neuroimaging signs of possible intracranial hypertension a mean age of 58 years (range 45–70) and 5/7 (71%) were fe- including intraparenchymal mass effect, thinning or erosions of males. All cases involved the cerebellum. Two cases (2/7, 29%) the calvarium and skull base, empty sella, optic sheath pseu- demonstrated T2 prolongation in the herniated tissue sugges- domeningoceles, and narrowing of the dural venous sinuses. tive of simple vasogenic edema (cases 2 and 11; Fig 1). One case (1/7, 14%) demonstrated T2 prolongation and enhance- Clinical Chart Analysis ment in the tissue subjacent to cerebellar herniation suggestive of more severe vasogenic edema related to ischemic injury (case The electronic medical record was systematically reviewed by 8; Fig 2). Four cases (4/7, 57%) demonstrated T2 prolongation, a board-certified neurologic surgeon (JJVG) with 5 years of ex- fluid-attenuated inversion recovery (FLAIR) suppression, and perience and clinical interest in skull base CSF leaks. Clinical T1 shortening in the parenchyma subjacent to the presumed histories were reviewed for any past or present neurologic sign AG suggestive of focal cerebellar encephalomalacia. Three of or symptom that could conceivably correlate with the MRI four (75%) cases of encephalomalacia were located along the finding(s). Confounding neurologic events (ie, prior trauma or occipital crest (cases 9, 10, and 12; Fig 3) and 1/4 (25%) was stroke) were noted, as well as underlying neurologic or systemic located along the posterior petrous face (case 7; Fig 4). The illness that could obscure the clinical and/or MRI picture. The clinical features and imaging features over time in this subset of neurosurgeon then categorized each independent presumed patients are detailed in the respective sections below. AG as unassociated or possibly associated to patient’s sign or symptom based on his clinical impression. Signs and symptoms Imaging Characteristics of Herniation and Underlying Brain were subcategorized by whether they were thought to be due Parenchyma over Time to the herniated parenchyma or presumed AG itself. Finally, clinical histories were also reviewed for any signs or symptoms Ten of 16 (63%) patients with BHAG had serial imaging span- of intracranial hypertension in the past or present. ning a mean interval of 42.6 months (range 4 to 128 months). In 8 (8/10, 80%) of these patients with a mean follow-up time of 43.4 months, the brain herniation was present on all exami- Results nations, without significant change in size over time. In case In all, we identified 16 patients with a total of 21 presumed 11, herniation of parenchyma appeared to alternate between AGs containing herniated brain parenchyma within our clin- present and absent on serial consecutive examinations over a ical practice. There were 11 females and 5 males with an 72.5 month interval (Fig 1). In case 14, the cerebellar herniation age range of 5–83 years (median, 61.5 years; mean, 56 years). increased in size on serial examinations over a 6-month interval High-resolution imaging was performed on 14/16 patients and corresponding with increased mass effect from enlarging cere- all patients received intravenous gadolinium on at least one bellar metastases in the posterior fossa (Fig 5). These assertions examination. High-resolution sequences varied depending on were based on the review of serial consecutive axial T2- and the system manufacturer and indication for the patient being FLAIR-weighted images, upon which spatial resolution was felt

2 Journal of Neuroimaging Vol 00 No 0 xxxx 2016 Table 1. Tabulated Summary of Cases from Our Collective Clinical Experience

Imaging Serial IH or Patient Location Lobe abnormality Symptomatology Imaging Mass Effect

1 TS Temporal No Epilepsy NA Remote history of meningitis 2 TS Cerebellum Edema in herniation None NA No 3 OC Cerebellum No None Unchanged over Boney remodeling; partial 57.5 months empty sella 4 OC Cerebellum No None Unchanged over 15 Supratentorial tumor OC Cerebellum No months resection complicated by OC Cerebellum No hemorrhage requiring surgical decompression 5 Torcula Cerebellum No None NA No 6 PPF Cerebellum No CSF otorrhea NA Partial empty sella 7 PPF Cerebellum Subjacent Hearing loss, Unchanged over No encephalomalacia tinnitus, vertigo 13.5 months 8 OC Cerebellum Subjacent edema Debilitating HA, Edema and Boney remodeling; partial and enhancement neck pain, enhancement empty sella OC Cerebellum No diplopia, fatigue resolves over 4 months; herniation persists 9 TS Cerebellum Subjacent None Unchanged over Boney remodeling; partial encephalomalacia 42.5 months empty sella 10 OC Cerebellum Subjacent None NA No encephalomalacia 11 OC Cerebellum Edema in herniation None Herniation alternates History of bacterial on serial exams meningitis over 72.5 months; edema increases 12 OC Cerebellum Subjacent None Unchanged over 128 No encephalomalacia months OC Cerebellum No 13 OC Cerebellum No None NA No 14 OC Cerebellum No None Herniations increase Multifocal cerebellar OC Cerebellum No in size over 6 metastases; increasing months mass effect over time 15 TS Occipital No None Unchanged over 4 Vestibular schwannoma months resection complicated by venous infarction and hydrocephalus; partial empty sella 16 OC Cerebellum No None Unchanged over 83 Partial empty sella months

TS = transverse sinus; OC = occipital crest; PPF = posterior petrous face; HA = headache; NA = not applicable. to be sufficient to differentiate between structural differences with indeterminant symptoms that could have been attributable and differences related to slice selection. to either. Overall, 5/7 (71%) patients with abnormal MRI signal or en- In case 1, a 5-year old female presented with complex par- cephalomalacia of the involved parenchyma had serial imag- tial or secondarily generalized epilepsy in the setting of a left ing with a mean interval of 52.1 months (range 4 months to temporal lobe herniation into the transverse sinus. Electroen- 128 months). In case 8, the enhancing T2 signal subjacent cephalogram (EEG) findings were suggestive of a left frontal to the cerebellar herniation subsided over 4 months. In case origin, though the inferior temporal lobe is notoriously under- 11, the nonenhancing T2 hyperintensity within the herniated sampled by bipolar EEG montage19 and the possibility of the parenchyma increased over 72.5 months (Fig 1). In the remain- left temporal herniation being the epileptogenic source could ing three cases (cases 7, 9, and 12), focal encephalomalacia was not be entirely excluded. No other structural abnormalities were present on the earliest available imaging studies and remained identified. unchanged over the imaging intervals. In case 6, a 66-year old female presented with CSF otorrhea in the setting of focal erosion along the left posterior petrous face. At surgery, the dura was found to be thin but Clinical Chart Analysis intact, and a large meningoencephalocele was directly visual- Overall, 4 patients with BHAG (4/16, 25%) were identified ized in the mastoid. Erosion from a presumed AG was thought by neurosurgical assessment as having possible associated to be the source of the patient’s symptoms. AGs have long been symptoms—1 (25%) directly related to the brain herniation, 2 recognized as a causative factor in CSF otorrhea in the medical 14,20–22 (50%) directly related to the presumed AG location, and 1 (25%) literature.

Liebo et al: Brain Herniation into Arachnoid Granulations 3 Fig 1. Alternating degree of herniation over time. Case 11: Herniation of the left cerebellar hemisphere (→) into a presumed arachnoid granulation along the left occipital crest (arrowheads) alternates between present and absent over serial axial T2-weighted examinations, both preceding (A), during (B), and following (C-E) his episode of meningitis. Note the development of abnormal T2 signal within the herniation on the ﬁnal examination (E).

Fig 2. Enhancing T2-weighted signal abnormality involving the herniation and subjacent brain. Case 8: 45-year-old woman with 10 days of rapidly progressive headache, stiff neck, and diplopia. (A) Initial noncontrast head CT demonstrates large erosions of the occipital bones, typical in appearance for arachnoid granulations (AGs), but no evidence of acute intracranial abnormality. After 2 weeks, her symptoms spontaneously resolved, and a follow-up MRI was obtained. (B) Axial T2-weighted FLAIR and (C) sagittal postcontrast T1-weighted MR images demonstrate protrusion of the cerebellar hemispheres into the occipital AGs (arrowheads) with subjacent edema (B,→) and enhancement (C,→) within the herniation and along the medial right cerebellar hemisphere.

In case 7 (Fig 4), a 54-year old female presented with hearing postulated that clinical and imaging findings could have been loss, tinnitus, and vertigo in the setting of a cystic right cerebel- related to transient strangulation of the herniated parenchyma lopontine angle mass, erosion of the posterior petrous face, and verses transient venous congestion from mass effect of the her- adjacent encephalomalacia. The patient underwent resection niation on the adjacent torcula and transverse sinuses. and arachnoid and neuroglial tissues were found on pathologic Overall, 2/7 (29%) cases with signal abnormality of the herni- analysis, consistent with herniation into an AG. Given the lo- ated or underlying brain had possible associated signs or symp- cation, the presumed AG may have led to dysfunction of the toms, while the majority did not. None of the 4 patients with endolymphatic sac, which could account for the patient’s symp- focal cerebellar encephalomalacia subjacent to the presumed toms. AG had current or prior signs or symptoms that could be at- In case 8 (Fig 2), a 45-year old female presented for a second tributed to the cerebellar herniation or encephalomalacic brain opinion regarding an acute episode of debilitating head and parenchyma, including history of prior cerebellar stroke or tran- neck pain, diplopia, and fatigue in the setting of a herniation sient acute cerebellar symptoms for which the patient sought into the right occipital crest with underlying cerebellar T2 hy- medical attention. Additionally, both patients with cerebellar perintensity and enhancement. By the time of her appointment T2 hyperintensity without enhancement had no corresponding several weeks later, she reported complete spontaneous symp- signs or symptoms. tomatic resolution. The details of the prior outside hospitaliza- tion remain vague; however, she never received any pharma- Intracranial Hypertension cological treatment. By 4 month follow-up, the MRI signal and Ten patients (10/16, 63%) had either imaging or clinical ab- enhancement had also resolved. While never confirmed, it was normalities that could be associated with current or prior

4 Journal of Neuroimaging Vol 00 No 0 xxxx 2016 Fig 3. Focal encephalomalacia underlying herniations into the occipital crests. (A) Case 12: high-resolution T2-weighted axial MRI demonstrates focal encephalomalacia (→) subjacent to the left cerebellar herniation, which extends into a large arachnoid granulation (AG) along the left occipital crest. Additional smaller herniations of the right cerebellar hemisphere into right occipital AGs are also seen (arrowheads = presumed AGs). Case 10: (B) High-resolution T2-weighted coronal and (C) axial T2-weighted FLAIR MR images demonstrate focal encephalomalacia (→) subjacent to the a large AG along the right occipital crest (arrowheads). The herniated parenchyma appears atretic, presumably secondary to scarring/ﬁbrosis.

Fig 4. Focal encephalomalacia underlying herniation into a presumed arachnoid granulation along the posterior petrous face. Case 7: (A) Axial temporal bone CT demonstrates bony erosion along the posterior right petrous face (arrowheads). (B) Axial T2-weighted MRI demonstrates a nonenhancing multiloculated cystic mass within the right cerebellopontine cistern (arrowheads). Abnormal cerebellar T2 signal partially surrounds the lesion (→). (C) At 3-month follow-up, T2-weighted MRI shows complete resection of the cystic lesion with persistent subtle T2-weighted signal hyperintensity around the surgical margin (→) consistent with encephalomalacia. Pathologic analysis of the lesion yielded both arachnoid and neuronal elements.

Fig 5. Enlarging cerebellar herniations in association with local mass effect. Case 14: (A-C) Serial axial T2-weighted images over a 6-month interval demonstrate increased herniation of the bilateral cerebellar hemispheres (→) into presumed arachnoid granulations along the occipital crests. The increase in herniation correlates with growing metastases and worsening mass effect (arrowheads) in the posterior fossa. intracranial hypertension, but confirmatory pressure monitor- tumors with or without postoperative complications resulting ing was never performed. Six of these patients (6/10, 60%) had in transient mass effect and effacement of the local-regional indirect imaging findings suggestive of chronic intracranial hy- subarachnoid space. All three had serial imaging available: pertension, including thinning/erosion of the calvarium and In case 14, the cerebellar herniations increased in size in di- skull base and partial empty sella. Five of these 6 patients had rect correlation with increasing edema in the posterior fossa serial imaging available, upon which the herniations were sta- (Fig 5); in case 15, the cerebellar herniation was stable in size ble over time. Two patients (2/10, 20%) had a clinical history of despite the transient mass effect from postoperative infratento- meningitis suggestive of transient elevation of intracranial pres- rial venous infarction and hemorrhage; in case 4, the cerebellar sure. One of these patients had prior imaging available, and the herniations were also stable in size, though the mass effect from presence of the herniation seemed to alternate over time (case postoperative hemorrhage was largely limited to the supraten- 11; Fig 1). Three patients (3/10, 30%) had intraparenchymal torial compartment.

Liebo et al: Brain Herniation into Arachnoid Granulations 5 Discussion hemodynamics in the dural venous sinuses4,5,15 or eroding of The primary findings of this study are important for several the posterior petrous face.12–14 Our results support the assertion reasons. First, our findings identify several features that have that most BHAGs are asymptomatic, but associated signs or not been described previously to our knowledge including: (1) symptoms may be present in a small number of cases. Signs BHAG can be associated with imaging abnormalities of the or symptoms arising from the herniation are rare with patho- herniated and adjacent brain suggestive of edema, ischemic physiology likely mimicking that of other small encephaloce- injury, or encephalomalacia; (2) the degree of brain hernia- les previously reported to be associated with seizures or focal tion and signal abnormality of the underlying brain typically neurologic deficits.23–25 Perhaps, the low frequency of symp- remains stable but can occasionally vary with time; and (3) tomatic association, as well as the poor correlation of clinical these parenchymal abnormalities correlate poorly with semi- findings with BHAG-related edema or encephalomalacia, can ology and symptomatology. Second, our findings support the be explained by the proclivity of BHAG for the posterior fossa prior observation that these herniations have a marked procliv- and seldom involvement of eloquent cortex. BHAGs into the ity for the posteroinferior skull11 and further the understanding mastoid portion of the temporal bone via the posterior petrous of the relationship of BHAG with intracranial hypertension and faces are more frequently associated with symptoms, both in clinical symptoms. our study and in the literature, though these symptoms are typ- Prior to the large systematic study published by Battal ically secondary to erosion and CSF leak and unrelated to the et al11 in 2015, there were only a small number of case se- parenchymal herniation. ries and case reports in the English literature.2,4,5,9,10,13–17 Battal Some authors have proposed that BHAG occurs sponta- et al11 reviewed a total of 6,160 brain MRIs, identifying a total neously, while others have suggested that increased intracranial of 21 herniations into the calvarium or dural venous sinuses in pressure may be a predisposing condition.9–11,14,15 Our results 20 patients, establishing a prevalence of 0.32%. Battal et al11 were equivocal, with only half of patients having indirect imag- postulated that these occult herniations could occur sponta- ing findings or clinical history suggestive of chronic or transient neously, secondary to transient intracranial pressure shifts, or intracranial hypertension. Perhaps of greater interest were the due to preexisting AGs or penetrating veins that weakened the 2 patients with intraparenchymal mass effect and transient or dura at the herniation site. However, we consider the hernia- progressive effacement of the subarachnoid space in the same tions in our study to occur directly into preexisting AGs rather compartment as the BHAG. While one of these herniations re- than nonspecific weak points in the dura for several reasons. mained stable over time, the other increased in size in direct First, all cases of BHAG occurred in locations where AGs are correlation with the mass effect in the posterior fossa. Further common and the fluid surrounding the herniations was isoin- investigation is needed to confirm or refute a relationship be- tense to CSF, in direct communiation with the subarachnoid tween intracranial hypertension or local-regional mass effect space. Second, if the CSF-filled invaginations and erosions had and BHAG. Given the reported incidence of BHAG of well been found in these locations without herniated brain, they under 1%,11 it seems unlikely that a quarter of the patients in would have been likely referred to as AGs in clinical prac- our study would develop multiple lesions without some predis- tice. Thirdly, one resected lesion in the right cerebellopontine posing factor. angle yielded arachnoid and neuroglial elements on pathologi- Detection of BHAG likely will increase over time with im- cal analysis. Although CSF signal may vary within AGs without proved awareness and growing utilization of high-resolution herniation,2,3,8,11,18 the signal could potentially differ in AGs that MR techniques. The possibility of associated imaging abnor- can accommodate herniations. Since this assertion could be de- malities of the involved parenchyma raises the question of bated and pathology is lacking for most cases, we acknowledge follow-up or management recommendations. If asymptomatic, that the designation is presumptive. with or without abnormality of the herniated parenchyma, no The presence of MRI signal change within the herniated further follow-up may be needed. However, if symptomatic at or adjacent brain in our study stands in stark contrast to the presentation, there may be a role for follow-up imaging or, in current literature. The pathophysiology cannot be definitively rare cases, possibly elective surgical decompression. Further- established without pathologic correlation. Given the current more, it may be important to alert a surgeon to the presence results, we speculate that herniated brain may occasionally be- of a BHAG prior to surgical repair of posterior fossa CSF leaks come strangulated, resulting in simple vasogenic edema. More since this could affect operative approach. severe strangulation and subsequent ischemia may result in This study has a number of limitations. Since it is a retrospec- greater edema and breakdown of the blood-brain barrier. Our tive review of cases from our personal collection, it cannot be data indicates that this presumed ischemic change can resolve used to determine the absolute prevalence of any given imaging or progress over time. It also raises the possibility that it can or clinical finding. The potential relationship between BHAG evolve into focal encephalomalacia, although this progression and symptoms was based on clinical impression and thereby was not confirmed by serial imaging in any of our cases. Regard- presumptive, based on the opinion of one neurosurgeon alone. less, it is important to recognize these findings as possibly asso- Direct causal relationship or association cannot be confirmed. ciated with BHAG and to not confuse with other pathology that Assessment of possible intracranial hypertension was based on could prompt unnecessary work-up for alternate explanations. interpretation of secondary clinical and neuroimaging signs, Battal et al11 concluded that brain herniations into the du- never confirmed with intracranial pressure monitoring. Serial ral sinuses and/or calvarium are incidental findings with no exams were only performed on a subset of patients. Since cases definite symptomatic correlation, although the relationship to were obtained from encounters during routine clinical practice, headaches was considered uncertain.11 Other case reports have the imaging parameters are variable. More cases are needed suggested that BHAGs may cause symptoms by altering the before statistical analysis can be performed.

6 Journal of Neuroimaging Vol 00 No 0 xxxx 2016 Conclusion 10. Battal B, Castillo M. Brain herniations into the dural venous sinuses BHAG has a proclivity for the posteroinferior skull and can or calvarium: MRI of a recently recognized entity. Neuroradiol J be associated with imaging abnormalities of the herniated and 2014;27(1):55-62. 11. Battal B, Hamcan S, Akgun V, et al. Brain herniations into the adjacent brain suggestive of vasogenic edema, ischemic injury, dural venous sinus or calvarium: MRI findings, possible causes and chronic encephalomalacia. Symptoms correlate poorly to and clinical significance. Eur Radiol 2016;26:1723-31. the presence or absence of parenchymal signal abnormality, 12. VandeVyer F, Lemmerling M, De Foer B, et al. Arachnoid gran- and most cases are likely asymptomatic. Indirect evidence of ulations of the posterior temporal bone wall: imaging appear- intracranial hypertension is present in a majority of patients, ance and differential diagnosis. AJNR Am J Neuroradiol 2007;28: but is completely lacking in others. Direct mass effect from an 610-2. 13. Kim SW, Choi JH. Cerebrospinal fluid otorrhea caused by arach- adjacent space occupying lesion can contribute to the degree noid granulation. Korean J Audiol 2012;16:152-5. of herniation. Multiple BHAGs may be seen in a single patient, 14. Nadaraja GS, Monfared A, Jackler RK. Spontaneous cerebrospinal which raises the possibility of a predisposing factor. fluid leak through the posterior aspect of the petrous bone. J Neurol Surg B Skull Base 2012;73:71-5. 15. Coban G, Yildirim E, Horasanli B, et al. Unusual cause of dizzi- References ness: occult temporal lobe encephalocele into transverse sinus. Clin 1. Leach JL, Jones BV, Tomsick TA, et al. Normal appearance of Neurol Neurosurg 2013;115:1911-3. arachnoid granulations on contrast-enhanced CT and MR of the 16. Asadi H, Morokoff A, Gillard F. Occult temporal lobe en- brain: differentiation from dural sinus disease. AJNR Am J Neuro- cephalocele into the transverse sinus. J Clin Neurosci 2015;22(7): radiol 1996;17:1523-32. 1202-4. 2. Liang L, Korogi Y, Sugahara T, et al. Normal structures in the 17. Kocyigit A, Herek D, Balci YI. Focal herniation of cerebral intracranial dural sinuses: delineation with 3D contrast-enhanced parenchyma into transverse sinus. J Neuroradiol 2015;42:126-7. magnetization prepared rapid acquisition gradient-echo imaging 18. Roche J, Warner D. Arachnoid granulations in the transverse and sequence. AJNR Am J Neuroradiol 2002;23:1739-46. sigmoid sinuses: CT, MR, and MR angiographic appearance of a 3. Trimble CR, Harnsberger HR, Castillo M, et al. “Giant” arach- normal anatomic variation. AJNR Am J Neuroradiol 1996;17:677- noid granulations just like CSF?: Not!! AJNR Am J Neuroradiol 83. 2010;31:1724-8. 19. Rosenzweig I, Fogarasi A, Johnsen B, et al. Beyond the double 4. Karatag O, Cosar M, Kizildag B, et al. Dural sinus fill- banana: improved recognition of temporal lobe seizures in long- ing defect: intrasigmoid encephalocele. BMJ Case Rep 2013. term EEG. J Clin Neurophysiol 2014;31(1):1-9. Available at: http://casereports/bmj.com/content/2013/bcr-2013- 20. Gacek RR, Gacek MR, Tart R. Adult spontaneous cere- 201616. Accessed April 25, 2016. brospinal fluid otorrhea: diagnosis and management. Am J Otol 5. Kollar C, Johnston I, Parker G, et al. Dural arteriovenous fistula in 1999;20(6):770-6. association with heterotopic brain nodule in the transverse sinus. 21. Lee MH, Kim HJ, Lee IH, et al. Prevalence and appearance of the AJNR Am J Neuroradiol 1998;19(6):1126-8. posterior wall defects of the temporal bone caused by presumed 6. Arjona A, Delgado F, Fernandez-Romero E. Intracranial hyperten- arachnoid granulations and their clinical significance: CT findings. sion secondary to giant arachnoid granulations. J Neurol Neurosurg AJNR Am J Neuroradiol 2008;29(9):1704-7. Psychiatry 2003;74:418. 22. Yew M, Dubbs B, Tong O, et al. Arachnoid granulations of the 7. Kan P, Stevens EA, Couldwell WT. Incidental giant arachnoid temporal bone: a histologic study of dural and osseous penetration. granulation. AJNR Am J Neuroradiol 2006;27:1491-2. Otol Neurotol 2011;32(4):602-9. 8. Leach JL, Meyer K, Jones BV, et al. Large arachnoid granula- 23. Loumiotis I, Jones L, Diehn F, et al. Symptomatic left in- tions involving the dorsal superior sagittal sinus: findings on MR tradiploic encephalocele. Neurology 2010;75(11):1027. Available imaging and MR venography. AJNR Am J Neuroradiol 2008;29: at: www.neurology.org/content/75/11/1027/. Accessed April 25, 1335-9. 2016. 9. Chan WC, Lai V, Wong YC, et al. Focal brain herniation into 24. Froelich S, Botelho C, Abu Eid M, et al. Intradiploic encephalo- giant arachnoid granulation: a rare occurrence. Eur J Radiol cele in an adult: a case report and review of the literature. Neu- Extra 2011;78(2):e111-3. Available at: http://www.sciencedirect. rochirurgie 2006;52(6):551-4. com/science/article/pii/S1571467511000393. Accessed April 25, 25. Kosnik EJ, Meagher JN, Quenemoen LR. Parietal intradiploic en- 2016. cephalocele: case report. J Neurosurg 1976;44:617-9.

Liebo et al: Brain Herniation into Arachnoid Granulations 7