Relationship Between Quantitative Magnetic Resonance Imaging Measures and Functional Changes in Patients With Duchenne Muscular Dystrophy

ABSTRACT

Introduction/Aims

Improved methodologies to monitor the progression of Duchenne muscular dystrophy (DMD) are needed, especially in the context of clinical trials. We report changes in muscle magnetic resonance imaging (MRI) parameters in participants with DMD, including changes in lean muscle volume (LMV), muscle fat fraction (MFF), and muscle fat infiltration (MFI) and their relationship to changes in functional parameters.

Methods

MRI data were obtained as part of a clinical study (NCT02310763) of domagrozumab, an antibody‐targeting myostatin that negatively regulates skeletal muscle mass. This post hoc analysis evaluated participants with Dixon MRI data and corresponding functional data at baseline and weeks 49 and 97. Images were analyzed to evaluate changes in adductors, hamstrings, and quadriceps.

Results

There was a positive correlation between increases in LMV and function. LMV changes in adductors (R = 0.51) and quadriceps (R = 0.54) showed a stronger correlation with function than LMV changes in hamstrings (R = 0.30). There was a negative correlation between MFF and MFI, respectively, and function in adductors (R = −0.57, R = −0.42), quadriceps (R = −0.59, R = −0.50), and hamstrings (R = −0.53, R = −048). Participants with preserved North Star Ambulatory Assessment scores had high total LMV (LMV_tot) and low total MFI (MFI_tot). Low ratios of LMV_tot to MFI_tot, or participants with small LMV_tot and high MFI_tot, appeared to have a rapid decline in function and loss of ambulation.

Discussion

These findings support the use of MRI biomarkers as potential surrogate endpoints in clinical trials of patients with DMD.

Trial Registration: ClinicalTrials.gov identifiers: NCT02310763

Abbreviations

CFB

change from baseline

DMD

Duchenne muscular dystrophy

LMV

lean muscle volume

LMV_tot

total lean muscle volume

MFF

muscle fat fraction

MFF_tot

total muscle fat fraction

MFI

muscle fat infiltration

MFI_tot

total muscle fat infiltration

MRI

magnetic resonance imaging

NSAA

North Star Ambulatory Assessment

P.P.

percentage points

S.D.

standard deviation

1. Introduction

Progression of Duchenne muscular dystrophy (DMD) is typically evaluated using functional tests, such as changes in North Star Ambulatory Assessment (NSAA) scores and other timed function tests [1, 2]. While defining clinically meaningful changes using metrics of functional performance is appealing from an interpretability perspective, detecting changes in function often requires large clinical trials with long monitoring periods. Magnetic resonance imaging (MRI) has been proposed as a biomarker of progression of DMD [3]. In addition, quantitative MRI measures of skeletal muscles can inform future functional changes and may be a useful biomarker for future clinical trials [4, 5, 6].

Previous studies have investigated various skeletal muscle MRI parameters as potential biomarkers of future functional changes [3, 7, 8, 9, 10]. However, additional studies are needed to understand the optimal segmentation of images during evaluation (i.e., region of interest in a single muscle, muscle group, or the whole thigh). In addition, when analyzing whole muscle groups, the changes in quantitative MRI parameters over time in participants with DMD are unclear. Most studies to date have focused on longitudinal changes in fat fraction in a central section of individual muscles [7, 8, 11, 12, 13, 14]. Results have been promising; however, changes to the contractile muscle tissue volume measured over the whole muscle, as well as diffuse fat infiltration in the still viable regions of the muscle, of individual muscle groups and across the whole thigh have not been extensively explored. Moreover, these biomarkers have recently been applied in large‐scale clinical trials in facioscapulohumeral muscular dystrophy, using standardized and automated image analysis based on fat‐referenced MRI, but their applicability in DMD is yet to be explored and validated [15, 16]. To develop sensitive, image‐based biomarkers for clinical trial endpoints, it is important to understand the expected changes in MRI measures over time, as well as the relationship between changes in quantitative MRI biomarkers and function [7, 13, 14].

This report presents the results of a post hoc analysis of MRI scans that were obtained as part of a Phase 2 clinical study (NCT02310763) of domagrozumab [17]. Domagrozumab is a humanized monoclonal antibody designed to inhibit a negative regulator of skeletal muscle growth, myostatin. The development of domagrozumab was discontinued after it did not demonstrate a significant treatment effect in patients with DMD. This analysis examined the changes in quantitative MRI measurements of LMV (lean tissue volume of the entire muscle), MFF (fat fraction of the entire muscle), and MFI (fat fraction of viable muscle tissue [voxels with < 50% fat]) across thigh muscle groups. We presented the relationship between quantitative MRI parameters and functional performance over time. We also examined the potential for combining MRI biomarkers to understand the potential risk for loss of ambulation. Finally, we also investigated the progression of MRI biomarkers at various stages of disease progression.

2. Methods

2.1. Study Design

MRI scans used in this analysis were obtained as part of a randomized, two‐period, double‐blind, placebo‐controlled study in ambulatory participants diagnosed with DMD. Participants were aged 6 to < 16 years at the screening visit. Participants completed the 4‐stair climb in ≥ 2.5 to < 12 s at screening and received a stable dose of glucocorticoid steroids for 6 months, with a stable regimen for ≥ 3 months prior to guardians signing informed consent. Participants received multiple ascending doses of intravenous (IV) domagrozumab (5, 20, and 40 mg/kg) or placebo.

The primary objectives of the original study were to determine the safety and tolerability of domagrozumab and to evaluate the observed mean change from baseline (CFB) on function (4‐stair climb) compared with placebo following 49 weeks of treatment. The CFB in NSAA was a secondary endpoint in the study. Detailed information regarding study design, population demographics, and baseline disease states has been published elsewhere [17].

This post hoc analysis evaluated all available participants with Dixon MRI data and corresponding functional data at baseline, Week 49, and Week 97. This analysis was performed on all patients combined.

2.2. Image Analysis

All images were collected using a standardized imaging protocol. All staff were trained on the specific requirements of this study to ensure consistent image acquisition across sites. The complete imaging protocol, training procedures, and quality inspection steps during the clinical trial are described elsewhere [3, 6].

Dixon images from a single thigh covering the hip to the knee were acquired at baseline, Week 17, Week 33, Week 49, and Week 97. Images were analyzed using the cloud‐based service AMRA Researcher (AMRA Medical AB, Linköping, Sweden) to evaluate adductors, hamstrings, and quadriceps [15]. The image analysis consisted of (1) automatic calibration using fat‐referenced MRI, which converted the fat images into quantitative fat concentration maps; (2) automatic segmentation using a multiatlas segmentation; (3) dual‐operator manual quality control of image quality and segmentations; and (4) automatic calculation of muscle measurements. The quantitative measurements in each of the muscle groups were LMV, MFF, and MFI. Examples of quadriceps segmentation and an illustration of the biomarkers are shown in Figure S1. Measurements of multiple muscles were combined into whole thigh composite scores of total LMV (LMV_tot), total MFF (MFF_tot), and total MFI (MFI_tot) [18].

• ○
  ${\mathrm{LMV}}_{\mathrm{tot}}={\sum }_{i\in \text{Muscles}}{\mathrm{LMV}}_i$
• ○
  ${\mathrm{MFF}}_{\mathrm{tot}}=1-\frac{{\sum }_{i\in \text{Muscles}}{\mathrm{LMV}}_i}{{\sum }_{i\in \text{Muscles}}{\mathrm{TMV}}_{\mathrm{i}}}=1-\frac{{\sum }_{i\in \text{Muscles}}{\mathrm{LMV}}_i}{{\sum }_{i\in \text{Muscles}}\left({\mathrm{LMV}}_i/\left(1-{\mathrm{MFF}}_i\right)\right)}$
• ○
  ${\mathrm{MFI}}_{\mathrm{tot}}=\frac{{\sum }_{i\in \mathit{\text{Muscles}}}{\mathrm{MFI}}_i\cdot {\mathrm{LMV}}_i}{{\sum }_{i\in \text{Muscles}}{\mathrm{LMV}}_i}$

Composite scores were calculated across the three thigh muscles. For longitudinal changes, a muscle was included in the composite score only if it was measurable across all available timepoints.

2.3. Functional Measurements

All functional assessments were based on NSAA as this assessment is commonly used to measure functional performance in clinical trials. NSAA is a 17‐item test that measures gross motor function [2]. Each individual item receives a score of 0 = unable to perform independently, 1 = able to perform with assistance, or 2 = able to perform without assistance. The total score ranges from 0, complete loss of function, to 34, which indicates fully independent function [2]. Functional measurements were collected at baseline, Week 49, and Week 97. The 10‐m walk/run timed functional assessment was also evaluated as part of the NSAA evaluation. Loss of ambulation was captured as an adverse event by the principal investigator and retrospectively coded as “gait inability”.

2.4. Statistical Analysis

The relationship between the CFB in MRI parameters at Week 49 and the CFB in NSAA total scores at Week 97 was assessed. The relationship between baseline MFF (%) and percent CFB (%CFB) for LMV and CFB (percentage points) for MFF and MFI over 49 weeks was explored.

LMV was evaluated as %CFB to account for variability in body size and age, whereas MFF and MFI were evaluated as CFB in percentage points. For continuous variables, data were summarized using descriptive statistics. For categorical or ordinal variables, the number of participants (%) was summarized. The relationship between the MRI endpoints and function was examined by regression analyses, with Spearman's correlation coefficients. Strong or very strong correlations were defined as those with coefficients of 0.70–1.00, moderate correlations of 0.40–0.69, and weak correlations of 0.10–0.39 [19].

The 10‐m run/walk timed functional test was evaluated as a velocity to account for participants who could not complete this assessment and whose scores may bias the overall results.

3. Results

3.1. Relationship Between MRI Parameters and NSAA Total Score at Baseline

At baseline, there was a moderate positive correlation between LMV and NSAA total scores in adductors and quadriceps while the correlation in hamstrings was weak. LMV_tot and NSAA total scores had a moderate positive correlation. The negative correlation between MFF or MFF_tot and NSAA total scores was moderate in adductors, hamstrings, and quadriceps. The MFI_tot and NSAA total scores also had a moderate negative correlation (Figure 1).

FIGURE 1.

Relationships between MRI parameters and functional assessments at baseline. The correlations between NSAA total score and MRI parameters at baseline. Abbreviations: ALMV, adductor lean muscle volume; AMFF, adductor muscle far fraction; AMFI, adductor muscle fat infiltration; HLMV, hamstring lean muscle volume; HMFF, hamstring muscle fat fraction; HMFI, hamstring muscle fat infiltration; LMV_tot, total lean muscle volume; MFF_tot, total muscle fat fraction; MFI, quadriceps muscle fat infiltration; MFI_tot, total mean muscle fat infiltration; MRI, magnetic resonance imaging; NSAA, North Star Ambulatory Assessment; QR, correlation coefficient.

3.2. Changes in MRI Parameters Over 49 Weeks

Ninety‐eight participants were included for progression analysis, of whom 78 had Dixon MRI data recorded at baseline and 67 had longitudinal assessments between baseline and Week 49. LMV measured through %CFB (mean ± standard deviation [SD]) decreased in the adductors and quadriceps at Week 49. There was a small increase in the relative LMV in hamstrings (2.78% ± 13.95%) over the same period. The CFB in MFF was similar across all muscle groups at Week 49 9. The CFB in MFI was also similar across muscle groups, with an increase in relative MFI in adductors, quadriceps, and hamstrings at Week 49. Composite LMV_tot values were stable from baseline to Week 49, while there were increases in MFF_tot and MFI_tot in the same period (Table 1).

TABLE 1.

Summary statistics of MRI parameters of interest over the 49‐week monitoring period.

Muscle region	MRI variable	Baseline mean ± SD n = 78 a	At Week 49 mean ± SD n = 75 b	CFB at Week 49 c mean ± SD n = 67 b
Adductors	LMV, L	0.13 ± 0.05	0.12 ± 0.06	−6.76 ± 10.56
	MFF, %	39.42 ± 20.05	44.40 ± 20.90	6.85 ± 5.89
	MFI, %	20.87 ± 6.71	21.63 ± 6.76	1.31 ± 2.41
Hamstrings	LMV, L	0.22 ± 0.06	0.23 ± 0.06	2.78 ± 13.95
	MFF, %	32.36 ± 15.74	36.89 ± 16.77	6.68 ± 5.74
	MFI, %	18.49 ± 6.29	19.80 ± 6.24	1.91 ± 2.09
Quadriceps	LMV, L	0.36 ± 0.14	0.36 ± 0.15	−4.1 ± 12.31
	MFF, %	31.69 ± 18.06	36.15 ± 18.79	7.20 ± 5.48
	MFI, %	19.50 ± 8.33	21.55 ± 8.58	3.32 ± 3.07
Composite of all muscles	LMVtot MFFtot MFItot	0.71 ± 0.22 c 33.21 ± 17.27 c 19.10 ± 6.79 c	0.72 ± 0.25 d 37.77 ± 17.96 d 20.70 ± 6.90 d	−0.01 ± 0.08 e 6.97 ± 5.44 e 2.46 ± 2.34 e

Note: MRI parameters were analyzed as absolute values and CFB.

Abbreviations: CFB, change from baseline; LMV, lean muscle volume; LMV_tot, composite LMV of all muscles, MFF, muscle fat fraction; MFF_tot, composite MFF of all muscles; MFI, muscle fat infiltration; MFI_tot, composite MFI of all muscles; MRI, magnetic resonance imaging; SD, standard deviation.

^a n = 78 for adductors and quadriceps, n = 77 for hamstrings.

^b One participant was omitted in the analysis of hamstrings and quadriceps.

^c n = 77, MFF and MFI are shown as the change in percentage points from baseline, and LMV is the percentage change from baseline.

^d n = 74.

^e n = 66.

3.3. Relationship Between Changes in MRI Parameters and Functional Changes

Due to the early termination of the parent study, functional data at Week 97 for the correlation analysis were only recorded for 33 participants [3, 17]. Correlation coefficients of the CFB in MRI parameters at Week 49 with the CFB in functional performance at Week 49 and Week 97 are shown in Figure 2. A positive relationship between LMV and function was observed, and the increases in LMV correlated with improved function. LMV changes in adductors, quadriceps, and LMV_tot showed a stronger correlation with function than LMV changes in hamstrings.

FIGURE 2.

Correlation between MRI changes and functional changes. The correlation coefficient is displayed with positive correlations shown in red and negative correlations shown in blue. Imaging Week 49 CFB versus Function CFB Week 49 n = 64 and Week 97 n = 33. One participant was omitted in the analysis of hamstrings and quadriceps as the muscles were not analyzable due to image quality. Abbreviations: CFB, change from baseline; LMV, lean muscle volume; MFF, muscle fat fraction; MFI, muscle fat infiltration; NS, not significant; NSAA, North Star Ambulatory Assessment; quads, quadriceps; R, correlation coefficient; tot, total; Wk, Week.

The negative correlation between changes in MFF and function was similar across adductors, hamstrings, and quadriceps, as well as MFF_tot. The negative correlation between changes in MFI and function was similar across all muscle groups, with the strongest correlations observed for the quadriceps and MFI_tot regions (Figure 2).

MRI parameters in all regions generally had a stronger correlation with the CFB in function at Week 97 than at Week 49, except for LMV changes in hamstrings.

3.4. Relationship Between Muscle MRI Parameters, Age, and NSAA Total Scores

There was a clear progression with participants' age in MFF and MFI for all muscle regions. In general, high NSAA scores were observed when MFF and MFI were low and when LMV was high. The LMV in adductors decreased over time across all NSAA scores (Figure 3A). There was an increase over time in LMV in the hamstrings of participants with high NSAA scores (Figure 3B). Of note, there was a steep increase in MFI in the quadriceps between ages 8 and 11 years, which was not as prominent in other muscle groups (Figure 3G). In general, LMV showed a less predictable trajectory across ages compared with MFI and MFF.

FIGURE 3.

Muscle MRI variables by age. (A–D) LMV; (E–H) MFI; (I–L) MFF. The NSAA score was highlighted in a blue scale with dark blue reflecting high scores and light blue reflecting low scores. N = 98. Abbreviations: ALMV, adductor lean muscle volume; AMFF, adductor muscle fat fraction, AMFI, adductor muscle fat infiltration; HLMV, hamstring lean muscle volume; HMFF, hamstring muscle fat fraction; HMFI, hamstring muscle fat infiltration; LMV_tot, total lean muscle volume; MFF_tot, total muscle fat fraction; MFI_tot, total muscle fat infiltration; MRI, magnetic resonance imaging; NSAA, North Star Ambulatory Assessment, QLMV, quadriceps lean muscle volume; QMFF, quadriceps muscle fat fraction; QMFI, quadriceps muscle fat infiltration.

3.5. Relationship Between LMV, MFF, and MFI

Overall, participants with low LMV had high MFF and MFI and low NSAA total scores. Participants with high LMV and low MFF and MFI had high NSAA total scores. In general, participants with low MFF or low MFI had higher LMV measures (Figure 4A–H). There was a strong correlation between MFF and MFI, with some divergence in the adductors compared with other muscle groups (Figure 4I–L).

FIGURE 4.

The relationship between the LMV, MFF, and MFI. (A–D): LMV to MFF; (E–H): LMV to MFI; (I–L): MFF to MFI.NSAA total scores were highlighted in a blue scale with dark blue reflecting high scores and light blue reflecting low scores. N = 98. Abbreviations: ALMV, adductor lean muscle volume; AMFF, adductor muscle fat fraction, AMFI, adductor muscle fat infiltration; HLMV, hamstring lean muscle volume; HMFF, hamstring muscle fat fraction; HMFI, hamstring muscle fat infiltration; LMV_tot, total lean muscle volume; MFF_tot, total muscle fat fraction; MFI_tot, total muscle fat infiltration; MRI, magnetic resonance imaging; NSAA, North Star Ambulatory Assessment, QLMV, quadriceps lean muscle volume; QMFF, quadriceps muscle fat fraction; QMFI, quadriceps muscle fat infiltration.

3.6. Relationship Between Composite Muscle Parameters, NSAA Total Score, and Ambulatory Status

In general, participants with preserved NSAA scores had high LMV_tot and low MFI_tot. In addition to the assessment of LMV_tot and MFI_tot as independent variables, the ratio of LMV_tot to MFI_tot was examined as an additional method to determine muscle quality. Low ratios of LMV_tot to MFI_tot, or participants with small LMV_tot and high MFI_tot, may relate to a rapid decline in function and loss of ambulation (Figure 5).

FIGURE 5.

The relationship between the composite MRI parameters and ambulatory status by age. Data relating NSAA total scores to muscle MRI measures in total thigh has been separated based on the participants' age group. The age groups are 5–7 years, 7–9 years, 11–13 years, 13–15 years, and > 15 years. Ambulatory status is indicated with dots for ambulatory (AMB) and crosses for non‐ambulatory (Non‐Amb). Abbreviations: LMV_tot, total lean muscle volume; MFF_tot, total muscle fat fraction; MFI_tot, total muscle fat infiltration; MRI, magnetic resonance imaging; NSAA = North Star Ambulatory Assessment.

4. Relationship Between Baseline MFF and the CFB in LMV, MFF, and MFI

The CFB MFF versus baseline MFF peaked between 30% and 60% baseline MFF with a bell‐shaped distribution. The CFB in MFI was the highest at 10%–30% baseline MFF and lowest at 60%–100% baseline MFF. In muscles with MFF lower than 20%, the CFB in LMV was positive with a peak of 17% in muscles with MFF < 10%, whereas it was negative for muscles with higher MFF, with a peak of −14% at 60%–70% MFF (Figure 6).

FIGURE 6.

The relationship between baseline MFF and CFB in LMV, MFF, and MFI. A histogram of the CFB in p.p. in LMV, MFF, and MFI in relation to baseline MFF is shown. Error bars indicate standard error. Abbreviations: CFB, change from baseline; LMV, lean muscle volume; MFF, muscle fat fraction; MFI, muscle fat infiltration; p.p., percentage point.

5. Discussion

Consistent with previous studies, there was a correlation between changes in skeletal muscle MRI parameters and functional changes in participants with DMD [3, 6, 20, 21, 22]. We built upon these studies to evaluate the relationship between muscle group measures of volume or quality versus functional changes. To establish muscle MRI as a viable pharmacodynamic response biomarker in clinical trials, it is important that image analysis is feasible at a large scale and that muscle MRI measures can inform future functional changes.

We demonstrated that correlations between MRI parameters at Week 49 and functional changes at Week 97 were stronger than comparisons at Week 49. This suggests that MRI parameters of LMV, MFF, and MFI, either across the entire thigh or within a muscle group, may be informative of future functional changes. Together with previous literature reports of MFF, this supports the use of MRI biomarkers as potential endpoints in clinical trials of patients with DMD [3, 8, 9].

We used a recently introduced combination of three MRI biomarkers for assessing disease progression in DMD, namely, MFF, MFI, and LMV. While MFF measures all the fat in the muscle, MFI measures only the fat in the viable muscle tissue. As such, MFI assesses the quality of the viable muscle tissue and complements MFF with a more detailed description of the muscle fat distribution. LMV measures the amount of contractile tissue in the muscles and is altered in response to changes in both muscle size and fat replacement. Since it is directly linked to both these processes, it may be the most useful biomarker for certain clinical questions. Measuring LMV has the potential to assess and correct for natural muscle growth.

The MFF trajectories presented here are consistent with the sigmoid shape observed across similar age ranges in natural history studies [8, 9]. Participants with the lowest MFF and MFI tended to have the highest NSAA total scores, with function decreasing as MFF and MFI increased.

LMV was also linked to NSAA scores and functional testing, with larger muscle volumes correlating with higher scores. The association between LMV and MFF/MFI was weak, indicating that they measure different properties of the muscles. Whole thigh or whole muscle LMV has not been extensively characterized in the DMD population [10]. Although some studies have presented muscle cross‐sectional areas, few have demonstrated full assessments of upper‐leg muscle volumes [23, 24].

We stratified changes in MFF, MFI, and LMV by baseline MFF. Baseline MFF was used as a proxy for the general disease status. The changes in MFF at Week 49 mirror the sigmoidal shape observed in trajectory curves, with the largest MFF changes observed in moderately affected muscles at baseline. The largest changes in MFI occurred in the muscle with low baseline MFF, suggesting that MFI may change early in the disease progression. LMV increases were observed in the least severely affected muscles at baseline, suggesting the effect of muscle growth outweighed the effect of disease progression. Future studies could benefit from taking the baseline disease state of the muscles into account, as muscles in different disease states show different trajectories. This could also be combined with composite scores calculated by grouping muscles based on disease state, rather than anatomical location [16, 18].

Using the AMRA Researcher tool (see the ‘Image analysis’ section), full characterization of muscle groups and total thigh analysis was feasible. These detailed analyses could be important for informing future functional changes. For example, an observed LMV increase in the hamstrings may not lead to improved function, as quantified by NSAA, whereas increases in adductors and quadriceps will likely result in improved function in the coming years.

As muscle pathophysiology in DMD is complex, it is important to understand the relationship between MRI measures. Participants with both high muscle volume and low MFF or MFI had the highest NSAA total scores. At low fat amounts, higher LMV appears to be associated with higher functional ability, a finding that is consistent with previous analyses [3]. In future interventional studies, it may be important to monitor both fat accumulation and muscle volume to characterize therapeutic benefits.

We demonstrated the relationship between NSAA total scores and LMV and MFI. It was shown that low LMV and high MFI lead to poor functional performance and loss of ambulation. We utilized the ratio of LMV to MFI to consider both the amount and quality of the remaining viable muscle tissue that is essential for physical function [13]. MFF was not considered in this ratio as it only relates to regions of fat, which would not contribute to the functional capacity of the muscle. The potential to combine multiple measures of LMV and muscle quality could have implications for future clinical trial designs. First, improvement in LMV or MFI may help reduce the risk of rapid decline of NSAA total scores, as study participants above ~4000 mL/% typically had higher NSAA total score performance. Second, MRI‐based muscle classification [16, 18] could potentially be used to identify diversions in treatment effects and identify patients most at‐risk for becoming nonambulatory.

This analysis had some limitations. The MRI images were acquired as part of an interventional clinical trial, and all participants received escalating doses of study treatment or placebo over the first 49 weeks. At Week 49, the therapeutic regimen changed to/from domagrozumab or placebo; therefore, every participant was exposed to domagrozumab during the course of this study. It was previously demonstrated that domagrozumab had an effect on muscle volume measures; however, the treatment did not have a significant effect on fat fraction or functional performance [6, 17, 25]. The trajectory plots shown in this analysis should not be interpreted as natural history data. Rather, these data add to the sparse literature on volumetric MRI measurements in participants with DMD. In addition, as Dixon images were required, only sites with manufacturer‐supplied fat/water images were included, thereby reducing the overall sample size [3].

Monitoring these muscle MRI biomarkers during an interventional clinical trial may be important to identify therapeutic benefits or patients at high risk for rapid functional decline and as such could potentially also have a role in patient selection. In conclusion, this study has highlighted expected MRI biomarker trajectories, demonstrated the value of analyzing multiple MRI biomarkers, discussed image analysis methodologies, and expanded the understanding of MRI parameters and functional performance for patients with DMD.

Author Contributions

Sarah P. Sherlock: writing – review and editing. Allison McCrady: writing – review and editing. Jeffrey Palmer: writing – review and editing. Haleh Aghamolaey: writing – review and editing. André Ahlgren: writing – review and editing. Per Widholm: writing – review and editing. Olof Dahlqvist Leinhard: writing – review and editing. Markus Karlsson: writing – review and editing.

Ethics Statement

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Conflicts of Interest

S.P.S. was an employee of Pfizer and held stock and/or stock options at the time of the study. J.P. is an employee of and holds stock and/or stock options in Pfizer. A.M. was a contractor at Pfizer during this work and is a student at the University of Virginia. H.A. was an employee of Quanticate Clinical Research Organization and is now an employee at Université de Montréal. P.W., A.A., O.D.L., and M.K. are employees of and hold stock in AMRA Medical. Medical writing support was provided by Neel Misra, MSc, CMPP, and Rabea Graepel, PhD, of Engage Scientific Solutions and funded by Pfizer.

Supporting information

Figure S1 Example images showing MRI parameters. The segmentation of muscle MRI images into LMV, MFF, and MFI is shown in representative MRI scans. Abbreviations: LMV, lean muscle volume; MFF, muscle fat fraction; MFI, muscle fat infiltration; MRI, magnetic resonance imaging.

Data Availability Statement

Upon request, and subject to review, Pfizer will provide the data that support the findings of this study. Subject to certain criteria, conditions, and exceptions, Pfizer may also provide access to the related individual de‐identified participant data. See https://www.pfizer.com/science/clinical‐trials/trial‐data‐and‐results for more information.