Quantitative analysis of orbit volume and eyeball position after resection of hyperostotic sphenoid wing meningioma and simultaneous orbital wall reconstruction using individual technologies

Abstract

To conduct a quantitative analysis of orbit volume at different stages of preparation and surgical treatment of patients with cranio-orbital meningiomas undergoing resection with simultaneous orbital wall reconstruction using 3D modeling and 3D printing technologies. A prospective cohort non-randomized study included 24 patients with cranio-orbital meningiomas. The volumes were measured by segmenting the orbital structures using the planimetric method in the Inobitec PRO software package. Three expert neurosurgeons independently performed these measurements. The implants were modeled in Blender software. We used the intraclass correlation coefficient (ICC) and global similarity index (GSI) to analyze interrater agreement and ensure reproducibility of analysis. Interrater agreement on orbital markings was very high for both metrics (ICC and GSI). The ICC (A,3) for intact orbit volume was 0.99 (95% CI 0.981; 0.996, p=1.9962018^{-33}), 0.99 (95% CI 0.983; 0.996, p=1.903203^{-34}) for damaged orbit volume at the preoperative stage, 0.99 (95% CI 0.979; 0.995, p=3.5939828^{-32}) for damaged orbit volume at the stage of modeling of resection and reconstruction, 0.99 (95% CI 0.978; 0.995, p=1.1048941^{-30}) for damaged orbit volume in postoperative period. The ICC for measurements related to EI was 0.94-0.97 (very high). This analysis revealed a strong inverse relationship between EI and volume index at the preoperative stage (rho= -0.55, p=0.004987), as well as between dynamics of EI and volume indexes in perioperative period (rho= -0.59, p=0.003). We found significant relationship (p=0.006757) between implant displacement in the area of lateral orbital wall and differences of actual and theoretical volumes. The proposed method of planimetric contouring and segmentation of orbital volumes is highly accurate and reproducible. Significant patterns allow us to develop predictive models for preliminary calculation of target volume of the damaged orbit.