gms | German Medical Science

Objective: To determine the threshold of intraoperative visual evoked potentials (VEP) decrease that best predict postoperative visual field defects.

Method: Visual evoked potentials were monitored for 32 patients (64 eyes) undergoing brain tumour resection. Localization of brain lesions were intra-orbital (n=2), temporal (n=9), parietal (n=6), frontal (n=8), occipital (n=4), parasellar (n=1) and posterior fossa (n=2). Visual fields were assessed preoperatively, postoperatively and at three months after surgery by Goldmann perimetry. The neurosurgeon was told to stop the surgery when VEP decrease below 50%.

Results: In 75% of patients (21/32) and eyes (48/64), reliable VEP was recorded. In 66% of patients (14/21) and in 54% of eyes (26/48) no postoperative visual changes were observed when no signal was given (true negative). In 38% of patients (8/21) and in 25% of eyes (12/48) new visual field defects was observed postoperatively when this signal was given (true positive). In 9% of patients (2/21) and in 6% of eyes (3/48) no signal was given. Visual field defects were observed postoperatively (false negative). In 14% of patients (3/21) and in 14% of eyes (7/48) a signal was given while no change was found (false positive). The threshold of 50% to give a signal correctly identified 80% of all eyes with postoperative visual changes (sensitivity). Positive alarm predicted a postoperative visual impairment in 63% of cases (positive predictive value). The absence of signal predicted an absence of postoperative visual impairment in 90% of cases (negative predictive value).

Conclusions: Intraoperative monitoring of VEP is a very effective and highly sensitive tool to detect a patient’s risk of visual change following brain lesions resection. The next step will be to decrease the signal criteria below 50% to further prevent postoperative visual defects.