Treatment response prediction in major neurocognitive disorder with depression using the scalp-to-cortex distance of TMS targets: implications for personalized dose estimation

Abstract

Neuronavigated repetitive transcranial magnetic stimulation (rTMS) has shown encouraging efficacy in treating various brain disorders. However, substantial inter-individual heterogeneity in treatment response remains a major challenge. In this secondary analysis of a randomized clinical trial, we examined the predictive performance of transcranial features in relation to output measures and responses to rTMS treatment. Fifty-five elderly patients with depression and cognitive impairments were randomly assigned to receive a 3-week treatment of either 10 Hz active rTMS or sham rTMS. Left dorsolateral prefrontal cortex (DLPFC) and primary motor cortex (M1) were the predefined treatment targets. Scalp-to-cortex distance (SCD) was measured as the Euclidean distance between the scalp stimulation site and the corresponding cortical target. SCD-modified output difference (SCD-MOOD) was developed for quantifying the differences between the actual output and target-specific SCD-adjusted output. Responders to rTMS exhibited a greater SCD of left M1, a lower SCD DLPFC/M1 ratio and a higher SCD-MOOD than nonresponders. Logistic regression analyses revealed that clinical remission to active rTMS was significantly associated with SCD of left M1 ( p  = 0.038) and SCD-MOOD ( p  = 0.042). Moreover, using SCD metrics as predictive features, the responder-nonresponder classification model achieved an accuracy of 92.5% at week 3 and 86.2% at week 12. Pre-treatment SCD metrics of the stimulation targets can effectively predict early rTMS treatment responses. Predictive models integrating with MRI-derived SCD and SCD-MOOD showed strong specificity for active rTMS compared to sham stimulation. This approach may enable the development of effective, personalized protocols to improve region-specific TMS treatment outcomes in clinical practice. ChiCTR-IOR-16008191, registered on March 30, 2016. ClinicalTrials.gov Identifier: NCT05967390