Diagnostic performance of T1-Weighted MRI in bipolar disorder: A systematic review and meta-analysis

Bipolar affective disorder (BPAD) is a clinically defined psychiatric condition characterized by heterogeneous symptom patterns and a reliance on subjective assessments. Structural neuroimaging, particularly gray matter volume (GMV) from T1-weighted MRI, has been investigated as a complementary diagnostic tool. However, its clinical applicability remains uncertain. This systematic review and meta-analysis evaluated the diagnostic performance of GMV in BPAD by synthesizing sensitivity, specificity, and classification accuracy, while accounting for demographic and methodological variability.

A comprehensive PRISMA-based literature search identified 15 structural MRI studies including 1645 participants. Of these, eight studies (n = 780) met inclusion criteria for quantitative synthesis based on reporting diagnostic metrics such as sensitivity and specificity. Pooled estimates were calculated, and subgroup and meta-regression analyses were conducted to examine the influence of age, sex, and control group characteristics.

GMV demonstrated moderate diagnostic accuracy, with a pooled sensitivity of 0.69 and a specificity of 0.77. Classification performance was higher in older adults than in younger cohorts. Sex-related GMV differences were observed in healthy controls but were attenuated in BPAD. Meta-regression identified control group size as a significant moderator of heterogeneity, with additional variability likely due to medication exposure and illness subtype.

GMV may provide incremental diagnostic value when integrated with clinical evaluation, particularly in older adults. Nonetheless, methodological heterogeneity and the modest accuracy observed indicate that GMV alone is insufficient as a biomarker. By incorporating the most recent evidence, these findings reinforce the need for multimodal diagnostic approaches that combine structural MRI with genetic, functional, and clinical data.