Precision Medicine in the Depression Treatment

Overview

The BrainDrugs-D study uses multimodal neuroimaging combined with self-report measures, clinical and molecular markers to identify clinically relevant predictors that can identify subtypes of major depressive disorder (MDD) and, in a naturalistic setting, predict treatment response to standard antidepressive treatment. The cohorts are followed in nationwide health registries.

BrainDrugs-D is a cohort study of patients with major depressive disorder (MDD) who are deeply phenotyped with demographic, clinical, genetic, biochemical and neuroimaging modalities. These features are subsequently examined for their ability to identify subtypes of MDD and to predict treatment response. Treatment and study population All participants are phenotyped before initiating a standardized 'treatment package' in out-patient clinics within the Mental Health Services in the Capital Region of Denmark. The goal is to recruit a total of 800 patients with non-psychotic MDD. We use broad inclusion criteria to enable recruitment of representative adult out-patients with non-psychotic MDD who receive standard treatment in practice. As the study is designed with a high degree of ecological validity, it will not interfere with or delay the standard treatment package for depression. The treatment package is a national uniform package designed by Mental Health Services in the Capital Region, it has been in use since 2017, after several preceding years of clinical use and patient experience. Treatment for first-episode depression has a manualized group Cognitive-behavioural therapy (CBT) as the backbone: 2-3 hours of initial workup followed by 6 hours of individual therapy or 12 sessions of 2 hours of group therapy (8 patients per group), 1-2 hours of engagement and psychoeducation of relatives, 1-5 hours of medication clinic and 2 hours of relapse prevention. Antidepressant medication and individual psychotherapy are instituted, as needed. Groups The study comprises three groups: The entire cohort (n=800) will have basic clinical, cognitive, psychometric, and biological data available. A subgroup (Subcohort I, n=600) provided expanded clinical, cognitive, psychometric, and biological data as well as Magnetic Resonance Imaging (MRI) and Electroencephalogram (EEG). Subcohort II, (n=60) will be exclusively for patients unmedicated at initiation, consisting of the same investigation, and contributing Positron Emission Tomography imaging with the \[11C\]-UCB-J tracer of synaptic density. Follow-up All cohorts receive questionnaires assessing depression symptom severity, level of functioning, and QoL at the three follow-up time points. We also assess the side effects of psychological treatment and medication at the end of the treatment package. The cohorts are also followed in nationwide health registries. Outcomes The primary outcome is remission (QIDS ≤5) and clinical improvement (≥50% reduction in QIDS) after 6 months. Secondary endpoints include remission status 12 and 18 months after treatment start and change in QIDS, SCL10, WHO-5, and SDS scores from baseline to follow-ups. Analysis We will use machine learning algorithms to determine a combination of baseline characteristics that best predict treatment outcomes and statistical models to investigate the association between individual and clinical outcomes. We will also assess associations between patient characteristics, treatment choices, and clinical outcomes using path analysis, enabling us to estimate the effect of treatment choices and timing on the clinical outcome. Hypotheses for the whole cohort: Primary hypotheses: 1.1 Clinical, cognitive, psychometric, genetic, and blood biomarker measures at inclusion can predict clinical remission (defined as QIDS≤5) at the first follow-up. 1.2 Clinical, cognitive, psychometric, genetic, and blood biomarker measures at inclusion can predict clinical improvement (a ≥50% reduction in QIDS from pretreatment) at the first follow-up. Secondary hypotheses: 1.3 Composite scores across a range of clinical, cognitive, psychometric, genetic, and blood biomarker measures at inclusion can cluster patients into MDD subgroups associated with treatment trajectories and outcomes. 1.4 Clinical, cognitive, psychometric, genetic, and blood biomarker measures at inclusion are associated with clinical outcome defined as a change in QIDS. 1.5 Path analysis of baseline patient characteristics and treatment tracks can uncover causal paths for clinical improvements, i.e., estimate the effect of treatment on clinical outcomes. Hypotheses for Subcohort I Primary hypotheses: 2.1 MRI, fMRI, and EEG patterns at inclusion may be associated with depressive phenotypes. 2.2 Adding EEG, MRI, and fMRI measures at inclusion to the classifier model (defined in hypotheses 1.1 and 1.2) may significantly improve the prediction of clinical remission and improvement. Secondary hypotheses: 2.3 Adding EEG, MRI, and fMRI measures at inclusion to the composite score (defined in hypothesis 1.3) may significantly improve the clustering of patients into MDD subgroups. Hypotheses for Subcohort II Primary hypotheses: 3.1 Cerebral \[11C\]-UCB-J binding is lower in patients with MDD than in healthy controls. 3.2 Domain-specific cognitive function correlates positively with \[11C\]-UCB-J binding in associated cortical and subcortical areas. Secondary hypotheses: 3.1 Depression severity, anxiety, and anhedonia correlate with \[11C\]-UCB-J binding in associated cortical and subcortical areas. 3.2 Addition of \[11C\]-UCB-J binding, EEG, and MRI measures at inclusion to the composite score (defined in hypotheses 2.1) can significantly improve the prediction of clinical improvement and remission beyond clinical, cognitive, psychometric, fluid biomarker, EEG, and MRI measures in antidepressant naïve patients.