High-resolution laser surface scanning for patient registration in cranial computer-assisted surgery.

OBJECTIVE

Markerless patient registration is a new procedure that may reduce logistical efforts and possibly also the radiation load on the patients prior to a computer-assisted intervention. Congruent surfaces, such as bone surfaces or skin surfaces, represented in a data set and in the surgical site, can be overlapped using surface-matching. Previous studies describing this kind of markerless registration, however, show inaccuracies of up to 10 mm during computer-assisted navigation. Furthermore, these systems use less than 1000 surface points of the soft tissue surface in order to establish a correlation between the patient and the data set. Previous papers did not answer the question whether it is this scanning resolution that induces these inaccuracies in registration or rather intraoperative skin deformations.

STUDY DESIGN

In the present study therefore a new navigation system (SSN++) was used which is able to register up to 180,000 surface points of the surgical site. SSN++ is an infrared navigation system enlarged by a Minolta VI 900 3D volume digitizer. Three different kinds of laser scan-resolution were used for data correlation. An additional congruence analysis was performed in order to assess the geometry of the matched skin surfaces. 22 patients suffering from different cranial diseases (tumors, bony malformations, foreign bodies) were prepared for a computer-assisted intervention. Intraoral titanium-markers, rigidly fixed on the patients by a maxillary splint, were placed as targets while the CT data sets were made. These targets were - after markerless laser scan registration of the patients - supposed to serve for validating the new high-resolution navigation system SSN++.

RESULTS

The accuracy of markerless laser scan registration depends on the intraoperative laser scan's resolution. A high accuracy of the data correlation can be achieved if the number of the laser scan cloud points is about the same as the number of voxels of the corresponding surface on the CT data set. A reduction of the laser scan cloud points to less than 10 % compared to the number of voxels of the CT surface, however, leads to a significant loss of accuracy after markerless patient registration.

CONCLUSION

The markerless laser scan registration of the surgical site may achieve the same accuracy as a patient registration made by rigidly fixed titanium screws (mean accuracy: 1.2 mm) as long as a high-resolution laser scan is being used.