Personalized Medicine for the Treatment of Depression

Major depressive disorder (MDD) is biologically heterogeneous, leading to variable outcomes despite effective therapies. This chapter outlines a pragmatic personalized medicine framework that integrates multimodal evidence with the patient’s clinical and ecosocial context. We review the strengths and limitations of biomarker-informed care, emphasizing that syndrome-based subtyping lacks reliable predictive value and that artificial intelligence (AI)-driven tools must augment clinical judgment, not replace it. Genetic and pharmacogenomic findings, notably CYP2D6/CYP2C19, offer actionable dose guidance, yielding small but significant gains at the population level, while epigenetic markers remain preliminary. Blood-based signals highlight inflammation (C-reactive protein [CRP], interleukin-6 [IL-6]) and neurotrophic pathways (brain-derived neurotrophic factor [BDNF]) as promising research candidates for stratification and monitoring, but they are not yet established as routine tools for antidepressant selection or longitudinal monitoring in standard care. Circuit-level biomarkers use functional magnetic resonance imaging (fMRI)-defined “biotypes” to correlate network function with symptom clusters and treatment response, though prospective validation is essential for routine adoption. Digital phenotyping and ecological momentary assessment (EMA) provide scalable, continuous measures for symptom tracking and risk assessment, whereas robust, generalizable individual-level relapse prediction remains preliminary, requiring ongoing standardization and user engagement. Treatment personalization is reviewed across psychotherapy (e.g., personalized advantage index [PAI]-guided matching), rapid-acting antidepressants (ketamine/esketamine), and neuromodulation, where connectivity-guided transcranial magnetic stimulation (TMS) targeting shows promise but requires further confirmation and remains investigational for routine targeting. Finally, we highlight digital therapeutics and a translational agenda, including harmonized pipelines, external validation, and clinically deployable decision support, all necessary to move from heterogeneous signals to patient-centered, mechanism-informed care.