Can DTI fiber tracking of the optic radiations predict visual deficit after surgery?

Sparing optic radiations can be of paramount importance during epilepsy surgery of the temporal lobe. The anatomical heterogeneity of the Meyer's loop of the optic radiations could be assessed by means of diffusion tensor tractography. We used temporal lobe surgery as a lesion model to validate this method.

We analyzed the distance between the temporal pole (TP) and Meyer's loop (ML) and the correlation between visual impairment and the percentage of virtual fibers injured. MRI studies were performed in 18 patients and 13 controls. Diffusion tensor imaging (DTI) with fiber tracking was performed using four different algorithms and various gradient directions (15 or 32) and fractional anisotropy (FA) thresholds (0.18, 0.20, and 0.22). To find the best DTI model, we tested each gradient direction and FA threshold on 16 operated patients by pre- and post-operative visual field testing that analyzed the percentage of virtual fibers damaged on 3-month-post-operative MRIs.

Marked individual differences were noted in the TP-ML distances (mean: 25.4mm; range 18.2-38.3mm; standard deviation: 4.7) but with no significant difference between patients and controls (p=0.9). The percentage of virtual fibers reconstructed by tracking and damaged by surgery was correlated with visual impairment. Significant differences appeared between algorithm types. The tensor-line algorithm with 15-direction resolution and an anisotropy threshold of 0.18 seemed to be the most relevant. A threshold of 5.5% of injured virtual fiber could predict a visual defect with a sensitivity of 71.4% and a specificity of 87.5%.

Optic radiation tractography by DTI could be a useful method to assess an individual patient's risk of postoperative visual deficit.