Marieke van Rosmalen

General discussion 133 8 identification of the monoclonal gammopathy, and presence of a possible underlying hematologic disorder (e.g. antibodies to myelin associated glycoprotein, Waldenströms macroglobulinemia or amyloidosis). 42 These clinical and laboratory features will often guide a clinician to the right diagnosis without the use of nerve imaging techniques. Optimalization of MRI for these neuropathies may therefore be less relevant for clinical practice. Patients with axonal neuropathies normally do not have thickened nerve roots on nerve ultrasound or MRI. 31,32 The results of our study may therefore be representative to the group of axonal neuropathies as well. Another important result of our study is that combinations of measurements of nerve root size are probably more useful than a fixed cut-off. The inflammatory changes in the cervical nerve roots are not always homogeneous, but often have a patchy distribution. Variable cut-off values are therefore preferred over fixed cut-off values. We further found that 6 bilateral measurements close to the ganglion of root C5, C6 and C7 in coronal plane were easy to implement in routine practice (3 minutes per subject) and resulted in optimal test characteristics with high specificity levels (95%). We implemented these results in a practical risk chart that predicts the chance of having an inflammatory neuropathy ( Figure 4.3 ). Sensitivity levels of quantitative assessment of brachial plexus MRI were lower (approximately 27%) than those reported in qualitative studies (approximately 51%). 34,35 This may be explained by a degree of inclusion bias in earlier studies, as shown by another recent prospective cohort study that reported a relatively low sensitivity (36%) of qualitative brachial and lumbosacral plexus MRI in patients with suspected CIDP. 43 Validation of our results in a future prospective study would further establish that a quantitative assessment is preferred over a qualitative assessment of brachial plexus MRI. Our study in chapter 4 primary focused on thickness and did not evaluate T2 signal of the nerve roots. As earlier described, an increase of signal intensity of nerve roots on T2-weighted imaging is, combined with nerve root thickening, one of the abnormalities that can be found on (brachial) plexus MRI in CIDP and MMN. 25,44 Tissues appear bright on T2-weighted images when they have a long T2 relaxation time and therefore retain more signal, e.g. water. Hyperintense nerve roots indicate that there is an increase of the amount of water, for example due to oedema caused by an inflammatory reaction. Previous qualitative MRI studies of the brachial plexus reported that these T2 hyperintensities occur in 58-74% of patients with CIDP and 33-43% of patients with MMN. 30,45 Although T2 hyperintensity also occurs without the presence of nerve root thickening, most patients (70% of the patients with CIDP and 82% of the patients with MMN) have T2 hyperintensities accompanied by thickening of the cervical nerve roots. 45