Visual Evoked Potential Monitoring Predicts Vision Outcomes in Recurrent Craniopharyngioma Surgery

Intraoperative visual evoked potential monitoring during extended endoscopic endonasal surgery for recurrent craniopharyngiomas provides critical predictive value for postoperative visual outcomes, establishing a distinct electrophysiological threshold that signals an elevated risk of functional visual decline. In a retrospective cohort study conducted at Beijing Tiantan Hospital, neurosurgical researchers demonstrated that an intraoperative amplitude reduction of 36.59 percent or greater in the N75-P100 wave complex represents the sole independent predictor of postoperative visual dysfunction. Across forty-two consecutive patients undergoing secondary resection through an extended endoscopic endonasal approach, postoperative visual impairment developed in eight operated eyes, representing 9.52 percent of all eighty-four monitored eyes and 16.67 percent of the patient cohort. Based on receiver operating characteristic analysis and multivariate logistic regression modeling, the research team recommended adopting a one-third reduction in the N75-P100 amplitude as a clinically practical intraoperative early warning criterion. These findings show that secondary surgical interventions around the sellar region demand substantially more sensitive electrophysiological monitoring thresholds than those applied during primary tumor resections, reflecting the unique vulnerability of previously operated optic apparatus.

Craniopharyngiomas are rare developmental neoplasms arising in the sellar and suprasellar regions, categorized pathologically into adamantinomatous and papillary subtypes. Although categorized as benign according to the World Health Organization classification system, approximately fifty percent of these tumors exhibit locally invasive growth patterns and maintain a notorious propensity for recurrence following initial intervention. Their intimate anatomical proximity to vital neurovascular structures—specifically the optic chiasm, bilateral optic nerves, pituitary stalk, and hypothalamus—makes radical gross-total surgical resection technically formidable. Because complete tumor extirpation carries significant hazards of severe neurological and endocrinological morbidity, some neurosurgical teams have increasingly favored subtotal resection coupled with adjuvant radiotherapy as an alternative treatment strategy. However, the introduction and refinement of the extended endoscopic endonasal approach has fundamentally transformed the management paradigm for craniopharyngiomas over recent decades. This direct corridor offers neurosurgeons panoramic visualization of the subchiasmatic space, facilitating improved gross-total resection rates while concurrently mitigating surgical complication profiles compared to traditional open transcranial routes.

Despite the structural visualization benefits provided by the extended endoscopic endonasal corridor, postoperative visual dysfunction remains among the most frequent and disabling complications encountered during craniopharyngioma resection. Visual deterioration following surgery does not stem solely from residual mass effect or incomplete decompression; direct mechanical manipulation, traction, thermal injury, and microvascular disruption of the optic apparatus during dissection play critical roles in causing iatrogenic neural damage. To protect neural pathways from irreversible injury, intraoperative visual evoked potential monitoring has emerged as an indispensable electrophysiological surveillance technique. By evaluating real-time functional conduction along the anterior visual pathway from the retina through the optic chiasm to the visual cortex, visual evoked potential monitoring alerts neurosurgeons to impending neural compromise before structural damage becomes permanent. The surgical team at Beijing Tiantan Hospital had previously implemented visual evoked potential monitoring for primary craniopharyngiomas, demonstrating its efficacy and proposing early warning guidelines. However, whether those original warning thresholds could be effectively transferred to recurrent craniopharyngiomas remained an unanswered clinical question.

The clinical development of intraoperative visual evoked potential monitoring spans more than half a century, tracing back to 1973 when Wright and colleagues first attempted electrophysiological visual surveillance during intraorbital operations. Early intraoperative monitoring endeavors faced major technical limitations, characterized by unstable baseline waveforms, erratic recording quality, and high vulnerability to electromagnetic interference that restricted widespread clinical adoption. Over subsequent decades, substantial technical innovations across photic stimulation, electrode configurations, and signal processing transformed the stability and predictive validity of visual monitoring. The introduction of high-intensity light-emitting diode stimulators provided stable, reproducible retinal excitation, significantly enhancing waveform clarity and amplitude constancy during skull base operations. Simultaneously, advancements in modern neuroanesthesia, particularly total intravenous anesthesia regimens, eliminated the severe suppressive effects that volatile inhalational anesthetics traditionally exerted on cortical visual responses. Together, these technological and pharmacological breakthroughs established intraoperative visual evoked potential monitoring as a dependable, safe, and increasingly sensitive tool for preserving vision during delicate endoscopic resections.

To evaluate visual evoked potential monitoring in recurrent disease, the researchers evaluated forty-two consecutive patients who underwent extended endoscopic endonasal resection for recurrent craniopharyngiomas between July 2019 and August 2022. The patient cohort consisted of seventeen men and twenty-five women, with an overall mean age of 44.07 ± 13.14 years and a range spanning from 22 to 67 years. Histopathological evaluation confirmed twenty-nine cases (69.0 percent) of adamantinomatous craniopharyngioma and thirteen cases (31.0 percent) of papillary craniopharyngioma. Prior to their secondary operations, the enrolled patients experienced a mean duration of preoperative visual impairment of 3.48 ± 4.60 months, ranging from 0 to 18 months. The mean recurrence interval between initial surgical treatment and secondary intervention was 4.58 ± 4.77 years, with a broad range from 0.25 to 21.00 years. Enrolled subjects satisfied strict inclusion criteria, including pathologically confirmed recurrent craniopharyngioma, surgical execution via the extended endonasal corridor, absence of severe preoperative visual deficits, successful continuous visual evoked potential recording with stable baseline waveforms, and complete clinical documentation under institutional ethical approval.

The intraoperative neurophysiological protocol integrated simultaneous visual evoked potential and electroretinogram recordings using the Nim-Eclipse monitoring system. Visual stimulation was delivered through dedicated light-emitting diode goggles placed over closed eyelids, configured with an illuminance range between 500 and 10,000 lux and a stimulus duration of 84 milliseconds. Cortical visual evoked potentials were captured through corkscrew electrodes positioned bilaterally at O1, O2, and Oz scalp locations in accordance with the International 10–20 system. Recording parameters included a band-pass filter of 1 to 100 Hertz, signal averaging of fewer than fifty sweeps, and an analysis window of 300 milliseconds. Concurrently, electroretinogram electrodes placed two centimeters lateral to the external canthi monitored retinal activation, confirming consistent photic delivery. Baseline visual evoked potentials were established following dura mater incision, with subsequent waveform acquisitions performed at two-minute intervals during optic canal decompression and tumor resection. Amplitude reduction ratios for the N75-P100 and P100-N145 complexes were calculated relative to baseline values, with negative changes recorded as zero.

Anesthetic depth and physiological parameters were tightly regulated throughout surgery to avoid non-surgical interference with evoked potential amplitudes. Total intravenous anesthesia was initiated using propofol at 2 to 2.5 milligrams per kilogram and sufentanil at 0.3 to 0.4 micrograms per kilogram, maintained by continuous infusions of propofol at 4 to 12 milligrams per kilogram per hour and remifentanil at 0.05 to 0.2 micrograms per kilogram per minute. Depth of anesthesia was monitored to maintain bispectral index scores between 40 and 60, alongside continuous tracking of standard vital signs. Postoperative ophthalmic evaluation was conducted by certified optometrists using standardized logarithmic visual acuity charts at five meters under controlled lighting, measured preoperatively and at least six months following surgery. Postoperative visual dysfunction was defined as a decrease of at least one line on the logarithmic chart. Baseline comparisons between patients who developed visual dysfunction and those who retained stable vision revealed no significant differences regarding age, sex, tumor histology, symptom duration, or recurrence intervals.

Reliable and reproducible visual evoked potential waveforms were acquired across all eighty-four monitored eyes in the forty-two patients. Postoperative visual dysfunction occurred in eight eyes (9.52 percent) from seven patients (16.67 percent). Eyes exhibiting visual dysfunction demonstrated significantly greater amplitude reduction ratios than unaffected eyes for both the N75-P100 complex (48.99 ± 30.43 percent versus 13.02 ± 18.00 percent, p < 0.001) and the P100-N145 complex (41.49 ± 27.89 percent versus 12.92 ± 17.19 percent, p = 0.002). Receiver operating characteristic curve analyses determined that an amplitude reduction threshold of 36.59 percent for N75-P100 yielded an area under the curve of 0.862 (p < 0.001), while a 36.65 percent threshold for P100-N145 achieved an area under the curve of 0.791 (p = 0.007). In univariate binary logistic regression, abnormal reductions in both complexes correlated significantly with visual dysfunction (p < 0.001 and p = 0.001). However, multivariate logistic regression demonstrated that an abnormal N75-P100 amplitude reduction remained the sole independent predictor of postoperative visual dysfunction, carrying an odds ratio of 9.257 with a 95 percent confidence interval of 1.124 to 76.263 (p = 0.039).

These statistical findings highlight a key divergence between electrophysiological monitoring criteria for recurrent craniopharyngiomas compared to primary lesions. In primary craniopharyngiomas, the research group previously identified warning thresholds of 50 percent for N75-P100 and 40 percent for P100-N145, with an optimal predictive threshold of 51.76 percent for N75-P100. In recurrent tumors, however, the critical threshold dropped markedly to 36.59 percent. This substantial downward shift reflects the heightened surgical complexity of recurrent operations, where scar tissue formation, distorted anatomy, and dense adhesions between the tumor and the optic apparatus lower the structural tolerance of optic fibers. Pearson correlation analysis demonstrated moderate collinearity between N75-P100 and P100-N145 reductions (correlation coefficient 0.685, p < 0.001). The authors explained that while both components showed predictive utility in univariate testing, the inherently superior predictive accuracy of N75-P100, combined with collinearity and the limited number of postoperative visual events, accounted for the loss of independent significance for P100-N145 in multivariate modeling.

In light of their findings, the authors recommended adopting a one-third reduction in N75-P100 amplitude as an early warning criterion during extended endoscopic endonasal resection of recurrent craniopharyngiomas. The investigators also acknowledged several methodological limitations inherent to their study design. The retrospective single-center framework and modest cohort size of forty-two patients with eight affected eyes constrain statistical power and broader generalizability. Furthermore, ophthalmic assessments relied exclusively on visual acuity charts rather than systematic visual field perimetry, which is clinically more sensitive to compressive optic chiasm pathology, due to logistical constraints during long-term follow-up. Additionally, preoperative baseline visual evoked potentials were not routinely recorded to establish individualized initial values. Finally, the analysis evaluated net amplitude reduction between baseline and surgical completion, omitting the potential prognostic relevance of transient, reversible amplitude fluctuations during intraoperative tumor manipulation. The authors emphasized that prospective multi-center studies integrating standardized perimetry, preoperative evoked potentials, and dynamic real-time monitoring are necessary to validate and refine individualized intraoperative protocols.

Subject of Research: Intraoperative visual evoked potential monitoring during endoscopic resection of recurrent craniopharyngiomas

Article Title: The predictive value of intraoperative visual evoked potential monitoring for postoperative visual outcomes following extended endoscopic endonasal resection of recurrent craniopharyngiomas

Article References: Tao, X., Li, K., Yang, X., Liu, J., Yang, J., Shi, J., Liang, Y., Gui, S., Li, C., Fan, X., & Qiao, H. (2026). The predictive value of intraoperative visual evoked potential monitoring for postoperative visual outcomes following extended endoscopic endonasal resection of recurrent craniopharyngiomas. Chinese Neurosurgical Journal, 12(1), Article 6. https://doi.org/10.1186/s41016-026-00425-x

Image Credits: AI Generated

DOI: 10.1186/s41016-026-00425-x

Keywords: Visual Evoked Potentials, Recurrent Craniopharyngioma, Endoscopic Endonasal Approach, Postoperative Visual Dysfunction, Intraoperative Neurophysiological Monitoring, Beijing Tiantan Hospital, Optic Nerve Protection, Surgical Neurophysiology, predictive, value, intraoperative, visual