WHAT IS ALREADY KNOWN ON THIS TOPIC

The trigeminal nerve is one of the most involved cranial nerves in multiple sclerosis (MS). However, the prevalence of trigeminal involvement in MS varies considerably due to the resolution of MRI and biases inherent in cohort selection.

WHAT THIS STUDY ADDS

We enrolled 120 patients with MS who underwent 7T brain MRI scans from an ongoing China National Registry of Neuro-Inflammatory Diseases. We observed trigeminal lesions in 15.8% of patients with MS. Remarkably, 57.8% of the lesions were found in the root entry zone, and 26.9% of those exhibited unexpected central venous sign.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICYIntroduction

Cranial nerve involvement is a prominent characteristic of multiple sclerosis (MS) lesions, with the optic nerve being the most frequently affected, followed by the trigeminal nerve. The prevalence of trigeminal involvement varies widely, ranging from 2.9% to 23%, depending on MRI resolution and sequence selection but also disease duration.1–8 Additionally, patients with MS have a 15-fold increased risk of developing trigeminal neuralgia (TN), compared with the general neurological outpatient population.9

Trigeminal lesions predominantly concentrate along the cisternal portion of the trigeminal nerve, extending to the root entry zone (REZ) and even to the pontine-medullary nucleus.4 7 In a cohort of 1628 patients with MS, 26 of 28 patients with TN displayed a REZ plaque.1 Lesions of REZ exhibit a distinctive linear shape with their long axis parallel to the trigeminal nerve.4 10 This uniqueness of the shape and anatomical location questions the underlying reasons behind the development of such lesions along the trigeminal root to REZ.

Histologic evidence has demonstrated focal demyelination at REZ in MS-related TN. A general consensus exists on the close association of the trigeminal nerve root with vascular anatomy, which may trigger TN.11 The utilisation of ultra-high-field 7T MRI, with its increased signal-to-noise ratio (SNR) and enhanced susceptibility effects, yields superior visualisation of the lesions and provides further insights into their possible pathology. Previous studies have mainly focused on MS-related TN with consecutive routine MRI scans, leading to potential underestimation of the prevalence of trigeminal nerve lesions in asymptomatic patients. To address this gap in knowledge, we aimed to determine the prevalence of trigeminal lesions using ultra-high-resolution MRI based on data from the China National Registry of Neuro-Inflammatory Diseases (CNRID) and explore characteristic imaging features.

Materials and methodsParticipants

We conducted an ongoing CNRID to investigate Chinese patients with inflammatory demyelinating disease. Further details about this cohort are available in ClinicalTrials.gov (Identifier: NCT05154370). Ethical approval for this study was obtained from the Ethical Committee of Beijing Tiantan Hospital, and written informed consent was obtained from all participating individuals according to the Declaration of Helsinki. For our investigation, we enrolled a total of 120 consecutive patients with MS from CNRID between December 2021 and May 2023. The inclusion criteria were as follows: age 18 years or older, a diagnosis of MS according to the 2017 McDonald criteria and no clinical relapse within the preceding 3 months. Detailed data on disease duration, Expanded Disability Status Scale (EDSS) score, TN and sensory symptoms were also collected.

MRI acquisition

All participants underwent 7T MR system (MAGNETOM Terra, Siemens Healthcare, Erlangen, Germany) using a 32-channel Rx/8-channel Tx head coil (Nova Medical, Wilmington, Massachusetts, USA). The scanning protocol includes a 3D T1-weighted magnetisation-prepared rapid acquisition gradient echo (MPRAGE) sequence to cover the whole brain (Repetition Time (TR)=2200 ms, Echo Time (TE)=3.0 ms, Inversion Time (TI)=1050 ms, matrix=320×320, isotropic resolution of 0.7×0.7×0.7 mm3, total acquisition time=6 min 29 s), a FLAIR sequence (TR=6600 ms, TE=95 ms, TI=2200 ms, flip angle=145°, in-plane resolution of 0.3×0.3 mm2, slice thickness of 2 mm, total acquisition time=5 min 18 s) and a multiecho T2*-weighted (T2*W) spoiled gradient echo sequence (TR=2500 ms, TE=10, 20, 30 ms, flip angle=50°, in-plane resolution of 0.1×0.1 mm2, slice thickness of 1.2 mm, total acquisition time=9 min 34 s) to cover the infratentorial brain. Also, fluid and white matter suppression (FLAWS) based on the magnetisation-prepared 2 rapid acquisition (MP2RAGE) sequence (FLAWS-MP2RAGE) was adopted for better-detecting lesions12 (TR=5000 ms, TE=1.44 ms, TI1/TI2=700/1700 ms, flip angles=4/5°, field of view=192×192 mm2, isotropic resolution of 0.75×0.75×0.75 mm3, total acquisition time=8 min 52 s). The FLAWS-MP2RAGE sequence produces three sets of images including two inversion times (INV1 and INV2) and a corrected image (UNI),13 and the FLAWS image was calculated using the pixel-wise minimum of INV1 and INV2 (figure 1).

12 ,13

Figure 1

Representative image of trigeminal involvement in FLAWS. The arrowhead illustrated the cisternal trigeminal nerve, root entry zone (REZ) and nuclear zone (patient 15). From left to right are INV2, INV1, UNI and FLAWS images from the FLAWS-MP2RAGE sequence. FLAWS-MP2RAGE, fluid and white matter suppression based on the magnetisation-prepared 2 rapid acquisition gradient echoes.

Data processing

Lesions were identified using T1-MPRAGE, T2-FLAIR and FLAWS images. Additionally, FLAIR* and T2*W images were employed to identify central vein signs (CVSs) in the trigeminal lesions.14 The FLAIR* was generated using the FLAIR and the third echo of T2*, by coregistering the FLAIR to T2* and performing image multiplication using the FSL toolbox (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/).15 MPRAGE-based magnetic resonance angiography (MRA) was adopted as an alternative imaging technique to evaluate neurovascular contact together with T2.16 The affected area of the trigeminal nerve was divided into three parts: the cisternal segment, REZ and nuclear zone, based on anatomical and signal abnormalities (figure 1). The cisternal segment begins where the trigeminal nerve enters the prepontine cistern, and REZ measures approximately 2 mm long and is located approximately 5–7 mm from the surface of the pons. All MRI images were independently analysed by two experienced neuroradiologists (LS and D-CT), with 8 and 11 years of neuroradiology experience, respectively. These MRI readers were blinded to patients’ clinical data.

Statistical analysis

Statistical analysis was performed using SPSS for Windows V.25.0 software (SPSS, Inc, Chicago, Illinois, USA). Descriptive statistics were primarily adopted in this study. Frequency distribution was used for lesion location, including the cisternal segment, REZ and the pontine-medullary nucleus. For continuous data, age was reported as mean±SD, while EDSS and disease duration were presented as median with range. Categorical variables were expressed as absolute and relative frequencies.

Data availability

The deidentified data will be available for investigators following approval from the Institutional Review Board of China National Clinical Research Center for Neurological Diseases (Beijing, China).

Results

A total of 120 participants with definite MS (111 relapsing-remitting, 9 progressive) underwent 7T MRI scans. 20 patients (16.7%) exhibited trigeminal involvement, with 19 having relapsing-remitting MS and 1 with secondary progressive MS. The female to male ratio was 1:1. The median disease duration was 3.5 years, with a range from 0.3 years to 15 years. The median EDSS score was 2.0, with a range from 0 to 7.0. At the time of imaging, seven patients received siponimod, one fingolimod, two ofatumumab, two teriflunomide and two dimethyl fumarate. Among the 20 patients with trigeminal involvement on the MRI, 6/20 (30%) reported paresthesia, primarily located in the maxillary division of the trigeminal nerve, while only 2 patients suffered from TN. Detailed clinical characteristics of each patient are listed in table 1.

Table 1

Clinical and trigeminal nerve involvement characteristics

The affected locations of the trigeminal nerve were clearly visualised on the FLAWS sequence, extending from the cistern segment to the trigeminal nucleus (figure 1). We observed abnormal signals in cisternal trigeminal nerve in 9/20 (45%) patients but did not classify them as trigeminal lesions. Of the 120 patients, trigeminal high signals on FLAWS were detected in 20 patients. 19/120 (15.8%) patients had a total of 45 trigeminal lesions in REZ and pontine-medullary nucleus, of which 11 of 19 (57.9%) were bilateral. Among these lesions, 24/45 (53.3%) lesions were located on the right, and 21/45 (46.7%) lesions were on the left side. The classification of trigeminal involvement is shown in figure 1. The lesions observed extended along the trigeminal nerve, with 26/45 lesions (57.8%) in REZ and 19/45 lesions (42.2%) in the pontine-medullary nucleus. Meantime, 8/9 (88.9%) of the abnormal signals in the cisternal trigeminal nerve are associated with either a REZ or pontine lesion (table 1).

A characteristic CVS was observed inside trigeminal lesions in 7 of 20 (35%) of patients, especially in REZ. A CVS was present in 7 of 26 (26.9%) of the lesions located in REZ (figure 2). The neurovascular contact was also evaluated via MPRAGE-based MRA, and four patients displayed neurovascular compression (figure 3). The interobserver agreement of trigeminal lesions and CVS was 0.93 and 0.90, respectively, while the intraobserver agreement was 0.94 and 0.95, respectively.

Figure 2

Central vein sign (CVS) in trigeminal lesions of multiple sclerosis on 7T MRI. (A) Representative image of trigeminal lesions and CVS (arrowheads) in patient 10. Sagittal (left panel) and axial (upper, right panel) T1 MPRAGE reveal lesions of the trigeminal root entry zone (REZ) (arrowheads). The dark vein can be visualised in axial and sagittal T2*-weighted image (arrowheads in middle and lower of right panel). (B) FLAIR* axial images of patients 3, 5, 6, 7, 13 and 15 demonstrate a dark vein located centrally in the majority of the trigeminal lesion (arrowheads), especially in REZ. MPRAGE, magnetisation-prepared rapid acquisition gradient echoes.

Figure 3

7T MPRAGE-MRA and T2-weighted image on the axial plane demonstrate neurovascular compression in patients 1, 4, 9 and 10. The arrowhead indicates the vessels touching the trigeminal nerve. MPRAGE-MRA, magnetisation-prepared rapid acquisition gradient echoes-magnetic resonance angiography.

A comprehensive review of PubMed using the keywords ‘multiple sclerosis’ and ‘trigeminal lesion’ allowed us to summarise the prevalence of trigeminal nerve involvement and the sites of injury in table 2.

Table 2

Summary of trigeminal nerve involvement in MS

Discussion

In this pioneering 7T MRI consecutive cohort of patients with early MS in China, we observed a notable proportion of trigeminal involvement, with 15.8% of patients exhibiting such lesions with most of these lesions being clinically silent. Moreover, a CVS was detected in 26.9% of lesions of REZ.

The prevalence of trigeminal nerve involvement in MS varies from 2.9% to 23%, owing to differences in selection bias in various cohorts and MRI resolution. In fact, most cohorts studying the prevalence of trigeminal involvement primarily focused on a history of MS-related TN, which introduces potential bias.2 6 7 Indeed, some patients with trigeminal nerve lesions may not manifest symptoms until several years after the lesion is detected on MRI, consistent with our finding that some patients had trigeminal nerve lesions without having TN.6

While prior studies used 0.1–3T MRI, we used 7T MRI that has high signal-to-noise ratio which significantly improves spatial resolution, thereby enabling the accurate detection of anatomical and pathological features. The combination of FLAWS with TI MPRAGE and FLAIR in our study facilitated the visualisation of both white and grey matter lesions, enhancing the identification of MS lesions. A recent study has verified the efficacy of 3D FLAWS to accurately diagnose infratentorial lesions in MS, thereby reducing false-negative findings.12

Regarding lesion distribution in our study, bilateral involvement of the trigeminal nerve was observed in 57.9% of the patients who had trigeminal nerve lesions. Notably, 57.8% of trigeminal nerve lesions in our cohort were located in the REZ region, which is the most commonly affected location in the literature, accounting for 51% to 100% of trigeminal nerve lesions (table 2). Histologic evidence has demonstrated focal demyelination at REZ from rhizotomy specimens, yet a consensus on the site specificity of trigeminal nerve demyelination in MS remains elusive.11 While vascular compression has been considered as an aetiological factor for TN, other possibilities include a demyelinating plaque within trigeminal REZ with coexisting neurovascular compression.9 17 18 However, many cases of MS-related TN have not shown apparent vascular compression.14

Notably, our study highlights the presence of MRI-detectable CVS in a linearly shaped demyelinating plaque located along the intramedullary trigeminal root, indicative of a central (oligodendroglia type of) myelin lesion. 57.8% of the lesions were located in REZ, of which 26.9% had a CVS. REZ represents the transition zone from central myelin to peripheral myelin. 88.9% of the abnormal signals in the cisternal trigeminal nerve are associated with either a REZ or pontine lesion, suggesting that the cisternal trigeminal nerve involvement may be a secondary change. There, oligodendrocytes supplying insulating myelin to the nerve fibres are gradually replaced by Schwann cells. This transition zone is more susceptible to neurovascular compression.11

The MRI-detectable CVS has been established as a biomarker of inflammatory demyelination.14 This vein serves as a privileged site for immune cell interactions that triggers an inflammatory cascade, leading to the formation of demyelinating lesions around the vein. The pathological basis of a classic trigeminal lesion with CVS has been attributed to the close proximity of blood vessels to the trigeminal root, with the peripheral Schwann cell transitioning to the central sheath of oligodendroglia in REZ.

Our study has certain limitations. First, FLAIR and T2*W images were not obtained volumetrically and isotropically and may result in producing suboptimal FLAIR* image. Future study will consider the optimised protocol with high-resolution isotropic 3D acquisition for both FLAIR and T2*. Second, high-resolution 7T Time-of-Flight (TOF) MRA should also be included in the protocol to better evaluate neurovascular compression. Furthermore, we did not describe the involvement of cranial nerves other than the trigeminal nerve, warranting the development of cranial nerve scanning protocol and further improvements in the resolution of infratentorial lesions in future investigations.

The strengths of our study include the use of 7T imaging in consecutive patients with relatively short MS duration. Moreover, the study is notable for including all patients regardless of trigeminal nerve symptoms or frank TN. In addition, it is a blinded reader evaluation of CVS with two independent investigators.

In conclusion, our study confirmed the high prevalence of trigeminal nerve on 7T MRI and highlighted the presence of a CVS in trigeminal nerve lesions. This study contributes to a deeper understanding of the pathophysiology and location-specific nature of trigeminal lesions, providing valuable insights into the high prevalence of trigeminal nerve involvement in MS. Future research should delve into trigeminal nerve lesions as one of the MS-specific lesion location for dissemination in space.

Abstract translationThis web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.Data availability statement

Data are available upon reasonable request. The deidentified data will be available for investigators following approval from the Institutional Review Board of China National Clinical Research Center for Neurological Diseases (Beijing, China).

Ethics statementsPatient consent for publication

Consent obtained directly from patient(s).

Ethics approval

This study involves human participants and was approved by the Ethics Committee of Beijing Tiantan Hospital, Capital Medical University, Beijing, China (KY2021-150-01). Participants gave informed consent to participate in the study before taking part.

Acknowledgments

We recognise colleagues from the Beijing-Tianjin Center for Neuroinflammation and Tiantan Neuroimaging Center of Excellence for their assorted support, and Elaine Shi for editorial assitence . We thank our patients for participating in the study.