Nerve Ultrasound Predicts Treatment Response in Chronic Inflammatory Demyelinating Polyradiculoneuropathy—a Prospective Follow-Up

As reliable biomarkers of disease activity are lacking, monitoring of therapeutic response in chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) remains a challenge. We sought to determine whether nerve ultrasound and electrophysiology scoring could close this gap. In CIDP patients (fulfilling EFNS/PNS criteria), we performed high-resolution nerve ultrasound to determine ultrasound pattern sum scores (UPSS) and predominant echotexture nerve conduction study scores (NCSS) as well as Medical Research Council sum scores (MRCSS) and inflammatory neuropathy cause and treatment disability scores (INCAT) at baseline and after 12 months of standard treatment. We retrospectively correlated ultrasound morphology with nerve histology when available. 72/80 CIDP patients featured multifocal nerve enlargement, and 35/80 were therapy-naïve. At baseline, clinical scores correlated with NCSS ( r 2 = 0.397 and r 2 = 0.443, p < 0.01), but not or hardly with UPSS (Medical Research Council sum scores MRCSS r 2 = 0.013, p = 0.332; inflammatory neuropathy cause and treatment disability scores INCAT r 2 = 0.053, p = 0.048). Longitudinal changes in clinical scores, however, correlated significantly with changes in both UPSS and NCSS ( r 2 = 0.272–0.414, p < 0.0001). Combining nerve/fascicle size with echointensity and histology at baseline, we noted 3 distinct classes: 1) hypoechoic enlargement, reflecting active inflammation and onion bulbs; 2) nerve enlargement with additional hyperechogenic fascicles/perifascicular tissue in > 50% of measured segments, possibly reflecting axonal degeneration; and 3) almost no enlargement, reflecting “burned-out” or “cured” disease without active inflammation. Clinical improvement after 12 months was best in patients with pattern 1 (up to 75% vs up to 43% in pattern 2/3, Fisher’s exact test p < 0.05). Nerve ultrasound has additional value not only for diagnosis, but also for classification of disease state and may predict treatment response.