Neuroanatomical Predictors of Response to Deep Brain Stimulation for Treatment Resistant Depression

Tejas Sankar1,2, M. Mallar Chakravarty3, Natasha Jawa2, Peter Giacobbe4, Sidney H. Kennedy4, Sakina J. Rizvi4, Clement Hamani2,3, Andres M. Lozano2

1. Division of Neurosurgery, Department of Surgery, University of Alberta, Edmonton, AB; 2. Division of Neurosurgery, Toronto Western Hospital, Toronto, ON; 3. Centre for Addiction and Mental Health, Toronto, ON; 4. Department of Psychiatry, University Health Network, Toronto, ON

Introduction: Deep brain stimulation (DBS) of the subcallosal cingulate gyrus (SCG) is an exciting and novel therapeutic approach to treatment-resistant depression. However, SCG DBS improves the symptoms of refractory depression in only some patients but not others. Identifying those patients who are likely to respond is therefore an important ongoing aim. We propose that there are pre-existing structural differences between the brains of responders and non-responders to SCG DBS, detectable prior to surgery using conventional, pre-operative structural magnetic resonance imaging (MRI) scans.

Methods: We studied pre-operative, T1-weighted MRI scans of twenty-five patients treated with SCG DBS within the last eight years at the Toronto Western Hospital. Responders (n=14) were patients whose 12-month Hamilton Rating Scale for Depression (HAMD-17) score improved by >50% from baseline prior to DBS. A trained observer blinded to patient identity measured the pre-operative volume of the SCG region in each patient. Automated measurements of hippocampal, thalamic, whole-brain, total grey matter, and total white matter volume were obtained. Automated whole-brain cortical thickness analysis was also performed.

Results: Baseline SCG and thalamic volumes were significantly larger in patients who responded to DBS. Hippocampal volume did not differ between groups. Interestingly, grey matter volume across the entire brain was significantly higher in non-responders, and the ratio of pre-DBS grey:white matter volume distinguished between eventual responders and non-responders with high sensitivity and specificity.

Discussion: Greater structural integrity of the target SCG region may correlate with response to DBS, while non-response may be related to a developmental phenotype in which the structure of the whole brain is affected.

Conclusions: There are indeed structural differences between the brains of depressed patients who respond to SCG DBS and those who do not. These structural differences may point to a subset of patients with treatment resistant depression characterized by a unique neurodevelopmental trajectory.