Phantom-based analysis of the accuracy of a new registration method in magnetic navigation of the paranasal sinuses

Introduction:

This prospective study compared the accuracy of two different registration methods (Fiagon GmbH, Germany) in electromagnetic navigation of the frontal skull base. A newly developed automatic image registration method promises increased user friendliness without loss accuracy, but to date, no study exists comparing the new method with the classic approach of tactile surface registration.

Methods:

A head phantom was prepared with 27 markers in sagittal, axial and coronary planes and their reference coordinates were determined using a navigational CT (low dose, slice 0.6 mm). Subsequently 20 runs (n = 20) of automatic image registration and tactile surface registration, respectively, were conducted and the resulting marker coordinates were compared to the reference coordinates. So the target registration error (TRE) of the 27 markers was assessed. The TRE of both methods were compared using a 2-factor ANOVA with repeated measure.

Results:

Mean TRE using surface registration was 1.97 mm (SD 0,57), while the mean TRE of the same number of registrations using automatic image registration was 1.55 mm (SD 0,24). This range of TRE values is within the range of previous studies (1.5 – 2 mm). The differences between the two errors was found to be highly significant (p < 0.001).

Conclusions:

In electromagnetic navigation, the new automatic image registration is a promising method for registration of the paranasal sinuses with significantly smaller TRE in comparison to the classic approach of tactile surface registration.

Publication History

Publication Date:
18 April 2018 (online)