Brachial plexus magnetic resonance imaging differentiates between inflammatory neuropathies and does not predict disease course

Abstract Objective The main objective of this study was to evaluate the correlation between the distribution of brachial plexus magnetic resonance imaging (MRI) abnormalities and clinical weakness, and to evaluate the value of brachial plexus MRI in predicting disease course and response to treatment in multifocal motor neuropathy (MMN), Lewis‐Sumner syndrome (LSS) and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). Methods Sixty‐seven patients with an inflammatory neuropathy diagnosed at our tertiary referral center for neuromuscular diseases had undergone bilateral T2‐weighted short tau inversion recovery (STIR) MRI of the brachial plexus. We obtained clinical follow‐up data and scored all MRIs for abnormalities and the symmetry of their distribution. Results Brachial plexus MRI abnormalities were detected in 45% of the patients. An abnormal MRI did not predict disease course in terms of patterns of weakness, sensory disturbances or response to treatment. Within the spectrum of radiological abnormalities, asymmetrical clinical syndromes, MMN and LSS were significantly associated with asymmetrical radiological abnormalities, whereas symmetrical abnormalities predominated in CIDP (p < .001, phi 0.791). Conclusion T2 STIR brachial plexus MRI abnormalities correspond with the distribution of neurological deficits in inflammatory neuropathies, but do not correlate with specific clinical characteristics, response to treatment or disease course.


| INTRODUCTION
Hyperintensity and hypertrophy of cervical nerves and nerve roots using T2-weighted short tau inversion recovery (STIR) magnetic resonance imaging (MRI) techniques have been described in patients with various types of demyelinating inflammatory neuropathies, including multifocal motor neuropathy (MMN) (Van Es et al., 1997), chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) (Adachi et al., 2011;Sinclair et al., 2011;Van Es et al., 1997), Lewis-Sumner syndrome (LSS) (Rajabally, Knopp, Martin-Lamb, & Morlese, 2014; Van den Berg-Vos et al., 2000), and polyneuropathy associated with IgM monoclonal gammopathy (Eurelings, Notermans, Van de Donk, & Lokhorst, 2001). Abnormalities on MR imaging of the brachial plexus are an argument for inflammatory neuropathies, particularly when nerve conduction studies are not fully conclusive. Therefore, positive findings on brachial plexus MRI have been included in the diagnostic consensus criteria for CIDP and MMN (Joint Task Force of the EFNS and the PNS for CIDP, 2010; Joint Task Force of the EFNS and the PNS for MMN, 2010).
It is unclear whether patterns of brachial plexus MRI abnormalities are associated with specific inflammatory neuropathies. This is important as different neuropathies may require specific therapeutic strategies (Nobile-Orazio, 2014;Pestronk et al., 1988;van Schaik, van den Berg, de Haan, & Vermeulen, 2005). Comparative studies of brachial plexus MRI have not been performed. A number of smaller studies have mainly suggested that asymmetry is associated with MMN or LSS rather than CIDP, but others could not confirm this (Rajabally et al., 2014;Van den Berg-Vos et al., 2000;Van Es et al., 1997). Moreover, it is not known whether MRI abnormalities predict disease course or response to treatment.
We therefore investigated patterns of brachial plexus MRI abnormalities in patients within the spectrum of inflammatory neuropathies, together with clinical follow-up data of all patients who had undergone an MRI to study their disease course and response to treatment.

| Patients
All patients who were diagnosed with MMN, LSS or CIDP at the tertiary referral neuromuscular outpatient clinic of the University Medical Center Utrecht between 1996-2015 were screened and those who had undergone bilateral T2 STIR MR imaging of the brachial plexus were included in this study (Van Es, 2001

| Clinical examination
Description of the neurological examination at the time of the MRI and at the last follow-up as reported in patient files was used to assess

| Brachial plexus MRI
MRIs of the brachial plexus were made between 1999 and 2015 using a previously described protocol, including bilateral T2 STIR imaging in the coronalplane on a 0.5-1.5T MRI scan, with slice thickness of 3.0 mm, a b-value of 0 and echo time of 60-100 ms (Van Es et al., 1997). All MRIs were made before and after the administration of gadolinium. Bilateral coronal T2 STIR images of the brachial plexus were used to identify abnormalities, that is signal hyperintensity and cervical nerve (root) thickening. All available MRIs were scored by an experienced neuroradiologist and all MRIs were re-examined by two of the authors (BJ and JB). In case of doubt, an experienced neuroradiologist (DR), blinded for the diagnosis, was consulted. All abnormal MRIs were scored as symmetrical or asymmetrical by visual interpretation of nerve intensities and nerve thickness.
The locally appointed ethics committee of the UMC Utrecht approved this study and they gave dispensation for informed consent from the included subjects as it is a retrospective medical file investigation.

| Statistical analysis
Statistical analyses were carried out using IBM SPSS 22.0 software.
For continuous data, between-group comparisons were made using Kruskal-Wallis H testing with post hoc Mann-Whitney U or Student's t-testing, as appropriate. Bonferroni correction for multiple testing was applied. Categorical variables were analyzed using a chi-square or Fisher's exact test. The phi coefficient was used as a measure of association between binary variables. A p-value <.05 was considered statistically significant.

| Patients
We identified a total of 267 patients with an inflammatory neuropathy, of which brachial plexus MRI had been performed in 67 patients; 40 patients had MMN, nine patients had LSS and 18 patients had CIDP. Of the CIDP patients, six had a pure-motor variant. Patient characteristics are listed in Table 1.
Multifocal motor neuropathy patients were significantly younger than CIDP patients at the time of diagnosis (p < .001). Time to diagnosis was significantly longer in MMN and LSS than for CIDP (p < .001).

| Brachial plexus MRI
MRI abnormalities were detected in 30/67 (45%) patients (Table 2, Evaluation of the correlation between clinical and radiological asymmetry revealed that asymmetric MRI abnormalities were associated with asymmetric weakness in 12 of 14 patients with MMN (86%) and in all five patients with LSS (100%) (Table 3). Similarly, symmetry of abnormalities on MRI and on clinical examination coincided (9/11 (82%)) in CIDP patients.

| Correlation clinical disease course versus MRI
No significant differences in clinical presentation, response to treatment and follow-up data were detected between patients with and without abnormalities on MRI per diagnosis (Table 4).
Follow-up studies of all 67 patients who underwent an MRI revealed that the diagnosis was altered only twice, namely from MMN into CIDP.

| Pooled data results
Pooled data analysis (Table 5) showed that clinical and brachial plexus MRI asymmetry concurred in the majority of MMN and LSS cases (17 of 19 (90%)). MRI abnormalities were symmetrical in 10 of 11 patients with CIDP (91%). Combining the data from our study showed that clinical and radiological asymmetry coincided (p < .001, phi 0.791).

| DISCUSSION
This study shows that (a)symmetry of abnormalities on brachial plexus MR imaging is associated with the distribution of clinical findings in patients with an inflammatory neuropathy. Hence, symmetrical distribution of MRI abnormalities is strongly associated with CIDP, whereas asymmetry suggests a diagnosis of MMN or LSS. Asymmetry of brachial plexus MRI abnormalities is associated with persisting asymmetry of weakness at follow-up indicating that radiological asymmetry is not a snapshot in time in the evolution to a symmetric disease, that is CIDP. Moreover, presence of (patterns of) MRI abnormalities do not predict a beneficial response to treatment with immunoglobulins.
Brachial plexus MRI is an established technique in the diagnostic work-up of patients suspected of an inflammatory neuropathy, but its predictive value for disease course or the relevance of specific patterns of weakness have not been studied. Moreover, comparative studies combining MR data of MMN, LSS and CIDP patients have not been performed. Symmetric brachial plexus abnormalities have been described using high-resolution ultrasound (HRUS) in asymmetric inflammatory neuropathies (Beekman et al., 2005). This discrepancy in results between HRUS and MRI may be caused by the lack of standardized MRI or HRUS criteria defining abnormality and asymmetry, but may also reflect differences in diagnostic sensitivity and specificity.