Quantitative Foot Muscle Magnetic Resonance Imaging Reliably Measures Disease Progression in Children and Adolescents with Charcot–Marie–Tooth Disease Type 1A

Quantitative muscle fat fraction (FF) responsiveness is lower in younger Charcot–Marie–Tooth disease type 1A (CMT1A) patients with lower baseline calf‐level FF. We investigated the practicality, validity, and responsiveness of foot‐level FF in this cohort involving 22 CMT1A patients and 14 controls. The mean baseline foot‐level FF was 25.9 ± 20.3% in CMT1A patients, and the 365‐day FF (n = 15) increased by 2.0 ± 2.4% (p < 0.001 vs controls). Intrinsic foot‐level FF demonstrated large responsiveness (12‐month standardized response mean (SRM) of 0.86) and correlated with the CMT examination score (ρ = 0.58, P = 0.01). Intrinsic foot‐level FF has the potential to be used as a biomarker in future clinical trials involving younger CMT1A patients. ANN NEUROL 2024;96:170–174

neuropathy affecting around 1 in 2,500 people, with length-dependent weakness, sensory loss, and muscular atrophy. 1 CMT1A is the commonest subtype, accounting for >50% of cases, and is due to a duplication of the PMP22 gene. 2,3While any potential treatment would likely first be tested in adults, children and young people have the most opportunity to benefit from treatments that slow the accumulation of disability, so suitable outcome measures are essential for this group of patients. 4s the possibility of genetic therapies for CMT advances, 5 quantification of intramuscular FF within calf muscles using lower limb magnetic resonance imaging (MRI) has been shown to be a reliable and responsive outcome measure of disease progression in adults with CMT1A, [6][7][8] CMTX1, CMT1B, CMT2A, 9 and hereditary sensory neuropathy type 1 (HSN1). 10However, subgroup analysis of adults with CMT1A, demonstrated lower responsiveness, as measured by the standardized response mean (SRM) in From the 1  patients who were younger, with baseline calf fat fraction <10% and with lower CMT examination scores (CMTES). 7ualitative and semiquantitative MRI data are available for this younger CMT1A cohort.One study undertook a cross-sectional analysis involving a 1.5-Tesla (T) MRI analysis of 55 children with CMT, including 44 with CMT1A, and showed correlation between volumetric analysis of the anterior calf on T1 weighted (w) images and a number of clinical measures. 11Another case series of lower limb muscle MRI in CMT1A included 3 children/adolescents aged 8-17 years, all of whom had normal calf muscle MRI, but abnormal intrinsic foot muscles on T1w images. 12However, neither quantitative nor longitudinal MRI of foot muscles has previously been reported in children with CMT1A.
We therefore assessed the practicality, validity, and responsiveness of intramuscular fat measurements of intrinsic foot muscles using quantitative MRI in children and adolescents with CMT1A.

Ethics Approval
The present study involves human participants, and has been approved by the local ethics committee and institutional review boards at the participating sites in London, UK, and Iowa, USA (London-Queen Square Research Ethics Committee, REC reference: 18/LO/1059, and HawkIRB, IRB ID: 201802722).Written informed consent was obtained from each participant.

Study Design and Participants
A total of 22 participants with CMT1A (mean age 13.1 years, range 6-20 years, 10 male) were recruited from the inherited neuropathy cohorts at the Queen Square Center for Neuromuscular Diseases at Great Ormond Street Hospital, London, UK, and at the University of Iowa Carver College of Medicine, IA, USA.Inclusion criteria were: age 5-20 years with genetically proven CMT1A or a clinical diagnosis of CMT1A, including neurophysiological and genetic confirmation of CMT1A in a first-degree relative.Exclusion criteria were safetyrelated contraindications to MRI, planned foot surgery, pregnancy, or concomitant neuromuscular disease.A total of 14 age-matched controls (mean age 14.5 years, range 6-20 years, 5 male) were recruited from friends and family of staff members and participants.Those at risk of CMT were excluded.

Procedures
Baseline and follow-up clinical assessments included Medical Research Council (MRC) grade, overall neuropathy limitations scale (ONLS), CMTES version 2 (CMTESv2), CMTES version 2 Rasch (CMTESv2-R), CMT Pediatric Scale (CMTPedS), and CMT Health Index (CMT-HI). 10,13Quantitative MRI of both calves was acquired, as previously reported. 6-10MRI of the left foot was acquired using a dedicated foot coil (Siemens Prisma, London, UK; A trained observer (C.M.D.) used ITK-SNAP software to manually select and segment the forefoot muscles in the coronal plane (Fig). 14This was performed on a single slice in the forefoot at the midpoint of the first metatarsal (Fig) .Manual segmentation was completed with the observer blinded to participant and scan order, and quantitative MRI parameters were extracted using our existing analysis pipeline and analyzed at baseline and follow-up.Calf-level 3-point Dixon MRI was performed centered at a fixed distance from the knee joint.A trained observer (L.F.O.) selected the slice for analysis as being the closest to a fixed distance from the lateral tibial plateau: 13 cm distal for the calf.Whole muscle segmentation was performed using Musclesense, a trained artificial neural network for lower limb segmentation, as previously published. 8,15atistical Analysis IBM-SPSS V.27.0 (IBM Corporation, Armonk, NY, USA) was used.Correlations were assessed using Spearman's correlation.Comparisons between patients and controls were made using students' two-tailed t tests.Longitudinal changes were assessed with paired samples t tests.

Results
The scans were well tolerated, with no children in the study unable to complete the 30-min protocol.A single baseline foot-level acquisition from a control was excluded due to failed generation of the FF map, and a calf-level FF of 14% was excluded due to associated bilateral Achilles tendon lengthening surgery.

Discussion
We have demonstrated the practicality, validity, and responsiveness of MRI quantification of intramuscular FF of foot muscles in children and adolescents with CMT1A.The protocol was successfully implemented across three sites and was well tolerated.Foot-level FF was significantly elevated in  children with CMT1A compared with healthy controls (25.9% vs 3.5%, p < 0.001) and correlated with age (ρ 0.59, p = 0.006).This is consistent with previous qualitative T1w MRI abnormalities in foot muscles in 3 children with CMT1A. 12By comparison, calf-level FF was only slightly elevated in CMT1A patients versus controls (2.7% vs 1.3%, p < 0.001), consistent with similarly low values reported in younger CMT1A patients previously. 7he validity of foot-level MRI as an outcome measure is evidenced by strong correlations with a number of validated outcome measures, including CMTES, CMTES version 2 Rasch, and 9-hole peg test.The strength of these correlations is similar to that previously seen for calf-level FF in adults with CMT1A, 6,7 although in children, it is present in foot muscles, where the range of foot muscle FF observed reflects the spectrum of disease severity in this age group.For the same reason, the responsiveness of foot-level FF over 12 months in children and adolescents in the present study was large (SRM 0.86), very similar to the large responsiveness of calf-level FF in adults (SRM 0.83). 6Conversely, calf-level FF did not progress over 12 months, and did not correlate with clinical measures in children with CMT1A.We did not observe a correlation with foot-level FF and the CMTPedS.This is likely explained by the CMTPedS containing items that do not relate to lower limb function, such as grip strength and upper limb functional dexterity. 13Similarly, items reflecting lower limb sensory function, such as pinprick and vibration sense, have weaker correlations with lower limb muscle MRI FF.Hence, the overall CMTPedS score, which includes these items, likely reduces the overall correlation observed.We have shown this in adult CMT1A patients, where calf-level FF correlated more strongly with the motor component of the CMTES than the overall CMTES (ρ = 0Á77, p < 0Á0001 vs ρ = 0Á63, p = 0Á003, respectively). 6 limitation of the present study was the exclusion of patients who were planning foot surgery during the course of the study.This was done to ensure there were no safety-related contraindications to MRI, and also to ensure that any postsurgical artefact did not interfere with the integrity of the MRI measurements.While necessary, this does limit the generalizability of our findings to all patients with CMT1A, especially those who have undergone or are considering foot surgery as part of their treatment plan.
The present study, combined with data from adults already published, completes a very clear pattern of disease progression on lower limb muscle MRI in CMT1A.This pattern is expected from the recognized length dependent clinical phenotype in CMT1A.In children and young adults, the widest spectrum of fat accumulation is seen in foot muscles, which provide the most responsive outcome measure for this group.In adults, with calf muscle FF >10%, calf-level FF is highly responsive (SRM 2.2) as an outcome measure. 7Finally, in those with end-stage calflevel FF of >70%, thigh-level FF is abnormal and responsive, as we recently demonstrated in other forms of CMT. 9 Therefore, if using quantitative muscle MRI as an outcome measure for disease progression in CMT1A, we recommend either inclusion criteria that ensures the same anatomical level that can be used for analysis in all participants, or utilizing a severity-appropriate anatomical level according to the baseline FF assessment.  of Rare Diseases (U54NS065712), and the Muscular Dystrophy Association (MDA510281) for their support.

Figure :
Figure: Examples of fat fraction (FF) acquisition at foot-and calf-level for patients and controls.Gradient echo sequence (performed as part of 3-point Dixon magnetic resonance imaging acquisition, echo time = 3.45 ms) showing slice selection at a single slice in the forefoot at the midpoint of the first metatarsal with regions of interest (represented in green) in the (A: upper left) axial plane, (A: upper right) sagittal plane, and (A: lower left) coronal plane.(A: lower right) When regions of interest were applied to the generated FF map in this Charcot-Marie-Tooth disease type 1A patient, a moderate-to-severe foot-level FF of 47.8% was acquired.This process was repeated in (B) another Charcot-Marie-Tooth disease type 1A patient, generating a mild-to-moderate foot-level FF of 20.2%, and (C) in a control patient, generating a normal foot-level FF of 1.9%.(D-F) Calf magnetic resonance imaging sequences acquired for the respective patients as detailed above.(D: upper) Gradient echo sequence (performed as part of 3-point Dixon magnetic resonance imaging acquisition, echo time = 3.45 ms) showing slice selection at a single slice at mid-calf level with regions of interest (represented in green) in the axial plane.(D: lower) The Musclesense-derived segmentation post-quality control checks were applied to the generated FF map, yielding an overall (left and right) calf-level mean FF of 2.3%.This process was repeated in (E) another Charcot-Marie-Tooth disease type 1A patient, generating a calf-level FF of 1.2%, and (F) in a control patient, generating a minimal calf-level FF of 0.5%.[Color figure can be viewed at www.annalsofneurology.org]
Center for Neuromuscular Diseases, Department of Neuromuscular Diseases, UCL Queen Square Institute of Neurology, London, UK; 2 Roy and Lucille Carver College of Medicine, University of Iowa, Iowa City, IA, USA; 3 Fondazione Serena Onlus, Centro Clinico NeMO Trento, Italy; 4 Lysholm Department of Radiology, National Hospital for Neurology and Neurosurgery, London, UK; 5 Dubowitz Neuromuscular Center, Great Ormond Street Hospital, London, UK;

Table 1 .
Magnetic Resonance Imaging Data Results are Given as mean AE standard deviation (SD).The p-values are independent samples 2-tailed t test comparing Charcot-Marie-Tooth disease type 1A patients and controls.NS = non-significant.

Table 2 .
Spearman's Correlation Coefficient Between Baseline Foot-Level Fat Fraction and Outcome Measures Performed in the Study for Charcot-Marie-Tooth Disease type 1A Patients