MRI features to aid the identification of lateral temporal bone cephaloceles

Objectives: To evaluate ancillary MRI features which may aid the identification of lateral temporal bone cephaloceles (LTBCs). Methods: A retrospective cohort study analysed patients with MRI evidence of surgically confirmed spontaneous LTBCs as defined by intracranial contents traversing the tegmen tympani or mastoideum. Cases were identified from radiology and surgical databases. Two observers analysed three-dimensional T 2W temporal bone and whole brain imaging according to a priori criteria by consensus, with emphasis on the relationship of any adjacent cerebrospinal fluid (CSF) cleft to the defect. The contents, location, and clinical features of the LTBCs were recorded. Results: Eighteen patients (11 female, 7 male; mean age 59.3 years, age range 42–86 years) with 20 surgically confirmed spontaneous LTBCs (2 bilateral;16 unilateral) were evaluated. A temporal lobe sulcus or other CSF cleft extending to or traversing the defect was identified in 19/20 (95%) cases. Isointense CSF tympanomastoid signal was present in 41.2% cases, whilst superior semi-circular canal dehiscence was found in 40% of cephaloceles. At least two MRI features of idiopathic intracranial hypertension were seen in 38.9% patients. Cephaloceles were most commonly centred on the tegmen tympani (55%). Meningoencephaloceles were present in 95% cases. Conclusion: A temporal lobe sulcus or CSF cleft extending to or traversing the defect may aid the identification of LTBCs. Isointense CSF tympanomastoid signal, superior semi-circular canal dehiscence and MRI features of idiopathic intracranial hypertension are only present in under half of LTBCs. Advances in knowledge: The study details novel ancillary MRI features of LTBCs which may aid their identification.


IntroduCtIon
[6][7] Spontaneous LTBCs are thought to manifest from a multifactorial process involving increased intracranial pressure in conjunction with an anatomical structural predisposition at the site of pneumatisation. 3,8The condition has been frequently associated with idiopathic intracranial hypertension (IIH). 9,10agnosis of the condition is often difficult due to the nonspecific constellation of presenting signs and symptoms.Commonly reported symptoms include hearing loss and cerebrospinal fluid (CSF) otorrhea, whilst rarer clinical presentations include seizures. 2,11,12Patients are at risk of potentially life-threatening infections such as meningitis, so timely identification is required. 13aging has an important role in the prompt and accurate diagnosis of the condition and to prevent further complications.Whilst CT can identify temporal bone defects and soft tissue opacification, MRI is optimal for directly demonstrating a LTBC.Thin section spin echo or gradient recalled echo (GRE) T 2 W sequences are the mainstay for evaluation.However, detecting the herniation of intracranial contents may be challenging, particularly in the objectives: To evaluate ancillary MRI features which may aid the identification of lateral temporal bone cephaloceles (LTBCs).Methods: A retrospective cohort study analysed patients with MRI evidence of surgically confirmed spontaneous LTBCs as defined by intracranial contents traversing the tegmen tympani or mastoideum.Cases were identified from radiology and surgical databases.Two observers analysed three-dimensional T 2 W temporal bone and whole brain imaging according to a priori criteria by consensus, with emphasis on the relationship of any adjacent cerebrospinal fluid (CSF) cleft to the defect.The contents, location, and clinical features of the LTBCs were recorded.results: Eighteen patients (11 female, 7 male; mean age 59.3 years, age range 42-86 years) with 20 surgically confirmed spontaneous LTBCs (2 bilateral;16 unilateral) were evaluated.A temporal lobe sulcus or other CSF cleft extending to or traversing the defect was identified in 19/20 (95%) cases.Isointense CSF tympanomastoid signal was present in 41.2% cases, whilst superior semi-circular canal dehiscence was found in 40% of cephaloceles.At least two MRI features of idiopathic intracranial hypertension were seen in 38.9% patients.Cephaloceles were most commonly centred on the tegmen tympani (55%).Meningoencephaloceles were present in 95% cases.

Conclusion:
A temporal lobe sulcus or CSF cleft extending to or traversing the defect may aid the identification of LTBCs.Isointense CSF tympanomastoid signal, superior semi-circular canal dehiscence and MRI features of idiopathic intracranial hypertension are only present in under half of LTBCs.advances in knowledge: The study details novel ancillary MRI features of LTBCs which may aid their identification.
absence of a meningocele.It is therefore important to evaluate ancillary imaging features which aid the identification.][18] Recent studies have described certain clinical and radiological findings in spontaneous LTBCs but have not focussed specifically on the MRI features of this condition. 1,19e primary objective of the study was to determine ancillary MRI findings which aid the identification of surgically confirmed spontaneous LTBCs whilst the secondary objectives were to define their contents, location and clinical features.

Patient population
Institutional approval (project number 13453) to conduct a service improvement project was obtained and patient consent was waived.This was a retrospective cohort study undertaken in a tertiary referral centre.Searches of a radiology imaging system and a dedicated surgical database were conducted from 01/01/2006 to 09/02/2022.The Boolean search comprised the keywords 'cephalocele' AND 'petrous' OR 'temporal' OR 'ear' OR 'mastoid' .Exclusion criteria were cephaloceles secondary to trauma, surgery, tumours or developmental defects, cephaloceles at the petrous apex, previous tympano-mastoid surgery, those not confirmed surgically and those without three-dimensional (3D) T 2 W imaging of the temporal bones.The study flowchart shown in Figure 1 summarises the patient selection process.
During this period, 25 MRI studies were found from a dedicated surgical database and supplemented with 779 studies on an additional imaging search.Following this, 786 studies were excluded and MRI studies of 18 individual patients were included for analysis.Demographic data including age, gender and surgical history were collated.

Clinical and surgical data
Clinical features including presenting signs and symptoms were collected.A diagnosis of IIH was confirmed according to clinical documentation of the patient meeting the Modified Dandy Criteria. 20,21Surgical data including the operation date, site, surgical approach and type of cephalocele were collated.

MRI protocols and image analysis
Axial whole brain T 2 W imaging and 3D T 2 W dedicated axial imaging of the temporal bones was performed on 1.5 Tesla (T) or 3T MRI systems.With respect to the 3D T 2 W axial imaging, CISS (constructive interference in steady state) was performed on 1.5T MAGNETOM Aera or Avanto scanners (Siemens Healthcare; Erlangen, Germany) in 16/20 (80%) cases, with one case (5%) performed on a 3T MAGNETOM Skyra scanner (Siemens Healthcare; Erlangen, Germany), whilst 3D T 2 W SPACE (sampling perfection with application optimised contrasts using different flip angle evaluation) was performed on 3T MAGNETOM Skyra scanners (Siemens Healthcare; Erlangen, Germany) in 3/20 (15%) cases.Whilst there was some variation between MRI scanners, the most frequently applied parameters were repetition time 5.8 ms, echo time 2.6 ms, one signal average, FOV 180 x 180 mm, slice thickness 0.7 mm, flip angle 66 0 , acquisition matrix 320 × 320, 420 pixel bandwidth Hz/ pixel for the CISS sequence and repetition time 1400 ms, echo time 155 ms, two signal averages, FOV 150 x 150 mm, slice thickness 0.7 mm, flip angle 120 0 , acquisition matrix 320 × 320, 290 pixel bandwidth Hz/pixel for the SPACE sequence.Two head and neck radiologists with 30 years (SC) and 5 years (RS) experience analysed the MRI studies with the aid of multiplanar reconstructions and according to a priori criteria by consensus.
MRI ancillary features and associations of LTBCs were evaluated.Firstly, the presence of either an adjacent temporal lobe sulcus or another CSF cleft overlying the temporal lobe was recorded if it extended to (Figure 2) or traversed (Figure 3) the defect.Such a CSF cleft may extend to or traverse the margin of the defect (Figures 2a and 3a) or alternatively may extend to or traverse the centre of the defect (Figures 2c and 3c).The presence of adjacent sulcal widening or asymmetric enlargement of the adjacent intracranial subarachnoid space was also recorded.Secondly, in cases with 3D CISS sequences available, the tympanomastoid signal was compared to that of CSF (CSF isointense, partially CSF isointense or completely different relative to CSF signal) (Figure 4).Thirdly, the presence of SSCD was evaluated on oblique coronal and sagittal 3D multiplanar reformats.Fourthly, the presence of ≥2 MRI features of IIH was recorded (Table 1). 18,21Finally, the MRI features of the cephalocele were recorded with respect to the contents (meningocele or meningoencephalocele) and location (tegmen tympani, medial or lateral tegmen mastoideum, alone or in combination).

Statistical analysis
Descriptive statistics were applied using Microsoft   the defect is seen in Figure 5. Adjacent sulcal widening or asymmetric enlargement of the adjacent intracranial subarachnoid space was observed in 17/20 (85%) LTBCs.
Eight of 20 (40%) cases in the cohort were found to demonstrate SSCD on MRI (Figure 6).
At least two MRI imaging features of IIH were demonstrated in 7/18 (38.9%) patients.There was one patient (14.3%) with IIH imaging features without a corresponding clinical diagnosis.In total, nine (50%) patients were found to have clinically confirmed IIH.
Table 2 summarises the ancillary MRI features of spontaneous LTBCS in our cohort of cases.

dISCuSSIon
A high T 2 W CSF signal cleft was identified to extend to or traverse the bony defect in almost all (95%) surgically confirmed LTBCs on MRI and represents a key MRI feature which may aid their identification.Adjacent LTBC sulcal widening or asymmetric enlargement of an intracranial subarachnoid space was also seen in 85% of cases.Minor ancillary features such as tympanomastoid signal entirely isointense to CSF, SSCD and MRI features of IIH are only present in under half of LTBCs and their absence should therefore not deter the diagnosis.
The current radiological literature characterising the imaging features of spontaneous LTBCs is limited.Two recent studies investigating a cohort of spontaneous LTBCs predominantly focussed upon the clinical features and CT imaging findings. 1,19ur study adds to this literature by evaluating associated MRI features in a cohort of surgically confirmed cases which may aid their detection.Previous studies investigating the MRI appearances of cephaloceles more generally have included descriptions   regarding the associated traction of adjacent sulci towards a cephalocele.For example, this may manifest as an accompanying herniation of a pouch of CSF or asymmetric enlargement of the adjacent intracranial subarachnoid space. 3,14However, our study expands on this by specifically detailing the prevalence and description of a T 2 W CSF signal cleft extending to or traversing the defect in spontaneous LTBCs.
Additional ancillary MRI features were chosen for evaluation on the basis of known associations.Firstly, unilateral tympanomastoid effusion is known to be associated with LTBCs, both clinically and on imaging studies. 15,22GRE 3D T 2 W sequences such as CISS are optimal for the analysis of the differing fluid composition of tympanomastoid fluid due to the utilisation of heavier T2 weighting and increased signal-to-noise ratios. 23,24reen et al studied patients with temporal bone CSF leaks, although without cephaloceles in the majority of cases, and found that tympanomastoid signal isointense to CSF was 100% specific and 76% sensitive for the presence of a CSF leak. 24Interestingly, only 41.2% of our LTBCs demonstrated entirely CSF isointense tympanomastoid fluid.This discrepancy may relate to the partial or transient obstruction of the defect, in the presence of a cephalocele, resulting in an increased protein content of any tympanomastoid CSF leak. 25Secondly, SSCD and spontaneous LTBCs are both linked to thinning of the tegmen tympani and mastoideum. 16,17The high rate of SSCD (40%) in our cohort was similar to a recent study evaluating the prevalence in patients with lateral temporal bone cephaloceles, 26 and much higher than the estimated prevalence of 0.7-4.0% in asymptomatic individuals. 27,28Whilst SSCD was absent in the majority of cases, the finding is of potential clinical importance since concomitant SSCD and LTBC repair can be performed using either a transmastoid or middle fossa approach. 29,30Finally, there is a known relationship between skull base cephaloceles and IIH. 9,10,31In our cohort, 38.9% of patients had MRI features of IIH, whilst half had a clinically confirmed diagnosis.This is higher than previously described in a retrospective case-control study investigating the rate of spontaneous intracranial meningoceles in IIH patients. 9   MRI can also have an important role in delineating the intracranial contents within a cephalocele.Meningoencephaloceles accounted for the majority of cephaloceles in the study.No previous studies have delineated the type of cephalocele using MRI in a cohort of patients with spontaneous LTBCs.The finding is of particular importance since the type of cephalocele has implications on pre-operative planning and outcome. 32n addition, identifying the centre of the defect is another important factor in influencing surgical approach. 22The tegmen tympani was involved as the centre of defect, either alone or in conjunction with other sites, in the majority (80%) of LTBCs in our cohort.The rate of tegmen tympani involvement in our study is towards the higher end compared to recent studies. 1,19he non-specific constellation of clinical presenting features in patients with spontaneous LTBCs presents a diagnostic challenge.Hearing loss was the most common symptom and in line with the rate in other recent studies. 1,19CSF otorrhea and meningitis are recognised clinical presentations and featured in our cohort, whereas seizures did not. 1,19,33,34e authors appreciate there are some limitations to the methodology.Firstly, inherent to the nature of the retrospective study, there were some data points which were not possible to complete.For example, spontaneous temporal bone cephaloceles have been significantly associated with obesity. 35However, the body mass indices in most patients were not available.Secondly, the scoring of MRI features was performed by consensus rather than by multiple observers.Whilst a priori criteria were applied, the subjective nature of these observations would have benefited from independent scoring and calculation of agreement statistics.Thirdly, our proposed ancillary features have not been fully evaluated for their diagnostic utility since they were not applied to a control group.Finally, the low number of patients in our study somewhat limits the overall generalisability of the results.

ConCluSIon
In conclusion, a T 2 W CSF signal cleft extends to or traverses the temporal bone defect in almost all cases of surgically confirmed spontaneous LTBCs on MRI, whilst entirely CSF isointense tympanomastoid signal, SSCD and MRI features of IIH are only present in under half of LTBCs.Interrogating the MRI for CSF clefts at the interface of the temporal lobe and temporal bone tegmen may aid the detection of LTBCs in the appropriate clinical setting.

Figure 1 .
Figure 1.Flow chart summarising the patient selection process.

Figure 2 .
Figure 2. Lateral temporal meningocephaloceles with an adjacent temporal lobe sulcus or other CSF cleft extending to the margin of the defect.(a, c) Schematic diagrams depicting lateral temporal bone meningoencephaloceles involving an adjacent temporal lobe sulcus (striped) (a)extending to the margin of the defect or other CSF cleft overlying the temporal lobe sulcus (c) extending to the centre of the defect (interruption in solid line).(b, d) Coronal 3D CISS imaging of the temporal bones demonstrating examples of lateral temporal bone meningoencephaloceles (open arrows) involving either an adjacent temporal lobe sulcus (b) extending to the margin of the defect (arrow) or other CSF cleft overlying the temporal lobe sulcus (d) extending to the centre of the defect (arrowhead).3D, three-dimensional; CISS, constructive interference in steady state; CSF, cerebrospinal fluid.

Figure 3 .
Figure 3. Lateral temporal bone meningoencephaloceles with an adjacent temporal lobe sulcus or other CSF cleft traversing the defect.(a, c) Schematic diagrams depicting lateral temporal bone meningoencephaloceles depicting an adjacent temporal lobe sulcus (striped) traversing the margin of the defect (a) or other CSF cleft overlying the temporal lobe sulcus (c) traversing the centre of the defect (interruption in solid line).(b, d) Coronal 3D CISS imaging of the temporal bones demonstrating examples of right sided lateral temporal bone meningoencephaloeceles involving either a traversing adjacent temporal lobe sulcus (b) along the margin of the defect (arrow) or other CSF cleft overlying the temporal lobe sulcus (d) through the centre of the defect (arrowhead).3D, three-dimensional; CISS, constructive interference in steady state; CSF, cerebrospinal fluid.

Figure 4 .
Figure 4. Tympanomastoid signal abnormality in patients with lateral temporal bone cephaloceles.(a, b, c) Coronal 3D CISS imaging of the temporal bones demonstrating examples of tympanomastoid signal abnormality (arrows) relative to the surrounding CSF in patients with meningoencephaloceles.Examples of (a) entirely isointense, (b) partially isointense and (c) entirely different tympanomastoid signal relative to CSF are demonstrated.3D, three-dimensional; CISS, constructive interference in steady state; CSF, cerebrospinal fluid.

Figure 5 .
Figure 5. Coronal 3D CISS imaging demonstrates a left sided lateral temporal bone meningoencephalocele (arrow) centred upon the tegmen tympani.There is no associated CSF signal cleft extending to or traversing the defect, nor associated adjacent asymmetric sulcal widening.Small volume tympanomastoid signal abnormality (arrowheads) isointense relative to CSF is demonstrated.3D, three-dimensional; CISS, constructive interference in steady state; CSF, cerebrospinal fluid.

Figure 6 .
Figure 6.SSCD associated with a lateral temporal bone meningoencephalocele. Coronal oblique 3D CISS imaging demonstrates a left-sided SSCD (arrowhead; open arrows-intact superior semicircular canal) in a patient with a spontaneous lateral temporal bone meningoencephalocele and associated tympanomastoid signal abnormality (filled arrows) that is of different signal relative to CSF. 3D, three-dimensional; CISS, constructive interference in steady state; CSF, cerebrospinal fluid; SSCD, semi-circular canal dehiscence.

Figure 7 .
Figure 7. Bar chart demonstrating the anatomical distribution of the centre of the defect.

Table 1 .
Lists the specific MRI features of IIH that were investigated 18,21idiopathic intracranial hypertension.If at least two of the features were demonstrated, this was recorded as positive for IIH.18,21

Table 2 .
Summarising the ancillary MRI features of spontaneous LTBCS in our cohort of cases 3D, three-dimensional; CISS, constructive interference in steady state; CSF, cerebrospinal fluid; LTBCS, lateral temporal bone cephaloceles.
Authors acknowledge funding support from Wellcome/ Engineering and Physical Sciences Research Council Centre for Medical Engineering at King's College London (WT 203148/Z/16/Z); National Institute for Health Research Biomedical Research Centre at Guy's & St Thomas' Hospitals and King's College London; Cancer Research UK National Cancer Imaging Translational Accelerator (A27066); the UK Research & Innovation London Medical Imaging and Artificial Intelligence Centre.For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.