Brain Structural Connectomic Topology Predicts Medication Response in Youth with Bipolar Disorder: A Randomized Clinical Trial

Du Lei,Kun Qin,Wenbin Li,Walter H.L. Pinaya,Maxwell J. Tallman,Jingbo Zhang,L. Rodrigo Patino,Jeffrey R. Strawn,David E. Fleck,Christina C. Klein,Qiyong Gong,Caleb M. Adler,Andrea Mechelli,John A. Sweeney,Melissa P. DelBello

DOI: https://doi.org/10.1016/j.jad.2024.11.061

IF: 6.533

2024-01-01

Journal of Affective Disorders

Abstract:Background Response to pharmacotherapy varies considerably among youths with bipolar disorder (BD) and is poorly predicted by clinical or demographic features. It can take several weeks to determine whether medication for BD is clinically effective. Although neuroimaging biomarkers are promising predictors, few studies examined the predictive value of the brain connectomic topology. Methods BD-I youth (N = 121) with no prior psychopharmacotherapy were randomized to 6-weeks of double-blind quetiapine or lithium. Structural magnetic resonance imaging (MRI) was performed before medication and at one week after medication initiation. Brain structural connectome was established from the MRI scans, and topological metrics were calculated for each patient. Deep learning-based prediction model was built using baseline and one-week connectome topology to predict medication response at week 6. Results Both baseline topological metrics and one-week topological changes could predict treatment response with significant accuracy (73.8 % - 86.8 %). A longitudinally joint model combining baseline and one-week topology provided the highest accuracy (91.3 %). The transferability between models for quetiapine and lithium was relatively poor. In addition, predictions for the two drugs were driven by similar baseline but distinct one-week salient topological patterns. Limitations Independent replication is needed to validate our preliminary findings. Conclusion Brain structural connectomic topology at baseline and its acute changes within the first week enable accurate BD medication response prediction. The most contributive brain regions differed between prediction models for quetiapine and lithium after one week. These findings provide preliminary evidence for the development of neuroimaging-based biomarkers for guiding therapeutic interventions in youth with BD.