Quantitative Muscle Magnetic Resonance Imaging Demonstrates Resolution of Muscle Edema in Refractory Dermatomyositis: A Post Hoc Analysis of a Proof‐of‐Concept Clinical Trial of Tofacitinib

Abstract

Objective

To evaluate changes in muscle edema using short‐tau inversion recovery (STIR) imaging and quantitative diffusion‐weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping in patients with refractory dermatomyositis (DM) treated with tofacitinib.

Methods

This post hoc imaging analysis included patients from a 12‐week, open‐label trial of tofacitinib 11 mg daily in refractory DM. Bilateral thigh magnetic resonance imaging was obtained at baseline and week 12 using standardized STIR, T1‐weighted, and DWI sequences. Muscle edema (ME) and fascial edema were assessed qualitatively on STIR imaging. ADC values were measured in muscles demonstrating baseline STIR hyperintensity using three nonoverlapping regions of interest per muscle. Clinical outcomes included Manual Muscle Testing of eight muscle groups (MMT‐8).

Results

Ten patients completed the study. Baseline ME was present in 32 of 320 thigh muscles (10%) across four patients. Quantitative ADC analysis, performed in the 32 edematous muscles, demonstrated significant decreases in tissue diffusivity in 11 muscles (34.5%), including the rectus femoris, vastus lateralis, vastus medialis, sartorius, gracilis, and adductors. Absolute ADC reductions ranged from −0.18 to −0.37 × 10⁻³ mm²/s. Two patients with baseline proximal weakness showed parallel improvement in MMT‐8 scores and ADC values, whereas two patients with normal strength showed subclinical ADC improvement.

Conclusion

Tofacitinib resulted in marked clinical and radiologic improvement in refractory DM. Quantitative ADC mapping detected resolution of muscle inflammation, including subclinical improvement not evident on strength or enzyme testing. ADC metrics may serve as objective imaging biomarkers of treatment response in DM.

INTRODUCTION

Dermatomyositis (DM) is a chronic autoimmune condition characterized by inflammation of the skin, skeletal muscle, and other organs. ¹ , ² Advanced imaging with magnetic resonance imaging (MRI) can detect subclinical muscle inflammation ³ even in skin‐predominant DM who lack overt muscle weakness. Accurate and objective assessment of disease activity is essential to guide treatment decisions and monitor therapeutic response, particularly in the era of targeted immunomodulatory treatment.

Recent insights into DM pathogenesis have highlighted activation of type I interferon and the JAK/STAT signaling pathways, ⁴ , ⁵ prompting the evaluation of JAK inhibitors as novel therapeutic agents. In the first open‐label pilot study, tofacitinib demonstrated clinical efficacy in refractory skin‐predominant DM, with significant improvements in Cutaneous Dermatomyositis Disease Area and Severity Index (CDASI) scores, muscle enzymes, and overall disease activity. ⁶ However, these clinical measures, while informative, rely partly on subjective assessment and may not fully capture subclinical tissue‐level changes in the muscle.

Muscle MRI provides a noninvasive means of visualizing both active inflammation and chronic muscle damage. T1‐weighted sequences are optimal for detecting chronic changes such as fatty infiltration, whereas features of muscle edema (ME) and fascial edema (FE) appear hyperintense on short‐tau inversion recovery (STIR) or fat‐saturated T2‐weighted images. ⁷ , ⁸ Although STIR imaging is sensitive for detecting ME and FE, interpretation is qualitative and reader dependent.

Conversely, diffusion‐weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping provides quantitative assessments of water mobility within tissue, reflecting the degree of cellular inflammation and edema. Higher ADC values correspond to increased extracellular water and active inflammation, whereas lower values indicate edema resolution. ⁹

In this post hoc imaging study, we analyzed muscle MRI data from patients in an open‐label tofacitinib trial to evaluate changes in muscle inflammation using both conventional STIR imaging and quantitative DWI or ADC mapping. Our purpose was to determine whether these MRI parameters can serve as complementary, objective imaging biomarkers of treatment response and muscle disease activity in DM.

PATIENTS AND METHODS

Study design and patient characteristics

This was a post hoc analysis of a prospective, open‐label, single‐arm, 12‐week pilot trial conducted at the Johns Hopkins Myositis Center to evaluate the efficacy of tofacitinib, a pan‐JAK inhibitor, in patients with treatment‐refractory DM. ⁶ The study protocol was approved by the institutional review board, and all the patients provided written informed consent.

Ten adult patients (≥18 years) with definite or probable DM according to the Bohan and Peter ¹⁰ , ¹¹ criteria were enrolled. Inclusion criteria required patients to have a baseline CDASI ¹² score ≥5, a prior diagnostic muscle or skin biopsy consistent with DM, or a positive myositis‐specific antibody. All patients demonstrated refractory disease, defined as an active disease persisting despite a 12‐week course of glucocorticoids and at least one additional immunosuppressive agent, such as methotrexate, mycophenolate mofetil, or azathioprine. Patients were required to be on a stable prednisone dose of 20 mg/day or less for at least two weeks before enrollment. Exclusion criteria included overlap syndromes, malignancy‐associated myositis, juvenile DM, or late‐stage disease characterized by the irreversible muscle damage.

Following washout of steroid‐sparing immunosuppressive agents, patients received extended‐release tofacitinib at a dose of 11 mg orally once daily for 12 weeks. Prednisone tapering was initiated at week 8, while rescue prednisone up to 20 mg daily was permitted in the event of a disease flare.

Bilateral thigh muscle MRI acquisition and analysis

A standardized imaging protocol (Supplementary Table S1) was used to perform bilateral thigh MRIs at baseline and again at 12 weeks. ME and FE were evaluated using STIR sequences, whereas muscle architecture and fatty infiltration were assessed using T1‐weighted sequences. To provide thorough coverage, axial and coronal slices were obtained through the pelvis and bilateral thighs, hip to knee. In addition to standard sequences, DWI was obtained with b‐values of 50, 400, and 800 s/mm², and ADC maps were generated to quantify tissue diffusivity from these acquisitions.

All 320 thigh muscles (16 muscles × 2 thighs × 10 patients) were qualitatively screened for ME and FE on baseline STIR images (present or absent). ADC quantification was performed only in muscles that demonstrated STIR hyperintensity at baseline. For each such muscle, three nonoverlapping circular regions of interest (ROIs) were placed entirely within the muscle belly: the first centered on the voxel of maximal STIR signal, the second and third on the immediately cranial and caudal axial slices. The mean of the three values represented the muscle's ADC (×10⁻³ mm²/s). Fatty infiltration on T1‐weighted images was graded 0 = none, 1 = mild, 2 = moderate, 3 = severe, and muscle atrophy was scored as present or absent. All MRI assessments, including identification of STIR hyperintensity and manual ROI placement for ADC analysis, were performed by a single fellowship‐trained musculoskeletal radiologist (LF) with over 23 years of postfellowship experience, who serves as chief of musculoskeletal radiology at the Johns Hopkins Hospital. To confirm that the ROI was in muscle tissue and not subcutaneous fat, and within matched regions of edema‐like signal on both the DWI and STIR sequences, the axial STIR and T1‐weighted sequences were examined simultaneously. Both the average and minimum ADC values (measured in ×10⁻³ mm²/s) were noted. For context, reported ADC values in normal skeletal muscle typically range from 0.77 to 1.15 × 10⁻³ mm²/s, ¹³ , ¹⁴ whereas inflamed muscle in myositis has been associated with higher values generally 1.3 to 1.8 × 10⁻³ mm²/s. ¹⁵ , ¹⁶

Measurements were taken for key muscle groups including the gluteus maximus, gluteus medius, gluteus minimus, piriformis, obturator internus, rectus femoris, vastus intermedius, vastus lateralis, sartorius, adductor longus, adductor magnus, gracilis, biceps femoris long head, semitendinosus, semimembranosus, and vastus medialis.

Clinical assessments

Manual Muscle Testing of eight muscle groups (MMT‐8) ¹⁷ was performed throughout the study duration at four‐week intervals, but for the purposes of this study, baseline and endpoint (12 weeks) were evaluated.

Statistical analyses

Continuous variables were summarized as mean ± SD or median (interquartile range), as appropriate. Categorical variables were expressed as counts and percentages. Changes in CDASI scores from pre‐ and posttreatment were compared using paired sample t‐tests.

Quantitative diffusion (ADC) analyses were restricted to patients with baseline STIR‐positive ME. Because multiple muscles were assessed per patients, ADC values were summarized at the patient level (mean across STIR‐positive muscles), and pre‐ and posttreatment changes were evaluated descriptively. An exploratory Wilcoxon signed rank test was performed for ADC changes, recognizing limited sample size. All statistical tests were two sided. Statistical analyses were performed using Stata version 15.0 (StataCorp LLC).

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

RESULTS

Ten patients with treatment‐refractory DM were enrolled and completed the 12‐week trial. Baseline characteristics, including demographics, autoantibody profiles, and prior treatments, are summarized in Table 1. The mean age at enrollment was 45.6 ± 10.6 years, and 7 of the 10 patients (70%) were female. All patients had skin‐predominant disease at baseline, with minimal or no muscle weakness. The mean CDASI score at baseline was 28 and mean MMT‐8 score was 147 out of 150. Creatine kinase (CK) levels were within or near the normal range for all patients, consistent with the skin‐predominant phenotype.

Table 1.

Baseline clinical and MRI characteristics by presence of muscle edema in 10 study patients in the Study of Tofacitinib in Refractory Dermatomyositis trial^*

Baseline characteristics	Muscle edema shown in STIR sequence of muscle MRI (n = 4)	No muscle edema shown in STIR sequence of muscle MRI (n = 6)
Demographics
Age at study enrollment, mean ± SD, y	46.5 ± 12.4	50.3 ± 16.5
Female, n (%)	3 of 4 (75)	4 of 6 (66.7)
Race, n (%)
White	3 of 4 (75)	4 of 6 (66.7)
Black	0 of 4 (0)	1 of 6 (16.7)
Hispanic	0 of 4 (0)	1 of 6 (16.7)
Other	1 of 4 (25)	0 of 6 (0)
Disease duration, a mean ± SD, y	6.2 ± 5.7	3.5 ± 2.3
Muscle and global disease activity
Median CK (IQR), IU/L	100.5 (94.8–108.2)	85.0 (50.8–120.0)
Median aldolase (IQR), aldolase units: U/L	8.4 (7.6–9.4)	5.2 (4.5–5.3)
Median MMT‐8 (IQR), range 0–150	146.0 (139.8–148.5)	150 (150–150)
Median patient global VAS (IQR), range 0–10	3.9 (3.1–5.5)	5.2 (4.9–5.8)
Median physician global VAS (IQR), range 0–10	8.0 (7.2–8.2)	6.8 (6.4–7.0)
Median HAQ‐DI (IQR)	0.6 (0.6–0.9)	0.8 (0.4–0.9)
Median CDASI—Activity Score	34.0 (16.5–52.5)	21.0 (20.2–22.5)
Myositis autoantibody status, n (%)
Anti‐Mi‐2	1 of 4 (25)	0 of 6 (0)
Anti‐NXP2	2 of 4 (50)	0 of 6 (0)
Anti‐TIF‐1γ	1 of 4 (25)	5 of 6 (83.3)
Concurrent treatment during trial
Prednisone 20 mg daily at entry, n (%)	2 of 4 (50)	2 of 6 (33.3)
Quantitative muscle MRI
Global mean ADC, mean ± SD, ×10−3 mm2/s	1.41 ± 0.16	1.28 ± 0.14

^*ADC, apparent diffusion coefficient; CDASI, Cutaneous Dermatomyositis Assessment and Severity Index; CK, creatine kinase; HAQ‐DI, Health Assessment Questionnaire disability index; IQR, interquartile range; MMT‐8, Manual Muscle Testing of eight muscle groups; MRI, magnetic resonance imaging; STIR, short‐tau inversion recovery; VAS, visual analog scale.

^a Disease duration calculated from date of dermatomyositis diagnosis to study enrollment.

Clinical response

After 12 weeks of treatment with tofacitinib, clinical improvement was observed in all patients using the 2016 American College of Rheumatology (ACR)/EULAR Total Improvement Score (TIS) criteria. ¹⁸ Five patients achieved moderate improvement (TIS ≥ 40), and five patients achieved minimal improvement (TIS ≥ 20). Mean CDASI scores improved from a mean of 28 to 9.5 (P = 0.0005). Two patients who had baseline muscle weakness exhibited improved MMT‐8 scores from 127 to >136 and 144 to >149.

Muscle MRI–STIR imaging

A total of 20 bilateral thigh MRI examinations (baseline + week 12 for each patient) provided 640 individual muscle evaluations. At baseline, ME was present in 32 of 320 muscles (10%) across four patients, while FE was less common, observed in 14 of 320 muscles (4.4%).

Among the 32 ME‐positive muscles, the most frequently involved muscle groups were the rectus femoris, gracilis, vastus lateralis, sartorius, adductor longus, and adductor magnus. At the patient level, these muscles were involved in 8 of 10 patients (80%), 6 of 10 patients (60%), 6 of 10 patients (60%), 5 of 10 patients (50%), and 4 of 10 patients (40%), respectively. Notably, several of the most frequently involved muscles, including the gracilis and sartorius, are not primary contributors to knee extension and are therefore not directly assessed by the MMT‐8, which may contribute to discordance between imaging findings and strength‐based outcomes.

By week 12, all 32 ME‐positive muscles reverted to an ME‐negative score on STIR, indicating resolution of edema. Representative axial STIR and DWI images from one patient (Figure 1) illustrate decreased signal intensity after treatment, consistent with the resolution of ME and restricted diffusion.

Figure 1.

Axial magnetic resonance images of the midthighs from a representative patient with dermatomyositis before (A and C) and after (B and D) 12 weeks of treatment with tofacitinib. (A) Baseline STIR image shows diffuse hyperintensity in multiple thigh muscles indicating edema. (B) Follow‐up STIR image demonstrates decreased signal intensity, consistent with resolution of inflammation. (C) Baseline DWI shows restricted diffusion in affected muscles. (D) Follow‐up DWI shows normalization of signal, correlating with clinical and ADC improvement. These findings illustrate radiologic response to therapy, with parallel improvement on STIR and DWI. ADC, apparent diffusion coefficient; DWI, diffusion‐weighted imaging; STIR, short‐tau inversion recovery.

Quantitative ADC analysis

ADC analyses were restricted to patients with baseline STIR‐positive ME (n = 4), in whom the ADC values are biologically interpretable. Across these patients, 32 STIR‐positive muscles were evaluable at baseline and follow‐up.

When ADC values were summarized at the patient level (the mean across STIR‐positive muscles), all four patients demonstrated a reduction in ADC from baseline to 12 weeks, consistent with improvement in ME (Supplementary Table S2). An exploratory Wilcoxon signed rank test performed at the patient level did not reach statistical significance (P = 0.125).

At the muscle level, ADC values decreased from baseline to week 12 in 11 of the 32 (34.5%) STIR‐positive muscles. The gluteus minimus demonstrated no appreciable change, whereas the piriformis and obturator internus showed increases in ADC. Absolute ADC reductions ranging from −0.18 to −0.37 × 10⁻³ mm²/s were observed across STIR‐positive muscles, consistent with measurable resolution of inflammation.

Two patients with baseline proximal weakness accounted for all 32 edematous muscles and demonstrated ADC improvement across multiple muscle groups in parallel with MMT‐8 gains. The other two patients had normal strength (MMT‐8 = 150) yet also exhibited ADC reductions, indicating that MRI captured subclinical improvement not detected by strength testing. No new areas of fatty infiltration or muscle atrophy were detected on T1‐weighted sequences, and CK levels remained stable and low throughout the treatment period.

DISCUSSION

In this prospective open‐label pilot study, treatment with tofacitinib resulted in consistent clinical and radiologic improvements over 12 weeks in patients with treatment‐refractory DM. According to the 2016 ACR/EULAR myositis response criteria, all 10 patients met thresholds for clinical response, with half achieving moderate improvement on the TIS. Our post hoc MRI analysis provides additional evidence that MRI can detect both overt and subclinical muscle inflammation in DM using both STIR imaging for qualitative assessment and quantitative DWI for quantitative assessments. Notably, baseline ADC values (1.41 × 10⁻³ mm²/s) were only mildly elevated relative to published normative skeletal muscle ranges (0.77–1.15 × 10⁻³ mm²/s), ¹³ consistent with the skin‐predominant phenotype and limited clinical weakness of the cohort.

STIR imaging identified ME in 10% of the muscles across four patients at baseline, including two with normal strength and normal CK, reinforcing the sensitivity of MRI for occult muscle disease. Quantitative DWI and ADC mapping demonstrated reductions in tissue diffusivity after treatment, particularly in the rectus femoris, vastus lateralis, vastus medialis, vastus intermedius, sartorius, gracilis, adductor longus, and adductor magnus. Notably, absolute ADC reductions ranging from −0.18 to −0.37 × 10⁻³ mm²/s were observed across STIR‐positive muscles, consistent with measurable resolution of ME.

These quantitative MRI changes paralleled clinical improvement in the two patients with baseline weakness and revealed subclinical improvement in the two patients with normal strength, supporting the biologic relevance of quantitative MRI metrics in DM. However, muscle‐level correlation between MRI findings and strength could not be systematically assessed. MMT‐8 evaluates composite muscle groups rather than individual muscles and has limited sensitivity in patients with minimal baseline weakness, resulting in ceiling effects in this cohort. Consequently, changes in frequently involved muscles such as the gracilis, sartorius, and adductor groups, which are not primary contributors to the MMT‐8, may not be fully reflected in strength‐based outcomes.

Overall, this clinical–radiologic concordance supports the biologic validity of quantitative MRI as a biomarker of tissue‐level response in DM. These findings extend the previous work by applying DWI to chronic inflammatory myopathies and support ADC mapping as a sensitive biomarker of treatment‐related change. ¹⁹ , ²⁰ The significance of subclinical muscle involvement for long‐term strength outcomes is unclear, and further longitudinal studies are needed to address this question.

Importantly, CK remained normal and unchanged throughout the trial, highlighting a key limitation of enzyme‐based monitoring in skin‐predominant DM. MRI, particularly ADC mapping, provided objective and quantifiable evidence of tissue‐level response. These results support the utility of quantitative MRI as a complementary imaging biomarker in clinical trials and therapeutic monitoring. Furthermore, qualitative reductions in STIR signal paralleled quantitative ADC decreases, suggesting that a combined qualitative–quantitative imaging approach may provide a comprehensive framework for assessing therapeutic response in DM.

This study has several limitations that should be considered when interpreting the findings. The small sample size limits statistical power and generalizability. The open‐label, single‐arm design introduces potential bias and precludes direct comparison with untreated controls. MRI evaluation was confined to the thigh muscles, precluding assessment of deeper proximal muscle groups such as the pelvic girdle. ADC analysis was performed only in muscles with baseline edema and relied on manually defined ROIs, which may introduce selection bias. MRI interpretation and ROI placement were performed by a single expert reader, which may introduce reader‐dependent bias despite the radiologist's extensive experience. Only four patients had MRI‐detectable edema at baseline, further constraining interpretation. For contextual comparison, baseline and follow‐up ADC values from STIR‐negative muscles are provided in Supplementary Table 3. Finally, the 12‐week duration limits evaluation of long‐term treatment durability.

Despite these limitations, tofacitinib demonstrated clinically meaningful efficacy in refractory DM, with improvements in both cutaneous and, in a subset, muscular manifestations. Quantitative DWI and ADC mapping provided objective evidence of reduced ME, even in patients with normal strength and enzyme levels. These results highlight the potential of quantitative MRI, particularly ADC metrics, as reproducible biomarkers of disease activity and therapeutic response in DM. Larger, controlled, and longitudinal studies are warranted to validate these findings and establish standardized MRI‐based endpoints for future clinical trials.

AUTHOR CONTRIBUTIONS

All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Dr Paik confirms that all authors have provided the final approval of the version to be published and takes responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements.

ROLE OF THE STUDY SPONSOR

Pfizer had no role in the study design or in the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication. Publication of this article was not contingent upon approval by Pfizer.

Supporting information

Disclosure form.

Supplementary Table 1: Patient‐level changes in ADC, STIR status, and MMT‐8 among patients with baseline STIR‐positive muscle edema.

Supplementary Table 2: Standardized axial mid‐thigh MRI acquisition parameters (1.5 T Siemens) used at baseline and 12‐week follow‐up: STIR for edema, T1‐weighted for architecture/fat infiltration, and DWI

Supplementary Table 3: ADC values in STIR‐negative dermatomyositis patients (baseline STIR absent).