Structural brain network metrics as novel predictors of treatment response in restless legs syndrome

Abstract

Objective: This study aimed to investigate morphometric similarity networks in patients with newly diagnosed restless legs syndrome (RLS) compared with healthy controls and to examine their relationship with treatment response.

Methods: A total of 49 patients with newly diagnosed RLS and 58 healthy controls were prospectively enrolled. Brain magnetic resonance imaging was performed using a 3-T scanner, and morphometric similarity network analysis was conducted on T1-weighted images. The severity of RLS was assessed using the International RLS Scale at baseline and at three months post-treatment initiation. Patients were classified as good or poor responders based on a decrease of ≥5 points in RLS severity scores following treatment with either pramipexole or pregabalin.

Results: Although no significant differences were observed in morphometric similarity networks between patients with RLS and controls, both modularity and small-worldness indices were lower in the RLS group (0.218 vs. 0.258, p = 0.023; 0.841 vs. 0.861, p = 0.042). Among the 40 patients who completed follow-up evaluation, 27 were good responders and 13 were poor responders. Network diameter was significantly higher in good responders than in poor responders (7.061 vs. 6.552, p = 0.002). Similarly, eccentricity was elevated in good responders (5.875 vs. 5.385, p = 0.008). Receiver operating characteristic curve analysis revealed high predictive values for both diameter and eccentricity (AUC = 0.838, p < 0.001; AUC = 0.751, p = 0.002, respectively).

Conclusion: Network metrics, particularly diameter and eccentricity, demonstrate potential utility as biomarkers for predicting treatment response in patients with RLS.