Magnetic resonance imaging may be used for early evaluation of diabetic peripheral polyneuropathy.

Abstract

Diabetic polyneuropathy (DPN) is conventionally diagnosed based on a neurological examination, nerve conduction studies (NCS), and quantitative sensory testing (QST). More recently, to detect lesions of peripheral nerves, novel methods have been introduced such as ultrasound of peripheral nerves1,2 and magnetic resonance neurography (MRN).3 Changes in the flow of low-protein endoneurial fluid in the epineurium4 along the nerve makes lesions appear hyperintense3 in MRN images. For that reason, MRN of microanatomical structures representing fascicular nerve lesions is feasible with the use of clinical high field scanners (3 Tesla) and T2-weighted imaging sequences with strong fat suppression.5 We have performed preliminary MRN studies demonstrating larger sciatic nerve volume and increased nerve signal intensity in diabetic patients with polyneuropathy as compared to nonneuropathic diabetic patients and healthy control subjects. We have observed that an image resolution of 0.6 × 0.6 mm2 is sufficient to determine volumetric enlargement, and that 0.3 × 0.3 mm2 is required (partial-volume-effects) to identify increased signal intensity in lesioned nerves. High-resolution fascicular intensity changes in a diabetic patient with neuropathy (B) and without neuropathy (C) are shown in Figure 1. MRN depicts variations in nerve lesion patterns along the nerve, which suggests, that DPN may not be “a dying back polyneuropathy” as it is believed to be. Figure 1. (A) Coronal image of the body indicating the scan location of the MR images at the midthigh. T2-weighted fat saturated cross sectional SPAIR MR Image (0.3 × 0.3 × 3 mm3) of (B) a diabetic type 2 patient with neuropathy and (C) a diabetic ... Resolution of 0.3 × 0.3 mm2 is sufficient to visualize the nerve fascicles, however, it is insufficient to visualize myelin sheaths (o: 2-20 µm) and the smaller structures of the axons (o: 0.2-3 µm).6 Visualization of such small structures requires much higher resolution using more dedicated surface coils (higher order of channels), and improved signal to noise ratio preferable to a higher field strength (eg, 7T). Larger studies are needed to determine the sensitivity and specificity of MRN applications in diabetic patients with varying degrees of neuropathy. Furthermore, application of diffusion-weighted sequences to visualize nerve pathology may increase the reliability, sensitivity and specificity of MRN. Our pilot study has limitations. First, the study has currently only included a low number of patients and our findings need to be confirmed in a much larger patient cohort. Second, thorough examination of peripheral nerve structures is needed with a less user dependent segmentation method than manual segmentation. Third, the fat saturated sequence rapidly lose signal from long echo times and an evaluation of echo times is needed to optimize the sequence. Finally, MR imaging analysis should include the tibial, sural, and peroneal nerves to provide additional information about lesions in nerves evaluated by NCS. MRN appears to detect early and more widespread nerve fascicle lesions that remain undetected by NCS. This suggests that MRN may become a useful investigation for diagnosing peripheral neuropathies. Hence, early detection of nerve lesions may enable prevention and thereby early treatment which may alleviate DPN and thereby the occurrence of end-stage complications such as diabetic foot ulcers and amputations.