Impact of multiorgan fusion imaging and interactive 3-dimensional visualization for intraventricular neuroendoscopic surgery

Abstract

Background: Imaging technologies have evolved to meet the demand for improved presurgical simulations, particularly with the introduction of endoscopic surgery in the neurosurgical field.

Objective: To evaluate the effectiveness of a 3-dimensional interactive visualization method with a computer graphics model, which was created using hybrid rendering and multimodal fusion methods for neuroendoscopic surgery, and to assess whether the 2-dimensional interactive visualization method could effectively represent the microsurgical anatomical information necessary for endoscopic surgery compared with conventional 3-dimensional computer graphics models.

Methods: Ten patients scheduled for neuroendoscopic surgery for intraventricular lesions were included in the study. For the 3-dimensional interactive visualization method, a hybrid model of volume and surface rendering was created from magnetic resonance images combined with computed tomography and positron emission tomography. Preoperative radiographic images were fused with the normalized mutual information method. Visibility of anatomic structures was compared between the multifusion models and nonfusion models created from only heavy-T2-weighted images that rely solely on the surface rendering method.

Results: The average visibility score of the multifusion models was 97.5% (range, 95.6% to 100%), which was significantly higher than that for nonfusion models (35.9% to 64.1%; P = .002). The multifusion model represents an improved visualization method for preoperative virtual simulation for neuroendoscopic intraventricular surgery.

Conclusion: Our 3-dimensional imaging method is superior to conventional methods and will greatly improve the safety and effectiveness of neuroendoscopic surgical procedures for complex intraventricular lesions.