Abstract
The use of neuronavigation with superimposed mapping tools has enabled visualization of key fiber tracts and improved peri-operative planning. However, a limitation of these approaches is their reliance on a static underlying brain atlas, particularly in neurosurgical patients with brain tumors. A tool that enables qualification and quantification of brain region connectivity could refine approaches to surgical resection.
We utilized a machine learning imaging platform, Quicktome™, to generate individualized functional parcels and tracts that dynamically adapt to perioperative change. The connectome was derived from a combination of diffusion tensor imaging and resting-state function magnetic resonance imaging. Matrices were generated from the functional MRI of four patients with intracranial neoplasms and the pre- and post-operative parcellation values were compared. The individual correlation and strength of regions were quantified. Hypo- and hyper-connected regions were marked as anomalous.
We present a case series of four patients to illustrate the correlation of the anomaly matrices with post-operative neurological changes. These include: post-operative delirium originating associated with salience network hypoconnectivity; visual hemineglect linked to hypoconnectivity in the dorsal attention network; and quantifiable improvements in the language network following the resolution of expressive aphasia. All differences between pre-and post-operative paired correlation values were statistically significant.
We demonstrate a novel approach to quantifying the extent to which anomalies in the functional connectome correlate with post-operative neurological changes. This has relevance in post-operative prognostication, provision of specialist therapy services, and could serve as a useful tool in surgical education and pre-operative planning.
•Peri-operative connectomic networks in four brain tumor patients were analyzed.•DTI and resting-state fMRI images were fed into Quicktome™ software.•Correlation matrices are generated to quantify perioperative network differences.•Anomaly matrices enable symptom-driven pre- and post-operative analysis.