Acceso usuarios
Guocan Ma, Fengyi Hao, Soon Kiat Chiang, Dewen Zhou, Roger S. McIntyre, Roger C. Ho
Abstract Background and objectives Bipolar disorder (BD) is a severe mental disorder whose diagnosis heavily relies on subjective symptomatic assessments, thus a need for an objective tool to assist in the timely identification and treatment of BD.
Methods We systematically reviewed the performance of objective diagnostic biomarkers for classification of BD that presented sensitivity and specificity values. A search on Ovid MEDLINE® ALL, PubMed, as well as manual searching were performed for literature dating from December 2013 to February 2025.
Results Sixty-one studies were included in the review. Twenty-four of them reported mainly molecular, fluid-based biomarkers, twenty-five reported neurophysiological examinations as biomarkers, and six reported other forms of biomarkers. The most accurate biomarkers included voice features, apoptosis-related long non-coding RNAs, PIK3R1 (Phosphoinositide-3-kinase regulatory subunit 1) and FYN mRNAs, electroencephalography (EEG), functional near-infrared spectroscopy (fNIRS), multimodal magnetic resonance imaging (MRI), and serum VGF protein, with area under the receiver operating characteristic curve (AUC) or accuracy values of greater than 0.93. The majority (thirty-six) of the studies utilized machine learning-based classification algorithms.
Conclusions The results have been promising and replicated for some biomarkers, but these results still need to be validated in larger samples. Future studies should focus on constructing larger cohorts of specific clinical subtypes of BD, predictive utility studies for BD patients initially diagnosed as major depressive disorder (MDD), and utilization of multimodal assessment and machine learning techniques.
