Assessing accuracy using different image guidance registration methods

Abstract

Problem: To determine the accuracy of image guidance technology using point registration and surface match registration protocols. Methods: This was a prospective case series and experimental study using a plastic model. Results: Image guidance technology is rapidly becoming commonplace and provides a significant improvement in the management of complex and revision surgeries. One of the potential areas where error can be induced in the use of image guidance is in the registration process. We studied 3 types of registration (point registration, point registration with surface matching, and a pure surface-matching protocol using a plastic model) subsequently in a series of cases. Accuracy was assessed at an assigned surface point and at a deep anatomic point. For the model, this point was represented by a marker at the center of the model, and for patients the Eustachian tube orifice was used as a landmark. Registration using individual points yielded accuracy that varied with the number of points used. The larger number of registration points yielded greatest accuracy and produced a large volume of registered space. Newer pure surface-matching protocols using an LED-covered mask yielded equivalent accuracy to best point-registered models. The use of 3 anchor points with subsequent surface matching was less accurate. The use of similar registration protocols in patients showed that surface matching yielded accuracies of 1 + 0.23 mm at the Eustachian tube with point registration yielding accuracies of 1 + 0.34 mm. Conclusion: Surface match registration provides an accurate registration even at deep points within the surgical target. Significance: Surface match registration using an LED mask may make the use of image guidance more time-efficient. Support: None reported.