Typical CIDP, distal variant CIDP, and anti-MAG antibody neuropathy: An ultra-high frequency ultrasound comparison of nerve structure

To date, little is known about the usefulness of ultra-high frequency ultrasound (UHF-US, 50–70 MHz) in clinical practice for the diagnosis of dysimmune neuropathies. We present a prospective study aimed at comparing UHF-US alterations of nerves and fascicles in chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), distal CIDP (d-CIDP) and anti-MAG neuropathy and their relationships with clinical and electrodiagnostic (EDX) features. 28 patients were included (twelve CIDP, 6 d-CIDP and 10 anti-MAG) and ten healthy controls. Each patient underwent neurological examination, EDX and UHF-US study of median and ulnar nerves bilaterally. UHF-US was reliable in differentiating immune neuropathies from controls when using mean and/or segmental nerve and/or fascicle cross-sectional area (CSA); furthermore, fascicle ratio (fascicle/nerve CSA) was a reliable factor for differentiating d-CIDP from other types of polyneuropathies. The fascicle CSA appears to be more increased in CIDP and its variant than in anti-MAG neuropathy. UHF-US offers information beyond simple nerve CSA and allows for a better characterization of the different forms of dysimmune neuropathies.


Electrodiagnostic studies
EDX evaluation (Dantec Keypoint, Natus Medical Incorporated, Orlando, USA) was performed on all subjects between 1 and 2 months prior to the US.Four motor (tibial, peroneal, median, ulnar) and four sensory (superficial peroneal, sural, median, ulnar) nerves were examined bilaterally.The normative reference values used in our laboratory are summarized in Table 1.For each patient we ensured that skin temperature was at least 32 °C.An EDX score was created taking into consideration several criteria: for distal motor latency (DML) prolongation a score of 1 was assigned for a prolongation of 30-50%, 2 for a prolongation of 50-70%, and 3 for ≥ 70% of the upper limit of normal (ULN); for motor conduction velocity (MCV) reduction a score of 1 was given for a reduction of = 30%, 2 for a reduction between 30 and 50%, and 3 for a reduction of > 50% of the lower limit of normal (LLN); for compound motor action potential (CMAP) amplitude decrease a score of 1 was given for a reduction of ≤ 50%, a score of 2 for a reduction between 50 and 70%, a score of 3 for a reduction of ≥ 70%, and a score of 4 if CMAP was unobtainable; for conduction block (defined as ≥ 50% reduction of proximal relative to distal negative peak CMAP amplitude, excluding the tibial nerve and Erb's point stimulation) a score of 1 was assigned; if EDX studies were considered normal a score of 0 was assigned (Table 1).

Ultrasound assessment
Ultrasound evaluations of the peripheral nerves were performed with two very high-resolution probes (UHF48 and UHF70 MHz, Vevo, VD, Visualtronics, Toronto, Canada).The highest frequency probe, UHF70, was used with the "General" preset, with a 29-71 MHz frequency band.When using UHF70 exploration depth was limited to 7.5 mm.The second probe, UHF48, was used with the "General" preset, with a 20-48 MHz frequency band.When using UHF48 exploration depth was limited to 14.5 mm.We used the UHF70 whenever the depth of the explored nerve was below 7.5 mm, usually at the wrist and elbow, The median nerve (MN) and ulnar nerve (UN) were examined from the wrist to the middle-arm.The probe was adjusted to be perpendicular to the nerves, no pressure was applied, and a neutral position was adopted for each limb except for the examination of the ulnar www.nature.com/scientificreports/nerve at the elbow, which was performed with the elbow flexed at 90°.Ultrasound examination of the nerve was performed bilaterally.The CSA for the MN was measured at the wrist, 10 cm from the distal wrist crease, at the antecubital fossa and mid-arm (Fig. 1).The CSA for the UN was measured at the wrist, 10 cm from the pisiform bone, at the elbow, 5 cm below and above the elbow and at mid-arm (Fig. 1).Assessment of nerve CSA (n-CSA) was performed on cross-sectional images using the manual tracking method by placing the cursor within the hyperechoic edge of the nerves.The depth function was not standardized but adjusted individually depending on the anatomy of the nerve at the examination site and the characteristics of the patient (amount of subcutaneous tissue) to allow visualization of the nerve in its entirety on cross-sections.The zoom function was not used.Fascicle identification and f-CSA measurement were performed on the same cross-sectional images used for n-CSA measurement.Fascicles were defined as hypoechoic areas surrounded by hyperechoic borders within the nerve and the CSA measurement was performed with the cursor placed within this hyperechoic margin 8 .
For each nerve, the largest f-CSA was measured and considered for statistical analysis.CSA reference values were obtained from the 10 healthy controls.Nerve and fascicle CSA were considered to be abnormal if 2 standard deviations (SD) greater than the upper limit.The fascicular ratio (FR) was calculated as f-CSA/n-CSA 10 .
To accommodate the potential variability in CSA distributions across different nerve segments and to enhance the diagnostic sensitivity of UHF-US in differentiating between dysimmune neuropathies, we also analyzed the largest nerve and fascicles CSA of the ulnar and median nerve for each pathology.

Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics for Windows, Version 26.0 (Armonk, NY: IBM Corp).Continuous variables with normal distribution are presented as mean ± standard deviation, while those with non-normal distribution are presented as median (interquartile range).Categorical variables are reported as the rate (percentage).For determining if there were differences between the four groups (controls, CIDP, d-CIDP, anti-MAG) concerning continuous data, one-way ANOVA tests and Kruskal-Wallis One-Way ANOVA on Ranks were conducted.Outliers were kept in the analysis, as they depicted true clinical scenarios.For determining the normality of data for each of the studied group, the Shapiro-Wilk test was performed.The homogeneity of variances was assessed by Levene's test of homogeneity of variances.For data for which the homogeneity of variances was violated, Welch ANOVA was used; in this case, post-hoc analysis was interpreted using the results of the Games-Howell post-hoc test.In cases in which the homogeneity of variances was met, post-hoc analysis was interpreted using the Tukey-Kramer post hoc test.When Kruskal-Wallis test was performed, post-hoc analysis was made using Dunn's (1964) procedure with a Bonferroni correction for multiple comparisons.For assessing the relationships between variables, Pearson correlation test was performed.In case of non-linearity, Spearman's rank-order correlation test was run.

Results
Thirty-eight participants were included in the current study: 12 with typical CIDP, 6 with distal variant CIDP (d-CIDP), and 10 with demyelinating neuropathy associated with anti-MAG antibodies (anti-MAG).Ten healthy controls were also included.Fifteen of those participating in the study (39.5%) were female, and 23 (60.5%) were male; there were more females in the control group than in the study group (70% vs. 28.6%).
The duration between the onset of neurological symptoms and first medical evaluation differed significantly between the three polyneuropathy groups, patients with anti-MAG antibodies having the longest duration of symptoms (19 months), followed by patients with CIDP (6.5 months), and by those with d-CIDP (3 months), χ2(2) = 8.942, p = 0.011.Post-hoc analysis revealed statistically significant differences in symptoms duration between the patients with d-CIDP and the ones with anti-MAG antibodies (p = 0.008) The demographic characteristics, clinical and electrophysiological scores are summarized in Table 2.
A Spearman's rank-order correlation was run to assess the relationship between EDX score and mean CSA of nerves and fascicles of ulnar and median nerves.There was a statistically significant, strong positive correlation between EDX scores and overall MN n-CSA r s (25) = 0.457, p = 0.017, and between EDX score and mean UN n-CSA r s (25) = 0.568, p = 0.002.There was no correlation between EDX score and f-CSAs.
When looking at the largest MN and UN CSA as measured by the UHF-US, significant between-group differences were seen for the studied nerves, suggesting substantial variations in nerve and fascicle enlargements in CIDP, d-CIDP and anti-MAG neuropathy compared to controls.Eta-squared values indicated a strong effect size for f-CSA (η 2 = 0.435), pointing to its potential as a sensitive biomarker for distinguishing neuropathic changes.(Table 3).
CIDP patients also had larger UN n-CSA, with a significant increase of the CSA of: 3.01 mm 3 , p = 0.033 above the elbow; 3.2 mm 3 (95% CI, 0.005-6.40),p = 0.049 below the elbow; and 4.24 mm 3 (95% CI, 0.24-8.24),p = 0.034 at the mid-arm level.UN f-CSA was also significantly larger in CIDP patients than in controls, with an increase of 0.62 mm 3 in the mid-forearm, p = 0.002, and of 1.16 mm 3 in the mid-arm site, p = 0.014.Overall mean UN n-CSA and mean UN f-CSA were significantly higher in CIDP patients than controls, with an increase of 1.78 mm 3 , p = 0.013, and respectively 0.69 mm 3 , p = 0.002.
These differences were consistent upon analysis of the largest median nerve (MN) n-CSA and f-CSA, with patients diagnosed with CIDP exhibiting a notably larger MN n-CSA (increase of 5.44 mm 3 , 95% CI (0.05, 10.84), p = 0.047) and f-CSA (increase of 4.19 mm 3 , 95% CI (1.71, 6.67), p < 0.001) compared to control subjects.Similar patterns were observed for the ulnar nerve (UN) n-CSA, where patients with CIDP demonstrated a significant increase of 2.17 mm 3 , p = 0.032 relative to controls.However, while the largest UN f-CSA was higher in the CIDP group compared to controls, this difference did not reach statistical significance (p = 0.182).www.nature.com/scientificreports/
When focusing on the largest CSA measurements of the studied nerves, it was observed that patients with anti-MAG neuropathy exhibited a trend toward larger MN n-CSA and f-CSA than controls, with an increase in MN n-CSA of 5.43 mm 3 ; however, this difference did not reach statistical significance (95% CI (−0.20, 11.06), p = 0.064).Similarly, the increase in MN f-CSA of 1.21 mm 3 (95% CI (−1.37, 3.81), p = 1.00) and the UN f-CSA of 0.66 mm 3 (p = 0.159) were larger in the anti-MAG group compared to controls but were not statistically significant.In contrast, the largest ulnar nerve (UN) n-CSA measurements proved to be a more robust differentiator, with patients with anti-MAG neuropathy displaying a statistically significant larger UN n-CSA (increase of 7.77 mm 3 , p = 0.001) compared to the control group.

D-CIDP vs. control
While no statistically significant differences were found for the MN and UN n-CSA, d-CIDP patients had significantly larger MN f-CSAs at the mid-arm site (increase of 3.44 mm 3 (95% CI, 0.739-6.15),p = 0.017) and UN f-CSAs below the elbow (increase of 3.05 mm 3 , p = 0.021).Furthermore, mean f-CSAs were significantly higher in d-CIDP patients, with an increase of 1.69 mm 3 , (95% CI, 0.26 to 3.12), p = 0.015 for the MN, and of 1.34 mm 3 , p < 0.001 for the UN respectively.Highest values of the CSAs of the studied nerves were also increased in patients with d-CIDP when compared with controls.Nonetheless, statistical significance was reached only in the case of the MN f-CSA (increase of 3.25 mm 3 , 95% CI (0.25, 6.25), p = 0.027) and UN f-CSA (increase of 3.47 mm 3 , p = 0.007) in d-CIDP patients when compared with controls.neuropathy, nerve enlargement is also seen in distal segments, without f-CSA increase.Analysis of the larger fascicular CSA allows one to clearly distinguish CIDP from anti-MAG neuropathy.To have an overall picture of these changes, the FR would be a suitable parameter, and based on our data a FR greater than 0.30 would be in favor of CIDP or its distal variant.This would suggest a different involvement of the different nerve structures by the specific inflammatory and immune mechanisms, which relates to the well-known clinical, EDX and treatment response differences of three different types of neuropathies 17 .Recent data has shown that the blood nerve barrier (BNB) is also subject to contrasting changes in the three neuropathies.The inflammatory process in CIDP and in d-CIDP leads to disruption of the tight junctions (TJ) 21 , while penetration of anti-MAG IgM antibodies through the BNB occurs without tight junction disruption and without increased permeability to small molecules (possibly through transcytosis) 22 .Since the TJs of endoneurial vessels are preserved, this would possibly explain the more reduced fascicular swelling.In d-CIDP, n-CSA values were overall lower than in the other neuropathies, while f-CSA were significantly increased, especially in proximal median nerve segments.Literature data regarding US in d-CIDP is scarce, due probably to two factors: firstly, most studies did not differentiate between subtypes or only included typical CIDP, and secondly a more definite definition was only recently made 1 .One report highlights diffuse n-CSA enlargement, especially in the distal sensory segments 23 .As far as our results are concerned, the fact that the n-CSA was not increased is not particularly surprising.This subgroup had a significantly shorter disease duration compared to the other groups and it has been shown that US changes are more evident the longer the patient's clinical history 20 .Also in our study, patients with a longer disease duration had a significantly larger fascicle area of the median nerve in the middle arm.On the other hand, these data seem to strengthen our hypothesis that in the case of CIDP and its distal variant, the inflammatory process electively affects the fascicles earlier in the disease and later the entire nerve.
In our study, no statistically significant association was found between clinical scores and the various groups of neuropathies.The absence of differences in clinical scores and different types of neuropathies was also noticed by Merola et al., while analyzing patients with CIDP and multifocal motor neuropathy 24 .However, an increase in mean MN et UN n-CSA was strongly correlated with an increase in EDX score, indicating that nerve enlargement is a marker of demyelination. 20,25.
The main limitations of our study are the small number of subjects in each subgroup and the different disease duration in each case.Furthermore, there is a predominance of women in the control group compared to the CIDP and anti-MAG group-however, we believe this is not a true limitation per se, as a statistically significant gender-related difference in MN and UN CSA has never been demonstrated 26 .
In conclusion, we believe our findings show the potential of UHF-US in differentiating between immune neuropathies, with several markers of internal nerve structure being of particular interest.Changes in fascicle size without or with increased n-CSA have already been described in the CIN 27 .The use of UHF probes allows better visualization of the nerve's internal structure and, based on our data, we could hypothesize that an increase in f-CSA with or without an increase in n-CSA would point the diagnosis towards CIDP or a variant thereof, whereas an increase in n-CSA even in the most distal parts of the nerve without an increase in f-CSA could be a marker of anti-MAG neuropathy.Further studies are needed to consolidate our findings and establish UHF-US as a useful biomarker to distinguish between immune neuropathies.

Figure 1 .
Figure 1.Median and Ulnar nerve in CIDP (A), in anti-MAG neuropathy (B), d-CIDP (C) and in healthy subject (D).In CIDP (A) thinning of perifascicular epineurium due to compression/dislocation from swollen fascicles.Reduced number of fascicles and swollen fascicles compared to healthy controls especially in more proximal nerve sites due to endoneuronal oedema.Focal/diffuse hypoechogenicity.In Anti-MAG neuropathy (B), reduced number of swollen fascicles with normal or slightly reduced perifascicular epineurium.Same or slightly reduced number of fascicles compared to healthy controls.Normal or slightly reduced echogenicity.In d-CIDP (C), normal number of swollen fascicles with normal or slightly reduced perifascicular epineurium.Normal or focal/diffuse hypoechogenicity.

Figure 2 .
Figure 2. Mean nerve CSA values (A: right; B: left) for the median and ulnar nerve in controls and neuropathy patients.

Figure 3 .
Figure 3. Mean fascicle CSA values (A: right; B: left) for the median and ulnar nerve in controls and neuropathy patients.

Table 1 .
EDX score.MN median nerve, UN ulnar nerve, TN tibial nerve, PN peroneal nerve, UL upper limbs, LL lower limbs, DML distal motor latency, MCV motor conduction velocity, cMAP compound motor action potential, CB conduction block, ms milliseconds, m/s meter/second, mV millivolts, NA not applicable, NR no response *values used in our laboratory.

Table 2 .
Clinical, demographical, and electrodiagnostic data.*Statistically significant difference M male, F female, INCAT Inflammatory Neuropathy Cause and Treatment Group (INCAT) disability scale, IQR interquartile range, NA not applicable, ONLS Overall Neuropathy Limitations Scale, SD Standard deviation.

Table 3 .
Differences between nerve and fascicle cross-sectional area (in mm 2 ) for median and ulnar nerve in controls and patients with immune neuropathies.Significance values in Bold.*Asterisks show statistical significance.