Diffusion tensor imaging MR neurography for the detection of polyneuropathy in type 1 diabetes

Purpose To evaluate if diffusion tensor imaging MR neurography (DTI‐MRN) can detect lesions of peripheral nerves in patients with type 1 diabetes. Materials and Methods Eleven type 1 diabetic patients with polyneuropathy (DPN), 10 type 1 diabetic patients without polyneuropathy (nDPN), and 10 healthy controls (HC) were investigated with a 3T MRI scanner. Clinical examinations, nerve‐conduction studies, and vibratory‐perception thresholds determined the presence of DPN. DTI‐MRN (voxel size: 1.4 × 1.4 × 3 mm 3 ; b‐values: 0, 800 s/mm 2 ) covered proximal (sciatic nerve) and distal regions of the lower extremity (tibial nerve). Fractional anisotropy (FA) and apparent diffusion coefficient (ADC) were calculated and compared to T 2 ‐relaxometry and proton‐spin density obtained from a multiecho turbo spin echo (TSE) sequence. Furthermore, we evaluated DTI reproducibility, repeatability, and diagnostic accuracy. Results DTI‐MRN could accurately discriminate between DPN, nDPN, and HC. The proximal FA was lowest in DPN (DPN 0.37 ± 0.06; nDPN 0.47 ± 0.03; HC 0.49 ± 0.06; P < 0.01). In addition, distal FA was lowest in DPN (DPN 0.31 ± 0.05; nDPN 0.41 ± 0.07; HC 0.43 ± 0.08; P < 0.01). Likewise, proximal ADC was highest in DPN (DPN 1.69 ± 0.25 × 10 −3 mm 2 /s; nDPN 1.50 ± 0.06 × 10 −3 mm 2 /s; HC 1.42 ± 0.12 × 10 −3 mm 2 /s; P < 0.01) as was distal ADC (DPN 1.87 ± 0.45 × 10 −3 mm 2 /s; nDPN 1.59 ± 0.19 × 10 −3 mm 2 /s; HC 1.57 ± 0.26 × 10 −3 mm 2 /s; P = 0.09). The combined interclass‐correlation (ICC) coefficient of DTI reproducibility and repeatability was high in the sciatic nerve (ICC: FA = 0.86; ADC = 0.85) and the tibial nerve (ICC: FA = 0.78; ADC = 0.66). T 2 ‐relaxometry and proton‐spin‐density did not enable detection of neuropathy. Conclusion DTI‐MRN accurately detects DPN by lower nerve FA and higher ADC. These alterations are likely to reflect proximal and distal nerve fiber pathology. Level of Evidence: 1 J. Magn. Reson. Imaging 2017;45:1125–1134