This retrospective cross-sectional study was conducted in the First Affiliated Hospital, Zhejiang University School of Medicine (FAHZJU) in China. In-hospital databases were searched for ¹⁸F-FDG PET/CT scans of the patients with IIM that were performed between January 1, 2017 and December 31, 2020. The inclusion criteria for this study were: (1) age over 18 years old. (2) The diagnosis of DM, polymyositis (PM), or amyopathic dermatomyositis (ADM) was based on the 2017 the American College of Rheumatology (ACR) and the European League Against Rheumatism (EULAR) classification criteria (17). (3) A ¹⁸F-FDG PET/CT scan was performed during hospitalization. The exclusion criteria were: (1) newly identified or unsolved malignancies; (2) clarified overlap syndrome with other connective tissue diseases; (3) myopathy related to thyroid dysfunction, strenuous exercise, inherited metabolic disorders, and drug-induced myositis; (4) loss of follow-up for any cause within 3 months after hospitalization. A written informed consent was acquired from all the included patients and the study was approved by the Institutional Review Board of FAHZJU (No. IIT202100194).

All the clinical and biological data were collected at the time of the inclusion. Clinical information was independently collected by two rheumatologists. Data collected were demographics, course of disease, clinical manifestations, complications, laboratory results, pulmonary function, myositis disease activity assessment (MYOACT) (18), and autoantibody status. The myositis-specific autoantibodies (MSAs, anti-MDA5, anti-Jo-1, anti-OJ, anti-PL-7, anti-PL-12, anti-EJ, anti-TIF1γ, anti-Mi-2α, anti-Mi-2β, anti-NXP2, anti-SRP, and anti-SAE1) and myositis-associated autoantibodies (MAAs, anti-Ro-52, anti-PM-Scl75, anti-PM-Scl100, and anti-Ku) were identified using the immunodot assays (Euroimmun, Lübeck, Germany) according to the instructions of the manufacturer. All the included patients received immune-suppressive medications. The immunosuppressive regimens used during hospitalization were categorized into five groups: (1) steroid monotherapy; (2) steroid + disease-modifying antirheumatic drugs (DMARDs); (3) steroid + intravenous immunoglobulin (IVIG); (4) steroid + DMARDs +IVIG; (5) steroid + Janus kinase (JAK) inhibitors. JAK inhibitors mainly referred to tofacitinib and baricitinib. The DMARDs included the usage of cyclophosphamide, mycophenolate mofetil, cyclosporine, tacrolimus, azathioprine, methotrexate, thalidomide, and hydroxychloroquine.

In each patient with IIM, the HRCT scan was performed within 1 month before or after the ¹⁸F-FDG PET/CT scans. Thin-section CT images were obtained in the supine position during breath-holding and deep inspiration. The chest HRCT images were re-assessed by an experienced radiologist and a respiratory specialist who were blinded to the clinical information of the patients. The radiological ILD patterns were defined as per the American Thoracic Society/European Respiratory Society criteria, which was a multidisciplinary decision made by the rheumatologists, respirologists, and radiologists (19). RP-ILD was defined as follows in this study: previous or concurrent diagnosis of ILD, presenting with progressive dyspnea or hypoxemia, and a worsening of interstitial change on chest HRCT radiograph within 1 month after the initial visit or onset of respiratory symptoms (20, 21).

¹⁸F-FDG PET/CT had to be performed within 2 weeks of diagnosis. The PET/CT acquisitions were carried out according to the European Association of Nuclear Medicine procedure guidelines (22). The patients were instructed to avoid exercise for at least 24 h and fast for at least 6 h prior to PET/CT detection. The mean serum glucose level was 5.5 mmol/L (range 5.2–6.1 mmol/L) and confirmed before the injection of 4.0 MBq/kg ¹⁸F-FDG. Then, 1 h after the intravenous injection, whole-body FDG-PET images were obtained using a dedicated PET scanner (Biograph, Sensation 16, Siemens systems, Germany). Low-dose CT was performed with 120 kV and 30 mAs prior to emission scanning. ¹⁸F-FDG uptake was calculated by the following formula: SUV (g/ml) = regional radioactivity concentration (Bq/ml)/[injected dose (Bq)/body weight (g)].

Visual examinations for the detection of increased ¹⁸F-FDG uptake were retrospectively performed by a single trained radiologist who was blinded to the diagnosis, complications, and outcome of all the patients involved. FDG uptake was calculated by using SUVmax of a circular region of interest (ROI) with a fixed diameter of 20 mm, containing lung, liver, spleen, bone marrow, heart, digestive system, cerebellum, and muscle. Bilaterally proximal muscle observed by PET/CT contained deltoideus, biceps brachii, trapezius, iliopsoas, gluteus maximus, gluteus medius, and quadriceps.

All the statistical analyses were performed using SPSS 22.0 (Chicago, IL, USA), R 3.6.1 and Graphpad Prism 8.0 (San Diego, CA, USA). The variables were expressed as mean ± SD or median (quartiles). The categorical variables were expressed as frequencies and percentages. The comparisons were performed by using the t-test or Mann–Whitney U-test for the quantitative variables. Unordered categorical variables were compared using the chi-square test or Fisher's exact. The survival curves were drawn by using the Kaplan–Meier method and compared by using the log-rank tests. The correlation between the two continuous variables was assessed utilizing Pearson's linear analysis. All the statistical tests were performed two-sided, and a P-value < 0.05 was considered statistically significant.

A total of 69 patients who satisfied the inclusion/exclusion criteria were included in the study (Figure 1). As characterized in Table 1, 26 patients with anti-MDA5+ DM were incorporated into this study, encompassing 17 with DM and 9 with ADM. Eighteen of them (69.2%) were women and the mean age was 55.92 ± 7.68 years old. The medium follow-up time was 11.9 (4.00, 23.80) months. The anti-MDA5+ and the anti-MDA5– group were similar with respect to age, gender, course of disease, and duration of diagnosis delay. Among the patients with anti-MDA5+ DM, 53.8% had fever, 38.5% had splenomegaly, and 15.4% had gastrointestinal hemorrhage. We found no difference in the prevalence of co-existing infectious complications between the anti-MDA5+ and the anti-MDA5– group. However, the prevalence of preceding carcinoma was higher in the anti-MDA5– group (p = 0.005).

The majority of the patients with anti-MDA5+ DM had ILD (96.2%) and a considerable proportion had RP-ILD (42.3%). About 83.7% patients had ILD in the anti-MDA5– group, with 23.3% patients had RP-ILD, owing to a considerable proportion of anti-synthetase syndrome (19/43). Furthermore, compared with the anti-MDA5– group, the patients in the anti-MDA5+ group had a higher MYOACT score (p = 0.003) and lower total lung capacity (p = 0.007).

Compared with the anti-MDA5– group, the patients in the anti-MDA5+ group showed lower level of white blood cell (WBC, p = 0.001), neutrophil (p = 0.002), platelet (p = 0.005), C-reactive protein (CRP, p = 0.010), and higher level of ferritin (p = 0.012). There were no differences in serum alanine aminotransferase (ALT), serum aspartate aminotransferase (AST), nor differences in serum lactate dehydrogenase (LDH), or creatine kinase (CK) between the anti-MDA5+ and the anti-MDA5– group. Higher percentage of ADM (p = 0.007) was found in the anti-MDA5+ group. In the Kaplan–Meier analysis, the log-rank test demonstrated a significant difference (p = 0.0362) in survival between the anti-MDA5+ and the anti-MDA5– group (Figure 2).

Distribution of MSAs and MAAs in the anti-MDA5+/– groups are presented in Supplementary Table 1. MSAs were reported in all the patients. Anti-synthetase antibodies were positive in 19 cases in anti-MDA5– group: anti-PL7 antibody (n = 7), anti-PL12 antibody (n = 2), anti-EJ antibody (n = 2), anti-OJ antibody (n = 1), and anti-Jo 1 antibody (n = 7). Anti-TIF1γ antibodies were positive in six cases (p = 0.076) and anti-NXP2 antibodies were positive in eight cases (p = 0.021) in the anti-MDA5– group. MAAs were present in 22 cases in anti-MDA5– group: anti-Ro-52 antibody (n = 16), anti-PM-Scl75 antibody (n = 4), and anti-Ku antibody (n = 2).

By comparing the patients with anti-MDA5+ DM and anti-MDA5–, we found a higher bilateral lung SUVmax (p = 0.029) in anti-MDA5+ group (Figures 3A,D). Besides, a significant positive correlation between serum ferritin and bilateral lung SUVmax (r = 0.449, p < 0.001, Figure 4A) in the patients with anti-MDA5+ DM. However, no significant correlation was observed between the ESR, serum CRP levels, MYOACT score, and bilateral lung SUVmax (Figures 4B–D, respectively). These results suggest that FDG uptake (SUVmax) in the lungs by FDG-PET/CT imaging may be useful to evaluate the location and activity of ILD in the patients with anti-MDA5+ DM.

Interestingly, the patients with anti-MDA5+ DM were found to have significantly higher SUVmax of spleen (p = 0.011) and bone marrow (p = 0.048) than those in the patients with anti-MDA5– IIM (Figures 3B,C,E,F, respectively). Significant correlations between the spleen SUVmax and serum ferritin levels (r = 0.398, p < 0.001), ESR (r = 0.274, p = 0.023), platelet count (r = – 0.265, p = 0.028), and MYOACT score (r = 0.332, p = 0.005) were as well observed (Figures 5A,B,D,E, respectively). However, no significant correlation was identified between the spleen SUVmax and serum CRP/WBC levels (Figures 5C,F, respectively). In addition, we recognized a significantly positive correlation between the spleen SUVmax and bone marrow SUVmax (r = 0.564, p < 0.001, Supplementary Figure 1A), as well as bilateral lung SUVmax (r = 0.393, p < 0.001, Supplementary Figure 1B). However, there existed no significant correlation between the bone marrow SUVmax and bilateral lung SUVmax (Supplementary Figure 1C) levels. These results suggested that the spleen FDG uptakes (SUVmax) might be valuable in the assessment of disease activity in the patients with anti-MDA5+ DM.

The patients with anti-MDA5– IIM had higher SUVmax of bilateral biceps (p = 0.006). In addition, we could not detect significant differences in SUVmax of the other organs: liver, heart, digestive system, cerebellum, and other proximal limb muscles between the patients with anti-MDA5+ and anti-MDA5– IIM (Table 1).

To explore the predictive value of PET/CT scans for the unfavorable short-term outcome (death within 3 months), we further analyzed the clinical data of 69 patients with IIM. The spleen SUVmax was found to be significantly correlated with the unfavorable short-term outcome (p = 0.001) (Supplementary Table 2). A ROC curve analysis revealed an area under the curve of 0.732 and a cutoff value of 2.319 (Supplementary Figure 2A). In addition, the spleen SUVmax was found to be significantly correlated with RP-ILD (p = 0.031) (Supplementary Table 3). A ROC curve analysis revealed an area under the curve of 0.684 and a cut-off value of 2.222 (Supplementary Figure 2B).