Quality and efficiency of a semi-automatic anatomy-based method for tractography of associative white matter fascicles in brain tumor patients, a case series

Tractography is a time consuming and operator-dependent process. Neuronavigation softwares can perform automatic segmentation of MRI data to recognize several anatomical structures to be included in surgical planning. Aim of the present investigation is to define if a semi-automatic (SA) method for tractography based on auto-segmented anatomical structures is non-inferior to standard ROI-based method (RB) for long associative white matter tracts in patients with brain tumor. Tractography of bilateral inferior fronto-occipital fascicle, inferior longitudinal fascicle, and superior longitudinal fascicle was performed in patients affected from brain tumor both with SA and with a standard RB technique with a commercially available neuronavigation software. The tracts produced were then compared in terms of quality and efficiency. In particular, two independent raters assigned a mark for the quality of each tract. Moreover, the time needed to perform the elaboration of each tract was evaluated. Forty patients were included in the study for a total of 240 tracts. The SA method was able to produce the planned tracts in all but one tract (99.6%). In comparison with RB tracts, the quality of the SA tracts was considered inadequate in two cases (0.8%), sufficient for neurosurgical planning in 15 tracts (6.4%) while slightly inferior or equivalent to RB in 37 (15.4%) and 185 (77.1%) tracts respectively. The time needed to elaborate the tracts was significantly inferior in the SA group (p < 0.001). A semi-automatic method was accurate and efficient in performing tractography of long associative white matter tracts in the context of neurosurgical planning for brain tumor patients. Clinical applications should be tested in future studies.