Biomarker-Guided Antidepressant Selection Boosts Response Rates by Nearly 67% in Precision Psychiatry Trial
Key Insights
A prospective biomarker-guided trial found that patients with favorable biological markers for both sertraline and bupropion achieved a 71.4% response rate versus 42.9% for those with no positive markers, a nearly 67% improvement.
The study used functional MRI brain connectivity, cognitive reward sensitivity, and clinical profiles to predict treatment outcomes in major depressive disorder (search), though the primary matched-vs-unmatched drug comparison did not reach statistical significance.
Researchers caution that the small sample size (fewer than 50 patients) and high cost of fMRI scans mean the approach is not yet ready for routine clinical deployment.
A landmark precision psychiatry study has demonstrated that using biological and behavioral markers to guide antidepressant treatment selection can substantially improve outcomes for patients with major depressive disorder (search) (MDD). Published in Nature Mental Health, the research found that patients possessing favorable biomarker signatures for two widely prescribed antidepressants achieved a 71.4% treatment response rate—a nearly 67% increase over the 42.9% response rate observed in patients lacking these biological indicators.
The study, led by researchers at the University of California, Irvine and Mass General Brigham (search)-affiliated McLean Hospital, represents one of the first prospective, double-blind trials to test biomarker-guided antidepressant treatment selection using two mechanistically distinct medications: sertraline (a selective serotonin reuptake inhibitor) and bupropion (which targets norepinephrine and dopamine).
"Depression treatment still relies far too heavily on trial and error," said Diego A. Pizzagalli, founding director of UC Irvine's Noel Drury, M.D. Institute for Translational Depression Discoveries and the study's senior author. "Patients often spend months cycling through medications before finding one that works, while symptoms worsen and suicide risk can increase. Our findings suggest we may be able to move psychiatry closer to precision medicine, where objective biological and behavioral data help guide treatment decisions from the outset."
The Clinical Challenge
Major depressive disorder (search) affects hundreds of millions of people worldwide, yet only 30% to 50% of patients respond to their first antidepressant. Even when medications eventually prove effective, individuals may endure weeks or months of debilitating symptoms, side effects, and uncertainty before improvement begins. Unlike many other fields of medicine, psychiatry still lacks objective laboratory tests or biomarkers that can reliably guide treatment decisions.
Building Predictive Algorithms from the EMBARC Study
To address this gap, the research team first developed predictive algorithms using data from the landmark national EMBARC study—a large, multisite clinical trial. The models incorporated functional MRI (fMRI) measurements of brain connectivity, cognitive control testing (Eriksen Flanker task), reward sensitivity metrics (Probabilistic Reward Task), depression severity, neuroticism, personality traits, and employment status.
The predictive model for bupropion response achieved a cross-validated area under the curve (AUC) of 0.86, with resting-state functional connectivity between the nucleus accumbens and rostral anterior cingulate cortex (NAcc–rACC) showing the largest standardized regression coefficient. The sertraline model achieved an AUC of 0.66, with employment status emerging as the strongest predictor.
The Prospective SMART-D Trial
These algorithms were then tested in the prospective SMART-D trial (NCT05537584), a single-site, double-blind, biomarker-guided study conducted at McLean Hospital. Forty-seven participants with MDD completed at least one week of treatment and were included in the primary intent-to-treat analyses. Participants underwent fMRI, cognitive testing, and clinical assessments before being randomized to receive either sertraline or bupropion, with treatment assignment either consistent or inconsistent with their biomarker profile.
The primary outcome was change in depression severity measured by the Montgomery–Åsberg Depression Rating Scale (MADRS) across the 8-week trial. Overall, the trial demonstrated strong treatment effects, with a response rate of 63.8% (30/47) and a remission rate of 53.2% (25/47) in the intent-to-treat sample. Among participants completing the full 8-week course, response and remission rates rose to 70.7% and 63.4%, respectively.
Key Findings: Biomarker Burden Matters
The study's most striking finding emerged when researchers examined overall biomarker patterns rather than drug-specific matching. The primary hypothesis—that participants receiving a drug consistent with their biomarker would outperform those receiving an inconsistent drug—did not reach statistical significance (time × consistent/inconsistent group, F(1,38.18) = 0.03, P = 0.87). Response rates were 66.7% in the inconsistent group versus 61.5% in the consistent group.
However, a significant difference emerged when comparing depression symptom trajectories based on the number of positive biomarkers a patient possessed (time × group F(2,38.32) = 3.85, P = 0.029). Patients with two negative markers showed the smallest reduction in depressive symptoms regardless of drug assignment. Response rates reached 71.4% among patients with positive biomarkers for both medications, compared with 65.4% for those with one positive marker and 42.9% for those with no positive markers—representing a 66.8% boost in response rate.
"This is important because it reinforces the idea that depression is not a single uniform illness," Pizzagalli said. "Different biological pathways likely contribute to symptoms in different people. Understanding those differences could eventually allow us to tailor treatments much more effectively."
Shared Biological Signals Across Drug Classes
The findings point to a shared biomarker signal predicting response across both sertraline and bupropion, despite their different mechanisms of action. Both drugs act on monoaminergic pathways—sertraline by inhibiting serotonin reuptake and bupropion by inhibiting dopamine and norepinephrine reuptake. The researchers note that reward processing and cognitive control, both mediated by monoaminergic signaling in the brain, may serve as markers of treatment responsiveness for monoaminergic medications broadly.
Limitations and Clinical Applicability
The researchers caution that the technology is not yet ready for routine clinical use. The final analytical sample included fewer than 50 patients, limiting statistical power to detect drug-specific differentiation. Additionally, some predictive measures relied on expensive fMRI scans that are not currently practical for most clinical settings.
"This study is an early but important proof of concept," Pizzagalli said. "It lays the groundwork for larger studies that could ultimately transform how we treat depression."
The implications could extend beyond medication selection. In the future, biomarker-guided approaches might help clinicians identify patients unlikely to respond to conventional antidepressants, allowing them to move more quickly toward alternatives such as psychotherapy, brain stimulation therapies, or ketamine-based treatments.
The research was funded by the National Institute of Mental Health, which supported the EMBARC study, and by Wellcome Leap's Multi-Channel Psych program, which supported the UC Irvine-led clinical trial.