To identify pre-treatment brain structural biomarkers predictive of rTMS efficacy in subjective tinnitus.
Approach:
Study Design: Prospective enrollment of 64 patients with subjective tinnitus and 18 healthy controls; patients underwent a 2-week course of rTMS.
Data Collection: High-resolution T1-weighted structural MRI was acquired, and 242 whole-brain morphometric features were extracted.
Analysis: Univariate analysis identified features differing between responders and non-responders; a machine learning model was constructed and evaluated.
Key Findings:
Thirty-six patients (56.25%) were classified as responders based on their response to rTMS treatment.
Ten regional features distinguished responders from non-responders, including areas in prefrontal, limbic, sensorimotor, and parietal networks.
The predictive model achieved an AUC of 0.85, accuracy of 0.77, precision of 0.71, recall of 0.97, and F1-score of 0.82, indicating strong predictive performance.
Right pars triangularis of the inferior frontal gyrus (IFGtriang-R) gray matter volume was identified as the top predictor of response, with significant differences observed between groups.
Interpretation:
The enlargement of the IFGtriang-R in responders suggests a potential structural characteristic associated with treatment response, warranting further investigation.
Limitations:
No significant correlation was found between IFGtriang-R volume and clinical improvement scores after Bonferroni correction, indicating that structural features may not directly predict clinical outcomes.
No robust association was observed between structural features and baseline clinical measures, suggesting limitations in the predictive power of these biomarkers.
Conclusion:
Pre-treatment sMRI assessment of the IFGtriang-R may facilitate patient stratification for rTMS treatment.
