In this study, we developed a unique AR navigation system, which applied detailed three-dimensional computer graphics (3DCG) created from the preoperative image and using the intuitive real-time adjustment method. The aim of this study was to evaluate the efficacy of this augmented reality (AR) neuronavigation system during ETS for parasellar tumors, based on the 3DCG model created from the preoperative radiographic data.

CT and MRI with gadolinium enhancement were acquired before surgery in all patients and were used to create the 3DCG VR model. CT was performed with a 320-slice CT scanner (Aquilion ONE TSX-306A; Canon Medical Systems, Tokyo, Japan) using the following parameters: collimation, 0.5 mm; tube voltage, 120 kVp; tube current (volume exposure control), 100–500 mA; rotation time, 0.75 s; reconstruction section width, 0.5 mm; reconstruction interval, 0.5 mm; number of slices, 640; and voxel size, 0.43 mm × 0.43 mm × 0.5 mm. MRI was performed with a Signa 3.0-T system (GE Healthcare, Milwaukee, WI, USA). Imaging parameters for gadolinium-enhanced T1 weighted images were as follows: repetition time, 7.0 ms; echo time, 3.2 ms; slice thickness, 0.7 to 1.4 mm; field of view, 22 cm; matrix size, 352 × 256; flip angle, 13 degrees; voxel size, 0.63 mm × 0.85 mm × 0.7 mm; and timing, 4 min after contrast injection. The imaging parameters for Fast Imaging Employing Steady-State Acquisition (FIESTA) were as follows: repetition time, 5.6 ms; echo time, 2.7 ms; slice thickness, 0.4 to 0.8 mm; field of view, 20 cm; matrix size, 320 × 320; flip angle, 50 degrees; voxel size, 0.63 mm × 0.63 mm × 0.4 mm; and timing, just after contrast injection. Imaging parameters for 3D time-of-flight magnetic resonance angiography (TOFMRA) were as follows: repetition time, 26 ms; echo time, 2.7 ms; slice thickness, 0.5 to 1.0 mm; field of view, 20 cm; matrix size, 384 mm × 224; flip angle, 20 degrees; and voxel size, 0.63 mm × 0.89 mm × 0.5 mm.

We used an original five-point scale to evaluate the efficacy of the AR navigation system (Table 2). The two neurosurgeons who performed the endoscopic skull base surgery and the three senior residents who participated in ETS evaluated the efficacy of the AR navigation system after the surgery.

A 58-year-old man underwent ETS for a growing pituitary neuroendocrine tumor (PitNET). Preoperative MR imaging showed an intrasellar tumor with suprasellar extension (Figure 4a). After removal of bone of the sella turcica (Figure 4b) in ETS for a PitNET, AR navigation showed that the tumor extended into the right cavernous sinus and displaced the ICA laterally (Figure 4c). After additional bone removal from the right cavernous sinus (Figure 4d), the dura mater was incised, and the tumor was sufficiently exposed. The tumor was totally removed, and the cavernous segment of the right ICA was exposed (Figure 4e). The tumor was totally resected, and the patient had a good postoperative course (Figure 4f).

A 75-year-old woman had right visual disturbance due to recurrent meningioma. Preoperative MR imaging showed a tumor around the right optic nerve (Figure 6a). In a patient with recurrent tuberculum sellae meningioma, endoscopic images in the sphenoid sinus showed thick mucosa. After cutting the mucosa, a tumor-like structure was observed inferior to the right frontal lobe (white arrowhead) and medial to the right optic sheath (Figure 6b). There was a part of the tumor that was difficult to distinguish from the displaced optic nerve from the color or the consistency. However, the AR navigation indicated it as a “tumor” (Figure 6c,d), and we could resect this part with confidence (Figure 6e). The tumor was subtotally resected (Figure 6f), and the patient’s visual disturbance gradually improved.

In this report, we present the preliminary results of our novel AR navigation system in ETS. The average score overall indicated that the surgeons considered AR navigation to be more useful than conventional navigation for certain patients. On the other hand, in two cases it was evaluated that AR navigation was not as useful as conventional navigation, though by only one resident in each case. The main complaint involved perception of depth; there might be individual differences among surgeons regarding recognition of the virtual image as the ‘real’ situation, depending on their previous experience with similar cases in ETS.

In this study, we report the preliminary results of our novel AR navigation system during ETS for sellar and parasellar tumors. The resulting 3DCG images were identical to structures in the surgical field. Surgeons considered the present system more useful than the conventional navigation system for facilitating an immediate 3D understanding of the lesion and surrounding structures. Our AR navigation system facilitates intuitive comprehension of the skull base anatomy during ETS and helps not only young neurosurgeons in ETS for basic sellar lesions but also experienced neurosurgeons in ETS for complex skull base lesions. While further refinement of the image fusion techniques is essential, the preliminary results of this study suggest efficacy of our AR navigation in ETS for pituitary and parasellar tumors.