Direct-cortical visual evoked potential monitoring during brain tumor resection.

Background: Visual evoked potential (VEP) recording is traditionally regarded as an unreliable evoked potential monitoring technique, precluding widespread use in intracranial neurosurgery. However, VEPs can serve as a useful intraoperative adjunct for real-time detection of mechanical damage to optic apparatuses. The low obtainability and prognostic utility of VEPs are associated with transcranial recording, which typically provides non-focal information and poor signal-to-noise ratio. Direct cortical VEP (DC-VEP) recordings may offer a solution.

Methods: We evaluated the obtainability of DC-VEPs as well as their prognostic utility in predicting postoperative visual function deterioration in a series of brain tumor patients undergoing craniotomies for tumor resection. Patient records were retrospectively reviewed for all consecutive patients undergoing brain tumor resections with DC-VEP monitoring. Pre- and postoperative visual fields were characterized from patient charts and associated with the presence of intraoperative monitoring alerts to determine the sensitivity, specificity, and positive and negative predictive values (PPV, NPV) of DC-VEPs in detecting postoperative visual field deficits.

Results: Twenty-two patients (9 male, 13 female) were included, with a median age of 60 years. DC-VEPs were reliably detected in 19 of 23 included surgeries (82.6%). The reported sensitivity, specificity, PPV, and NPV in detecting postoperative visual field deficits was 60%, 92.9%, 75%, and 86.7%, respectively. There was a statistically significant association between monitoring alerts and the presence of visual field deterioration by Fischer's exact test (p = 0.0374).

Conclusions: DC-VEPs can be reliably obtained and are useful for detecting mechanical injury to optic areas and tracts during tumor resection.