Background

Immune checkpoint inhibitor (ICI) rechallenge after an initial rheumatic immune-related adverse event (irAE) remains a high-stakes clinical decision. Current guidelines lack biomarkers to stratify rechallenge risk, forcing oncologists and rheumatologists into empirical risk-benefit assessments. The paradox is notable: patients who developed rheumatic irAEs often had superior anti-tumor responses, yet rechallenge carries 40-65% irAE recurrence rates.

Hypothesis

We hypothesize that the ratio of serum soluble LAG-3 (sLAG-3) decline rate to circulating PD-1+TIM-3+LAG-3+ exhausted CD8+ T-cell frequency measured at baseline and weeks 2, 4, and 8 post-rechallenge constitutes a dynamic biomarker system that predicts paradoxical rheumatic irAE recurrence 4–12 weeks before musculoskeletal symptom onset with >80% sensitivity and >75% specificity.

Mechanistic Rationale

1. sLAG-3 as a decoy reservoir: Soluble LAG-3, shed by ADAM10/17-mediated cleavage, normally buffers MHC-II-mediated T-cell activation. Rapid sLAG-3 consumption (declining trajectory) during ICI rechallenge signals loss of this braking mechanism, preceding clinical autoimmunity.

2. Exhaustion-to-activation phenotypic reversal: Triple-positive exhausted CD8+ T cells (PD-1+TIM-3+LAG-3+) represent a metastable state. ICI rechallenge — particularly anti-PD-1 or anti-PD-L1 — selectively reverses exhaustion in this compartment, and cells expressing self-reactive TCRs that cross-react with joint antigens (molecular mimicry with tumor neoantigens) preferentially expand.

3. The ratio as a dynamic bifurcation indicator: The sLAG-3 decline-to-exhausted-cell-frequency ratio captures the balance between immune brake consumption and effector pool reactivation. A rapid ratio decline below a critical threshold (estimated at <0.3 normalized units by week 4) would indicate the system has crossed a bifurcation point toward autoimmune flare.

Testable Predictions

• P1: In a prospective cohort of ≥80 ICI rechallenge patients with prior rheumatic irAE, sLAG-3/exhausted-CD8 ratio at week 4 will discriminate flare vs non-flare with AUROC >0.82.
• P2: Patients whose ratio crosses the critical threshold by week 4 will develop musculoskeletal symptoms 4–12 weeks later, providing actionable lead time for prophylactic hydroxychloroquine or low-dose prednisone.
• P3: The biomarker will remain predictive independent of ICI class (anti-PD-1, anti-PD-L1, anti-CTLA-4) and tumor type, suggesting a shared immunological mechanism.
• P4: sLAG-3 kinetics alone will underperform the composite ratio (ΔAUROC >0.08), confirming that the cellular exhaustion compartment provides non-redundant information.

Study Design

Prospective observational cohort, multi-center, enrolling patients with documented rheumatic irAE (inflammatory arthritis, myositis, or vasculitis) being rechallenged with any ICI. Serial blood draws at baseline, weeks 2, 4, 8, 12, and 16. Primary endpoint: rheumatic irAE recurrence by CTCAE v5 grade ≥2. Flow cytometry panel: CD3/CD8/PD-1/TIM-3/LAG-3 with MHC-I tetramer controls. sLAG-3 by high-sensitivity ELISA.

Limitations

• sLAG-3 ELISA standardization varies across platforms; multi-center validation requires harmonized assays
• Exhausted T-cell phenotyping by flow cytometry is resource-intensive and not yet standard-of-care
• Confounders include concurrent immunosuppression (which may mask the biomarker signal) and tumor progression (which independently alters immune landscapes)
• The 0.3 threshold estimate is theoretical and requires empirical derivation from discovery cohorts
• Sample size of 80 may be underpowered for subgroup analyses by ICI class

Clinical Significance

If validated, this composite biomarker would transform ICI rechallenge decision-making from empirical to precision-guided. A 4–12 week early warning window enables prophylactic immunomodulation, potentially allowing continued cancer immunotherapy while preventing disabling rheumatic flares. This addresses a critical unmet need at the oncology-rheumatology interface, where neither specialty currently owns the rechallenge risk stratification.

LES AI • DeSci Rheumatology