Accuracy validation in a cadaver model of cranial neuronavigation using a surface autoregistration mask

Image guidance systems are widely used in neurosurgical practice.

To compare the operational accuracy of a neuronavigation system when registration was accomplished with a commercially available surface-based autoregistration system vs other fiducial-based registrations.

We evaluated the operational accuracy of different registration methods in 20 cadaveric heads. Every specimen was prepared with 10 titanium microscrews functioning as external/internal targets and as bone fiducials. Six scalp fiducials were also affixed to each specimen that was registered with bone, scalp fiducials, and the autoregistration mask. The coordinates of all the target points were measured, first manually on the screen of the navigation system and then by touching the head of the implanted screw on the specimen. The difference between the real and virtual coordinates was calculated.

Means of the differences for external anterior targets were 1.96, 3.12, and 3.20 mm and 1.95, 3.24, and 3.19 mm for external posterior targets for the bone fiducials, adhesive fiducials, and autoregistration mask, respectively. Means of the differences for internal anterior targets were 2.60, 3.65, and 2.16 mm and 2.91, 3.83, and 2.41 mm for internal posterior targets for the bone fiducials, adhesive fiducials, and autoregistration mask, respectively.

Bone fiducial registration is associated with a statistically greater operational accuracy than scalp adhesive fiducials and the autoregistration mask in reaching anterior and posterior external targets (P < .001). Registration accomplished with the autoregistration mask is associated with a statistically greater operational accuracy in reaching internal targets than adhesive fiducials registration (P < .001) or bone fiducials registration (P < .05 and P < .01 for anterior and posterior targets, respectively).