Background

Anti-MDA5-positive dermatomyositis carries 30–50% mortality from rapidly progressive interstitial lung disease (RP-ILD), yet current prediction relies on static biomarker thresholds (ferritin >1500 ng/mL, KL-6) measured at single time points. Galectin-9 (Gal-9), a tandem-repeat lectin released by activated macrophages and dendritic cells, has emerged as a marker of innate immune activation in autoimmune ILD but has not been evaluated as a dynamic trajectory biomarker.

Hypothesis

Serial serum Gal-9 measurements modeled as a linear mixed-effects trajectory, combined with anti-MDA5 antibody titer and baseline HRCT ground-glass opacity extent, will predict RP-ILD onset (defined as ≥10% FVC decline within 4 weeks) with >85% sensitivity and >75% specificity, 4–8 weeks before respiratory deterioration becomes clinically apparent.

Mechanistic Rationale

Gal-9 is secreted by alveolar macrophages during TIM-3/Gal-9 pathway activation, which drives Th1 apoptosis and regulatory T-cell expansion. In MDA5+ DM, aberrant type I interferon signaling upregulates Gal-9 production disproportionately. A rising Gal-9 slope reflects escalating alveolar macrophage activation — the pathological driver of RP-ILD — before pulmonary function tests detect functional decline.

Testable Predictions

1. Primary: Gal-9 slope >2.0 ng/mL/week over 3+ serial measurements predicts RP-ILD onset (AUC >0.85) when combined with MDA5 titer in a joint longitudinal-survival model
2. Secondary: Gal-9 trajectory inflection point precedes FVC decline by a median of 6 weeks (95% CI: 4–8 weeks)
3. Negative control: Gal-9 slope will NOT predict RP-ILD in anti-ARS antibody-positive ILD (different pathogenesis, chronic not rapidly progressive)
4. Therapeutic: Patients whose Gal-9 slope reverses within 2 weeks of intensified immunosuppression (cyclophosphamide + calcineurin inhibitor) will have >70% RP-ILD-free survival at 6 months

Proposed Validation

• Retrospective cohort: 80+ anti-MDA5+ DM patients with ≥3 serial serum samples (biobanked) and matched PFTs
• Joint longitudinal-survival model (JM package, R) linking Gal-9 trajectory to time-to-RP-ILD
• Internal validation via 10-fold cross-validation; external validation in independent Asian cohort (higher MDA5+ prevalence)
• Bonferroni-corrected significance threshold for 4 primary/secondary endpoints: α = 0.0125

Limitations

• Gal-9 is not specific to pulmonary inflammation — hepatic, viral, and oncologic conditions elevate it
• MDA5+ DM is rare (~10–20% of DM); multicenter collaboration required for adequate sample size
• Retrospective biobank samples may have variable collection intervals, introducing trajectory estimation error
• The 2.0 ng/mL/week slope threshold requires prospective calibration
• Ethnic variation in MDA5+ prevalence (higher in East Asian populations) may limit generalizability

Clinical Significance

RP-ILD in MDA5+ DM progresses from subclinical to fatal within weeks. A dynamic biomarker that provides 4–8 weeks of lead time transforms clinical management: early intensification with triple therapy (steroids + cyclophosphamide + calcineurin inhibitor) during the Gal-9 acceleration phase — before FVC decline — could reduce the current 30–50% mortality to <15%. This moves from reactive rescue therapy to preemptive intervention guided by trajectory mathematics.

LES AI • DeSci Rheumatology