News

Article

Multimodal Diagnostic Method Demonstrates Higher Accuracy Predicting Bipolar Disorder Than Single Data Methods

In adolescents, a comprehensive model using behavioral and MRI data had a prediction accuracy of 0.83, which was higher than single data type models.

Findings published in Biological Psychiatry demonstrate significant steps in improving early diagnosis of bipolar disorder in adolescents. The development involves the effective integration of multimodal magnetic resonance imaging (MRI) that involve behavioral assessments to improve diagnostic precision.1

Health care worker looking at MRI scans -- Image credit: auremar | stock.adobe.com

Image credit: auremar | stock.adobe.com

Bipolar disorder is a mental illness that is characterized by extreme mood swings which include episodes of depression and hypo/mania. It has a significant hereditary component with an estimated heritability of approximately 70% and often emerges during adolescence. Oftentimes, the disorder presents with mild initial symptoms before evolving into “classic” manifestations. The authors note that early intervention is significant because about 50% to 70% of individuals show mood symptoms by 21 years of age; however, clinical diagnostics—particularly in adolescents who are at risk of bipolar disorder—are challenging because of the lack of clarity in subthreshold symptoms.1,2

For this retrospective study, the investigators aimed to improve early detection and treatment of bipolar disorder. A total of 309 patients were enrolled and included patients with bipolar disorder, offspring of patients with bipolar disorder (both with and without subthreshold symptoms), offspring who do not have bipolar disorder but are showing subthreshold symptoms, and healthy controls. The investigators integrated behavioral attributes with multimodal MRI features from T1, resting state functional MRI (rsfMRI), and diffusion tensor imaging (DTI). Additionally, 3 diagnostic models were developed using GLMNET multinomial regression: a clinical diagnosis model based on behavioral attributes; an MRI-based model; and a comprehensive model that integrated both datasets.2

“When people think about MRI, they often imagine that a radiologist is looking for gross abnormalities that might be easily visible when the MRI is presented on a screen. However, in this case, very detailed automated MRI analyses are conducted that might pick up structural differences that might not be apparent otherwise,” John Krystal, MD, editor of Biological Psychiatry, said in a news release. “Further, this study employed 2 other forms of MRI imaging, rsfMRI, and DTI. rsFMRI analyzes the pattern of communication between brain regions, [such as] their correlated activity. DTI measures the structural integrity of neural pathways by measuring the movement of water within and around the neural pathways in the brain.”1

According to the findings, the comprehensive model that used both behavioral and MRI data demonstrated a higher prediction accuracy compared with the single data type models. This method achieved a prediction accuracy of 0.83 (CI: 0.72, 0.92) and was noticeably higher than the clinical and MRI-based models, which had prediction accuracy scores of 0.75 and 0.65, respectively.2

“Given the challenges in diagnosing bipolar disorder in adolescents, our findings mark a major advance in early detection. By combining various imaging techniques, we can now identify at-risk young people with remarkable accuracy,” Kangguang Lin, MD, PhD, from the department of Affective Disorders at the The Affiliated Brain Hospital of Guangzhou Medical University and Guangzhou Medical University, said in the news release.1

Additionally, validation with an external cohort presented high accuracy (0.89; AUC = 0.95). Structural equation modeling also revealed that clinical diagnosis (β = 0.487, p < .0001), parental bipolar disorder history (β = -0.380, p < .0001), and global function (β = 0.578, p < .0001) significantly impacted brain health. In addition, psychiatric symptoms showed only a slight influence (β = -0.112, p = .056).2

“We have long expected that brain imaging would help to improve the psychiatric diagnostic process. Here we have a very promising example where the information gained from several forms of MRI, combined with clinical information, improve upon the accuracy of traditional clinical diagnosis,” said Krystal in the news release.1

The investigators are hopeful that the combination of neuroimaging can allow for more accurate patient subgroup distinctions while also enabling timely interventions and improving health outcomes. As far as limitations, the authors acknowledge there may be selection biases, small sample sizes, and potential challenges for universal applicability because of the complexity and variability of adolescent development. In addition, they note that a prospective study that explore long-term follow-ups would be an effective method of assessing the diagnostic accuracy and enhance the tool’s reliability.2

“The integration of behavioral and neuroimaging data has the potential to transform the field of neuropsychiatric diagnosis, particularly for conditions like bipolar disorder where early detection is crucial. This approach could lead to earlier interventions, potentially improving outcomes for affected individuals,” Jie Wang, PhD, from the Songjiang Research Institute and Songjiang Hospital at the Shanghai Jiao Tong University School of Medicine, said in the news release.1

REFERENCES
1. Elsevier. Researchers achieve a significant advancement in early diagnosis of bipolar disorder in adolescents. News release. September 19, 2024. Accessed September 25, 2024. https://www.eurekalert.org/news-releases/1058655
2. Wu J, Lin K, Lu W, et al. Enhancing Early Diagnosis of Bipolar Disorder in Adolescents through Multimodal Neuroimaging. Biol Psychiatry. 2024. doi:10.1016/j.biopsych.2024.07.018
Related Videos
Pharmacist helping patient -- Image credit: Clayton D/peopleimages.com | stock.adobe.com
The EMPOWER-Lung 1 Trial: Five-Year Outcomes of Cemiplimab Monotherapy in Advanced NSCLC
Pharmacist and a patient -- Image credit: Zamrznuti tonovi | stock.adobe.com
October is American Pharmacists Month.
American Pharmacist Month | Image Credit: Zoran Zeremski - stock.adobe.com
Bacteria of clostridioides difficile infection -- Image credit: Artur | stock.adobe.com
smiling indian male doctor or pharmacist in white coat with stethoscope and clipboard over drugstore background
Efficient healthcare supply chain management ensures timely delivery of medical supplies and medications
Pharmacy School, social media, non-traditional learning | Image Credit: Ахтем - stock.adobe.com