816. Frontal structural network vulnerability and peripheral immune dysregulation in major depressive disorder: diagnostic and prognostic implications

Abstract

Abstract Background Immune dysregulation has been increasingly implicated in major depressive disorder (MDD), yet it remains unclear how peripheral immune alterations interact with brain structure and whether these mechanisms differentially contribute to diagnosis and prognosis. In particular, the relative clinical value of immune markers versus cortical structural vulnerability has not been systematically evaluated. Aims & Objectives To clarify causal immune–brain structural interactions in MDD and evaluate their diagnostic and prognostic utility. Method We integrated bidirectional Mendelian randomization (MR) and a longitudinal cohort. Large-scale genome-wide association study (GWAS) meta-analyses comprising 688,808 individuals with MDD and 4,364,225 controls, together with UK Biobank gray matter volume (GMV) phenotypes and 19 peripheral immune-related traits, were leveraged to assess causal relationships. A prospective cohort of 122 medication-naïve patients with MDD and 66 healthy controls underwent voxel-based morphometry, individualized structural covariance network analysis, and measurement of 40 peripheral inflammatory, immune, growth, and vascular markers. Diagnostic performance was assessed using receiver operating characteristic analyses, and remission during follow-up was evaluated using Cox regression and Kaplan–Meier curves. Results Genetically predicted MDD was causally associated with reduced IGF-1 levels (ORIVW = 0.95, pFDR < 0.001). Bidirectional causal associations were identified between peripheral immune signaling and regional GMV, particularly involving frontal, insular, cingulate, and striatal regions (all pFDR < 0.05). Disrupted frontal structural covariance networks, enabling stratification of MDD into three neurobiologically distinct subtypes with divergent immune profiles. Integrating immune markers with regional GMV significantly improved diagnostic discrimination between MDD and controls (AUC = 0.94), outperforming immune markers alone. In contrast, the Cox regression model showed that frontal network dysregulation predicted poorer prognosis (HR = 0.762, p = 0.003), whereas inflammatory and immune-related markers did not predict remission (all p > 0.05). Kaplan–Meier analyses demonstrated earlier and higher cumulative remission in one frontal network-defined subtype compared with the others. Discussion & Conclusions This study identifies genetically informed and possibly causal links between MDD, peripheral immune markers, and cortical structure, highlighting intrinsic immune–brain coupling at the population level. Clinically, immune markers facilitate biologically informed MDD subtyping, whereas frontal structural network vulnerability better predicts remission, indicating dissociable roles in diagnosis and prognosis.