Prediction of Non-Remission to Anti-Depressant Therapy Using Diffusion Tensor Imaging Running Title: Predicting Depression Nonremission with DTI

Objective: Over 50% of outpatients with nonpsychotic major depressive disorder (MDD) do not achieve remission with any single antidepressant medication (ADM). There are currently no clinically useful pre-treatment measures that inform the decision to prescribe or select ADMs. This report examines whether a biomarker based on diffusion tensor imaging (DTI) measures of brain connectivity can identify a subset of non-remitting patients with a sufficiently high degree of specificity that they could avoid a medication that is likely to fail. Method: MDD outpatients recruited from community and primary-care setting underwent pretreatment magnetic resonance imaging as part of the international Study to Predict Optimized Treatment for Depression (iSPOT-D; conducted Dec 2008 – June 2014). DSM-IV criteria and a 17-item Hamilton Rating Scale for Depression (HRSD17) score≥16 confirmed the primary diagnosis of non-psychotic MDD. Data from the first cohort of MDD patients (n=74) were used to calculate fractional anisotropy measures of the stria terminalis and cingulate portion of the cingulate bundle (CgC). Based on our previous data, we hypothesized that non-remission might be predicted using a ratio of these two values. Remission was defined as an HRSD17 of ≤7 following 8 weeks of open-label treatment with escitalopram, sertraline, or venlafaxine-extended release, randomized across participants. The second study cohort (n=83) was used for replication. Results: 35% of all participants achieved remission. A value>1.0 for the ratio of the fractional anisotropy of the stria terminalis over the CgC identified 38% of the non-remitting participants with an accuracy of 88% (test cohort; odds ratio [OR], 9.6; 95%CI, 2.0-45.9); 24% with an accuracy of 83% (replication cohort; OR, 1.8; 95%CI, 0.5-6.9) and 29% with an