Objective: Anti-synthetase syndrome (ASS) is characterized by a relapsing disease course, yet reliable biomarkers predicting relapse remain lacking. This study aimed to determine whether dynamic remodeling of peripheral immunoregulatory cells during cyclophosphamide (CYC) induction therapy predicts long-term clinical relapse in ASS.

Methods: We conducted a retrospective longitudinal cohort study including 32 newly diagnosed ASS patients treated with CYC-based induction therapy. All patients achieved clinical remission after induction and were followed for at least 18 months after the last CYC infusion. After CYC induction therapy, patients received sequential maintenance immunosuppressive therapy, mainly including mycophenolate mofetil or calcineurin inhibitors. Peripheral blood immune subsets were analyzed by flow cytometry at baseline and at the completion of induction therapy. The main populations evaluated included monocytic and polymorphonuclear myeloid-derived suppressor cells (m-MDSCs and PMN-MDSCs), regulatory T cells (Tregs), and T helper subsets (Th1, Th2, and Th17). Changes in immune cell proportions during induction were assessed using paired non-parametric tests, and the absolute changes were calculated. Clinical relapse was defined as treatment escalation due to worsening myositis or interstitial lung disease, including an increase in glucocorticoid dose or the addition of a new immunosuppressive agent. Time to relapse was calculated from the final CYC administration to the first documented flare. Associations between immune parameters and relapse were analyzed using between-group comparisons and time-to-event analyses. Optimal cut-off values were identified using maximally selected rank statistics.

Results: During follow-up, 10 of 32 patients (31.2%) developed clinical relapse. The mean follow-up duration was 29.06±5.10 months. Baseline demographic characteristics, clinical manifestations, and treatment exposure were similar between the relapse and non-relapse groups. However, patients who later relapsed had significantly higher baseline m-MDSC levels than those who remained relapse-free [1.64 (0.80, 2.23) % vs 0.86 (0.27, 1.40) %, P = 0.025]. In the overall cohort, most immune cell subsets did not change substantially during induction therapy. Treg proportions showed a slight, non-significant increase [7.45 (5.59, 9.50) % to 8.57 (6.61, 11.34) %, P = 0.067], and no significant changes were observed in MDSC subsets or effector T-cell populations (all P > 0.05). When stratified by clinical outcome, distinct patterns emerged. Patients who remained relapse-free showed significant expansion of both m-MDSCs [0.86 (0.27, 1.40) % to 1.55 (1.01, 2.52) %, P = 0.0078] and PMN-MDSCs (0.04 (0.01, 0.06) % to 0.10 (0.03, 0.36) %, P = 0.034), whereas no significant changes were observed in those who relapsed [m-MDSCs: 1.64 (0.80, 2.23) % to 1.00 (0.61, 1.80) %, P = 0.26; PMN-MDSCs: 0.08 (0.04, 0.15) % to 0.01 (0.00, 0.06) %, P = 0.15]. Treg proportions remained stable in both groups [non-relapse: 8.05 (5.63, 10.26) % to 7.88 (6.17, 9.30) %, P = 0.47; relapse: 6.62 (5.72, 8.21) % to 10.68 (8.68, 12.88), P = 0.067]. Further analysis showed that the increase in MDSCs during induction was significantly greater in the non-relapse group (P = 0.038 for m-MDSCs; P = 0.02 for PMN-MDSCs), while changes in Tregs and effector T-cell subsets were comparable between groups. In time-to-event analyses, patients with baseline m-MDSC levels >1.12% had significantly shorter relapse-free survival compared with those below this threshold (log-rank P = 0.034). Similarly, baseline Treg proportions <9.31% were associated with an increased risk of relapse (log-rank P = 0.018).

Conclusions: Impaired expansion of m-MDSCs during induction therapy, together with lower baseline Treg levels, is associated with an increased risk of clinical relapse in patients with anti-synthetase syndrome. These findings suggest that effective restoration of immunoregulatory balance during induction may be critical for maintaining disease remission. Longitudinal monitoring of m-MDSCs and Tregs may provide a practical approach for relapse risk stratification and could help guide post-induction management strategies.