Visual Field Outcome Reporting in Neurosurgery: Lessons Learned from a Prospective, Multicenter Study in Transsphenoidal Pituitary Surgery

Background Visual field (VF) outcomes are commonly reported in neurosurgical case series; however, detailed methodology is rarely reported.

Objectives To examine the methodology and reliability of VF outcomes ongoing, multicenter study of transsphenoidal pituitary surgery.

Methods Optional VF testing results were collected during a prospective, multicenter clinical trial evaluating patient outcomes after transsphenoidal surgery for nonfunctioning pituitary adenomas (TRANSSPHER Study; ClinicalTrials.gov NCT02357498). Both preoperative and postoperative VF studies were evaluated. Two independent ophthalmologists reviewed the VF studies for reliability and analyzed the results.

Results Suboptimal perimetry field settings were reported in 37% of VF studies. Preoperative perimetry testing was performed at a mean of 1.8 months (range: 0.03–12 months) before surgery. Postoperative perimetry testing was performed at a mean of 2.4 months (range: 0.2–13 months) after surgery. Automated reliability criteria characterized 25 to 29% of VF studies as unreliable, while evaluation by two independent ophthalmologists demonstrated that 16 to 30% of studies were unreliable. Agreement between automated criteria and raters on VF reliability was inconsistent (kappa coefficients, 0.42–0.83); however, agreement between the two raters was “substantial” to “almost perfect” (kappa coefficients, 0.78–0.83). According to preoperative and postoperative paired tests, most patients demonstrated improvement after surgery (Rater 1, 60%; Rater 2, 67%) with “substantial” rater agreement on outcomes for paired examinations (kappa coefficient, 0.62).

Conclusion VF outcome measures collected in a prospective, multicenter trial demonstrated significant heterogeneity in testing parameters and patient performance. VF test settings were variable both between patients and for some patients preoperatively compared with postoperatively. Furthermore, the timeline of postoperative follow-up was heterogeneous, indicating widely differing practice patterns across centers. Reliance on automated criteria alone did not substitute for independent ophthalmologist review regarding the reliability of studies. Standardized guidelines for VF data collection and reporting will increase the reliability of results and allow for better comparisons of outcomes in future studies.