Biomarkers and Pharmacogenomics for Precision Depression Therapy

1. Study Design and Overall Framework This is a prospective, randomized, parallel-controlled, single-center clinical trial, conducted in accordance with the Declaration of Helsinki and Chinese ethical regulations. A total of 220 patients diagnosed with major depressive disorder (MDD) per DSM-5 criteria will be enrolled, along with 80-100 age-, sex-, and education-matched healthy controls (retrospective data from an anonymized health check-up database). Patient participants will be randomized 1:1 via block randomization (block size 4-6) into: (1) Pharmacogenomics (PGx)-guided therapy group; (2) Conventional treatment group (guideline- and experience-based). The follow-up duration is 12 weeks, with assessments scheduled at baseline, Week 4, Week 8, and Week 12 (endpoint). Healthy controls provide only cross-sectional baseline data.
2. Participant Eligibility and Screening Inclusion criteria for patients: (1) MDD diagnosis confirmed by two senior psychiatrists; (2) HAMD-17 total score ≥17 at baseline; (3) aged 14-60 years; (4) at least junior high school education. Exclusion criteria include: other psychiatric disorders, organic CNS diseases, severe physical illnesses, substance abuse history, physical therapy within the past year, pregnancy/lactation, or significant abnormalities in ECG, routine blood tests, liver/kidney/thyroid function. Written informed consent is obtained from all patients (and legal guardians for minors or cognitively impaired individuals).
3. Interventions and Procedures (1) PGx-guided group: Buccal swab samples are collected at baseline for targeted genotyping (chip or sequencing) of key polymorphisms in drug-metabolizing enzymes, transporters, and targets, including CYP2D6, CYP2C19, CYP3A4/5, CYP1A2, CYP2B6, CYP2C9, HTR2A, and SLC6A4. The report classifies patients into ultrarapid, normal, intermediate, or poor metabolizer phenotypes and provides prescribing recommendations for SSRIs (escitalopram, sertraline, fluoxetine, paroxetine, fluvoxamine, citalopram), SNRIs (venlafaxine, duloxetine, milnacipran), NaSSAs (mirtazapine), and others (bupropion). Clinicians prioritize drugs predicted to offer better efficacy and lower risk, adjusting starting doses based on metabolic type. (2) Conventional group: Buccal swabs are collected but not analyzed (sham procedure to maintain blinding balance). Clinicians prescribe antidepressants following Chinese depression treatment guidelines, based on symptom profiles, prior medication history, and clinical judgment. Both groups allow dose adjustments during follow-up, with all changes documented.
4. Data Acquisition and Multi-Dimensional Biomarker Assessment All participants (including healthy controls) undergo baseline assessments: (1) Demographics and clinical history; (2) Peripheral blood for inflammatory markers (CRP, IL-6), endocrine function (thyroid panel: TSH, FT3, FT4), neurotrophic factor (BDNF), and routine biochemistry (liver/kidney function, glucose, lipids); (3) Urine and stool samples for future metabolomic and gut microbiome analyses; (4) Resting-state fMRI scanning, with primary extraction of voxel-mirrored homotopic connectivity (VMHC), and optional exploration of ALFF and FCD; (5) Scale assessments: HAMD-17, HAMA, and TESS. Follow-up visits (Weeks 4, 8, 12) for patient groups include only scale assessments and adverse event recording, without repeated blood draws or imaging.
5. Outcome Definitions Primary outcome: 12-week response rate, defined as the proportion of patients with ≥50% reduction in HAMD-17 total score from baseline (χ² test between groups).

Secondary outcomes: (1) 12-week remission rate (HAMD-17 ≤7); (2) Trajectory of depressive symptoms across time points (repeated-measures ANOVA); (3) Incidence of treatment-emergent adverse events (TESS); (4) Dropout rate due to adverse events; (5) Concordance between initial drug choice and PGx recommendations (descriptive, intervention group only); (6) Time to reach stable effective dose (survival analysis); (7) Frequency of medication adjustments (Mann-Whitney U test).

Exploratory outcomes: (1) Efficacy differences across distinct metabolic phenotypes; (2) Correlations of baseline VMHC and blood biomarkers with 12-week efficacy; (3) Case-control comparisons of all biomarker dimensions; (4) Construction of a comprehensive predictive model integrating clinical, imaging, blood, and genomic features using multivariate logistic regression or machine learning (e.g., random forest, SVM), with cross-validation for performance evaluation (AUC, sensitivity, specificity).

1. Sample Size Justification Sample size is calculated for the primary outcome (two independent proportions, χ² test) using PASS 15.0. Parameters: α=0.05 (two-sided), power 1-β=0.80. Based on published literature, the 12-week response rate for conventional treatment is estimated at 50% (P₁), and the PGx-guided group is expected to achieve 70% (P₂), an absolute difference of 20%. The required sample size is 86 per group. Accounting for a 20% dropout rate, the adjusted sample size is 107.5 per group, rounded up to 110 per group, yielding a total of 220 patients to ensure adequate statistical power.
2. Statistical Analysis Plan Analyses will be performed using SPSS 26.0 and R 4.0. Efficacy analysis follows the intention-to-treat (ITT) principle with multiple imputation for missing data; per-protocol analysis serves as a sensitivity check. Baseline comparisons use t-tests or Mann-Whitney U tests for continuous variables and χ² or Fisher's exact tests for categorical variables. Repeated measures are analyzed using mixed-effects models or repeated-measures ANOVA, adjusting for baseline values. Biomarker exploration: Case-control comparisons use ANCOVA (adjusting for age, sex, etc.); correlation analyses use Pearson/Spearman coefficients. Predictive modeling: Candidate variables with P<0.10 in univariate analysis enter multivariate logistic regression (forward stepwise) or machine learning algorithms. Internal validation employs 5-fold cross-validation, with AUC assessing discrimination and calibration curves assessing goodness-of-fit.
3. Data Management and Quality Control Data are recorded in an electronic data capture (EDC) system with double-entry verification. Biological samples are aliquoted, coded, and stored at -80°C with routine quality checks. MRI data are processed independently by two blinded neuroradiologists. All adverse events are assessed for severity and causality by study physicians and reported to the ethics committee. Independent monitoring audits are conducted periodically.
4. Ethical Approval and Informed Consent This study has been approved by the Ethics Committee of the Fourth Affiliated Hospital of Zhejiang University School of Medicine. Written informed consent is obtained from all patient participants (or their legal guardians) prior to any study procedures. For healthy controls, data are derived from an anonymized health check-up database archived before January 31, 2026; original consent covered research use, and the ethics committee granted a waiver of re-consent.
5. Study Timeline The total study duration is 36 months from ethics approval: 6 months for preparation and training, 24 months for enrollment and 12-week follow-up completion, and 6 months for data cleaning, statistical analysis, and final reporting.