Integrating Qualitative and Quantitative MRI Analysis for Optimizing DBS Candidate Selection in Patients with Disorders of Consciousness

Marina Raguz^1,2Petar Marčinković^1*Hana Chudy³Valentina Galkowski⁴Maja Majdak³Darko Orešković¹Darko Chudy^1,5

• ¹Department of Neurosurgery, Clinical Hospital Dubrava, Zagreb, Croatia
• ²School of Medicine, Catholic university of Croatia, Zagreb, Croatia
• ³Department of Neurology, Clinical Hospital Dubrava, Zagreb, Croatia
• ⁴Department of Diagnostic and Interventional Radiology, Clinical Hospital Dubrava, Zagreb, Croatia
• ⁵School of Medicine, University of Zagreb, Croatia, Zagreb, Croatia

Introduction: Disorders of consciousness (DoC) encompass a spectrum of clinical conditions with often indistinct boundaries, making accurate diagnosis and therapeutic decision-making particularly challenging. While advanced imaging techniques such as fMRI and PET reduce misdiagnosis risk, their limited availability in routine clinical settings underscores the need for alternative approaches. This study investigates whether the integration of qualitative and quantitative parameters derived from conventional MRI can improve diagnostic precision and support more accurate deep brain stimulation (DBS) candidate selection in DoC patients. Methods: Fifty consecutive DoC patients underwent comprehensive clinical, neurophysiological, and MRI assessment. Based on an integrated assessment of these findings, patients were classified as DBS candidates or non-candidates. MRI scans were qualitatively assessed for cortical and subcortical atrophy (including diffuse cortical, thalamic, and brainstem degeneration), ventricular enlargement, sulcal widening, leukoaraiosis, corpus callosum damage, gray-white matter border effacement, and extensive lesions (e.g., global ischemia or porencephalic cavities). Quantitative volumetric analysis was performed using the FreeSurfer pipeline. Results: Qualitative features such as leukoaraiosis, thalamic and cortical atrophy, ventricular enlargement, and corpus callosum lesions were significantly associated with DBS candidacy. Quantitative predictors included striatal volume, total gray matter, ventricular volume, CSF, and supratentorial volume. A combined model incorporating both qualitative and quantitative MRI data achieved high predictive accuracy (AUC=0.88) for DBS candidacy. Conclusion: Integrating conventional MRI-based qualitative and quantitative assessments with clinical and neurophysiological evaluation may substantially improve DBS candidate selection in DoC patients, especially where functional imaging is unavailable. These findings support the development of practical MRI-based decision frameworks and call for multicenter validation. Despite increasing research on imaging and neuromodulation in DoC, studies directly comparing qualitative and quantitative structural MRI in the context of DBS candidacy remain scarce, highlighting a critical gap in the field.