Article Text

Download PDFPDF

Neuromuscular disease
MRI shows increased sciatic nerve cross sectional area in inherited and inflammatory neuropathies
  1. C D J Sinclair1,2,
  2. M A Miranda2,3,
  3. P Cowley2,
  4. J M Morrow1,
  5. I Davagnanam2,
  6. H Mehta2,
  7. M G Hanna1,4,
  8. M Koltzenburg1,5,
  9. M M Reilly1,4,
  10. T A Yousry1,2,
  11. J S Thornton1,2
  1. 1MRC Centre for Neuromuscular Diseases, UCL Institute of Neurology, London, UK
  2. 2Academic Neuroradiological Unit, Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, UK
  3. 3Division of Radiology, Department of Medicine, School of Medicine, University of Panama, Panama City, Panama
  4. 4Department of Molecular Neurosciences, UCL Institute of Neurology, London, UK
  5. 5Sobell Department of Movement Neuroscience, UCL Institute of Neurology, London, UK
  1. Correspondence to Dr C D J Sinclair, Box 65, National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, UK; c.sinclair{at}


Measurements of the cross sectional area of the sciatic nerve are described in a group of 10 patients with genetically confirmed Charcot–Marie–Tooth disease type 1A (CMT1A), nine patients with chronic inflammatory demyelinating polyneuropathy (CIDP) and 10 healthy controls using MRI. One mid-thigh of each individual was imaged using a short tau inversion recovery sequence and the nerve appearance evaluated radiologically with respect to the signal intensity and visibility of the internal neural structure. The cross sectional area of the sciatic nerve of each individual was measured by defining irregular enclosing regions of interest on the MRI images. The sciatic nerve area was enlarged in both CMT1A (p<0.001) and CIDP (p=0.008) compared with controls and in CMT1A compared with CIDP (p<0.001). Median (interquartile range) areas were 67.6 (16.2) mm2 for the CIDP group, 135.9 (46.5) mm2 for the CMT1A group and 43.3 (19.9) mm2 for the control group. The critical upper value for discriminating pathologically enlarged nerves from normal controls with p<0.05 was 64.4 mm2. Quantification of sciatic nerve hypertrophy on MRI may be of assistance in cases where the diagnosis is still in doubt, providing an objective pathological marker complimenting other clinical investigations.

Statistics from


MRI of nerves is becoming increasingly important in the diagnosis and evaluation of peripheral neuropathies due to its excellent soft tissue contrast and sensitivity to pathology.1 2 In this work, we evaluated the sciatic nerve in two representative peripheral neuropathies: the hereditary disease Charcot–Marie–Tooth disease type 1A (CMT1A) and the acquired condition, chronic inflammatory demyelinating polyneuropathy (CIDP).

Charcot–Marie–Tooth disease describes a genetically transmitted group of peripheral neuropathies3 with an estimated prevalence of 1 in 2500, characterised by distal muscle weakness and atrophy, impaired sensation and diminished deep tendon reflexes. The most common form, CMT1A, results from duplication of the PMP22 gene on chromosome 17. Nerve hypertrophy in CMT1A has been demonstrated on MRI in cranial nerves,4 spinal nerve roots and cauda equina.5

CIDP is an immune mediated and treatable condition characterised by a chronic relapsing course of sensorimotor involvement of all four limbs.6 Hypertrophy and gadolinium enhancement have been described in CIDP on MRI of spinal nerve roots,7 cauda equina,8 brachial and lumbosacral plexuses,9 and cranial nerves.10 These changes are included as a supportive feature in the European Federation of Neurological Societies/Peripheral Nerve Society (EFNS/PNS) diagnostic criteria.11

Here we used MRI to examine the size of the sciatic nerve in groups comprising CIDP and CMT1A patients and healthy controls.


Ten patients with genetically confirmed CMT1A (mean age±SD 42.6±13.7 years), nine patients with CIDP (59.4±10.3 years) and 10 healthy controls (47.0±13.4 years) were studied. Seven of the CIDP patients were classified as definite CIDP according to the EFNS/PNS diagnostic criteria.11 The other two patients were classified as definite CIDP associated with concomitant disease due to a monoclonal gammopathy of undetermined significance.11 All subjects gave informed consent and the study was performed with the approval of the local ethics review committee. As an indication of disease severity, lower limb muscle strength was assessed using the Medical Research Council (MRC) grading scale12 by a single examiner (MMR).

MRI was performed at 1.5 T (General Electric Signa Lx, Milwaukee, Wisconsin, USA). A T2 weighted short inversion time (tau) inversion recovery (STIR) sequence with repetition time, echo time and inversion times (TR/TE/IR) of 6000 ms, 65 ms and 150 ms, respectively, was used to obtain a set of images of one thigh of each participant. Seventeen axial slices of 6 mm thickness and 6 mm separation were acquired with two signal averages, a 512×512 matrix and 25.6 cm field of view yielding an in-plane resolution of 0.5 mm. A ‘flex’ RF surface coil was used for signal reception. The most affected limb was imaged in asymmetrically affected patients. In CIDP, strength assessment was made within a week of the MRI, while in CMT, where ankle dorsiflexion has been shown to deteriorate only very gradually,13 their most recent clinical assessment was used.

Two radiologists (ID and PC) measured independently the area of the sciatic nerve for each subject, blinded to the diagnosis. The sciatic nerve cross sectional area was measured by defining an irregular region of interest around the nerve boundary on a slice chosen to most clearly depict the point of maximum area. The arithmetic mean of the two observers' measurements was calculated. A third radiologist (TAY), also blinded to diagnosis, evaluated the nerve appearance on the STIR images with respect to signal intensity (markedly hyperintense, mildly hyperintense and isointense) and fascicle visibility.

Statistical analyses were performed using SPSS 14. Pearson correlation and Bland–Altman analysis assessed interobserver agreement for the nerve cross sectional area. The Kruskal–Wallis test and the Mann–Whitney U test were used for intergroup comparisons. A Bonferroni corrected p value of 0.017 was considered significant, accounting for three post hoc comparisons.


Patient characteristics

Clinical data for the CMT1A and CIDP patients studied are summarised in table 1. There was a wide range of severity in the CMT1A patients, with ankle dorsiflexion strength ranging from 0 to 4 (MRC grade). The CIDP patients had more asymmetry in ankle dorsiflexion strength than CMT1A patients. One CIDP patient had normal ankle dorsiflexion strength at the time of the MRI scan but this varied with treatment and was MRC grade 4 bilaterally at time of recruitment. All of the CIDP patients were receiving intravenous immunoglobulin at regular intervals. All but one of the CMT1A patients had onset of symptoms in childhood.

Table 1

Summary of the demographic and clinical features of the patients studied

MRI appearance

The sciatic nerves of many of the CIDP and CMT1A patients were highly conspicuous on the STIR images, characterised by notable hypertrophy (greatest in the CMT1A patients) and conspicuity of the internal fascicular structure. Representative images of a control subject, CIDP patient and CMT1A patient are shown in figure 1A–C). The CIDP and CMT1A nerves are clearly visible and distinct from the surrounding musculature.

Figure 1

Axial STIR images at the mid-thigh level in representative (A) control, (B) chronic inflammatory demyelinating polyneuropathy (CIDP, patient No 16) and (C) Charcot–Marie–Tooth disease type 1A (CMT1A, patient No 2) cases. The sciatic nerve (white arrow) is conspicuously enlarged in both patients, the greater hypertrophy being in CMT1A. Details of the sciatic nerve are shown as insets. (D) Scatterplot summarising the sciatic nerve cross sectional areas of the control, CIDP and CMT1A subjects. Horizontal lines are median values for each group (*p=0.008, **p<0.001).

In contrast with the patients, no controls displayed markedly hyperintense sciatic nerve signal, as rated by the blinded radiologist. Six controls were rated as mildly hyperintense and four as isointense. In the patient groups, the distribution was more varied, with a mixture of isointense (three CIDP, four CMT1A), mildly hyperintense (two CIDP, two CMT1A) and markedly hyperintense (four CIDP, four CMT1A) signals compared with the background tissue. It was possible to visualise the sciatic nerve internal fascicular structure in six controls, seven CIDP cases and seven CMT1A cases. The remaining cases either showed no fascicular structure or, in two cases, where fascicle visibility was recorded as indeterminate, suboptimal image quality prevented a confident assessment of the nerve internal anatomy.

Sciatic nerve cross sectional area: interobserver agreement

The independent measurements of sciatic nerve cross sectional areas from the two observers were highly correlated across subjects, with a Pearson coefficient of 0.95 (p<0.01). Bland–Altman analysis yielded a mean bias between observers of 2.71 (95% CI −3 to 8.5) mm2 and upper and lower limits of agreement of 29.4 (19.5 to 39.4) and −26.4 (−27.7 to −25.0) mm2, respectively.

Sciatic nerve cross sectional area: group comparisons

Group median nerve areas were significantly different (p<0.001 with χ2(2)=20.1; Kruskal–Wallis test against the null hypothesis of identical median values for the three independent samples) and post hoc analysis revealed that that nerve areas of both the CMT1A and CIDP groups were significantly greater than controls (U test, U=1.0, p<0.001 and U=13.0, p=0.008, respectively), and those of the CMT1A group significantly greater than those of the CIDP patients (U=5, p=<0.001).

Under the hypothesis that these neuropathies cause only nerve enlargement rather than reduction, the control group data distribution suggested that for a one tailed z score of 1.643 and p=0.05, the upper limit of normal area would be 64.4 mm2. Using this upper cut-off value gives respective values for sensitivity and specificity of 55.6% and 90% for CIDP and 100% and 90% for CMT1A. The equivalent upper threshold of nerve size for the CIDP group is 87.5 mm2 (z=1.643, p=0.05) which would discriminate between CMT1A and CIDP with a sensitivity of 90% and a specificity of 89%.

No association was observed between nerve area and duration of disease since onset (CIDP: Spearman's ρ =−0.18, p=0.64, CMT1A: ρ=0.12, p=0.75) or with the strength of ankle dorsiflexion in the imaged limb (CIDP: ρ=−0.21, p=0.58, CMT1A: ρ=−0.34, p=0.75).


STIR-MRI of the sciatic nerve revealed nerve hypertrophy and hyperintensity in both CMT1A and CIDP patients compared with healthy control subjects. There were significant differences in the median sciatic nerve cross sectional areas between the three groups. Nerve hypertrophy was greatest in the CMT1A group whereas cross sectional areas were more modestly increased in the CIDP group.

The CMT1A group was well separated from the controls, with all cases exhibiting nerve areas greater than the critical upper value of 64.4 mm2. A previous MRI study of sciatic nerve size described three CMT1A patients with nerve cross sectional areas of 40.8, 48.5 and 80 mm2, respectively, and a single control patient with a sciatic nerve area of 50 mm2.5 Ultrasonography (US) has also been used to assess nerve area in patients with CMT1A demonstrating hypertrophy in the ulnar14 and median14 15 nerves. However, an earlier US study of 10 patients with CMT (six with CMT1A), examining the radial, ulnar, median and sciatic nerves, failed to show a significant difference in nerve size compared with reference ranges based on 50 healthy controls. Anatomical limitations meant that the sciatic nerve could not always be identified on US.16 In contrast, MRI is particularly well suited to the robust evaluation of deep nerves.

This is the first study to demonstrate a statistically significant increased sciatic nerve cross sectional area in patients with CIDP versus normal controls, albeit with a substantial overlap between groups. In a single reported CIDP case, the sciatic nerve cross sectional area was 44.5 mm2 compared with a control patient with an area of 50 mm2.5 A recent US study demonstrated significantly enlarged nerve size index (average of median and ulnar nerve cross sectional areas corrected for height) in 31 of 36 patients with CIDP.14

Unlike CMT1A, the diagnosis of CIDP relies on the synthesis of multiple data as there is no single discriminatory diagnostic investigation. Modest sciatic nerve hypertrophy may be useful as supporting evidence for a diagnosis in CIDP, in addition to enlargement of nerve roots or plexuses, which is already included as supportive evidence in the current EFNS/PNS diagnostic criteria.11 This may be especially true in cases where the diagnosis is in doubt. Future studies in CIDP comparing images of both upper and lower limbs bilaterally may be of assistance in this regard.

We have demonstrated significant sciatic nerve hypertrophy in CMT1A patients compared with healthy controls by objective, quantitative measurements performed on MRI, providing additional information to complement and augment that obtained on conventional radiological evaluation. A more modest nerve enlargement was observed in CIDP patients. Sciatic nerve cross sectional area may add diagnostic value in investigations of acquired and inherited peripheral neuropathies such as CIDP and CMT.


We are grateful to Lisa Strycharczuk and Caroline Andrews for expert radiographic assistance.



  • Funding Funding was partly provided by the Medical Research Council (MRC) Centre for Neuromuscular Diseases, UCL Institute of Neurology. MAM was partially supported by AlBan, High Level Scholarship Programme of the European Union (2004-2006) and Caja de Seguro Social (CSS), Panama. MMR is grateful to the MRC and the Muscular Dystrophy Campaign for their support. This work was undertaken at UCLH/UCL who received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme.

  • Competing interests M G Hanna is Deputy Editor of JNNP but had no role in the editorial process for this paper.

  • Ethics approval This study was conducted with the approval of the The National Hospital for Neurology and Neurosurgery and the Institute of Neurology Joint Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.