Towards a blood-based diagnostic panel for bipolar disorder.

Area of Science:

• Biomarker discovery
• Psychiatric diagnostics
• Proteomics

Background:

• Bipolar disorder (BD) is a severe mental health condition frequently misdiagnosed, leading to suboptimal treatment outcomes.
• Early and accurate diagnosis is crucial for managing BD and potentially preventing its progression.

Purpose of the Study:

• To investigate the diagnostic utility of a novel biomarker panel for bipolar disorder.
• To assess the panel's performance in differentiating BD from other psychiatric conditions and in identifying individuals at risk.

Main Methods:

• A meta-analysis of eight case-control studies was conducted to identify a diagnostic biomarker panel for BD.
• Machine learning (lasso regression with cross-validation) was used to select 20 protein analytes from 87 available.
• The panel's predictive performance was evaluated on established BD patients, pre-diagnostic BD patients, schizophrenia, major depressive disorder (MDD), and controls.

Main Results:

• A 20-protein biomarker panel demonstrated excellent predictive performance (AUC ≥ 0.90) for BD diagnosis.
• The panel showed good performance (AUC 0.84) in distinguishing misdiagnosed BD patients from first-onset MDD.
• Fair to good performance (AUC 0.79) was observed in differentiating pre-diagnostic BD patients from controls, with demonstrated disease specificity.

Conclusions:

• The developed biomarker panel shows significant potential as a diagnostic tool for bipolar disorder.
• Early and accurate diagnosis facilitated by this panel could delay or prevent the onset of BD.
• This study highlights the clinical utility of a proteomic biomarker approach for psychiatric diagnostics.