[Robotics]

Image-to-patient registration by natural anatomical surfaces of the head

The use of registration markers in computer-assisted surgery is combined with high logistic costs and efforts.

During the preparation of image guided surgery, automated markerless patient-to-image registration based on anatomical surfaces allows a significant reduction of preoperative effort and of the radiation dose the patient is exposed to. Placement and measurement of radio-opaque fiducial markers becomes unnecessary. The usability of face, auricle, maxilla and mandible for surface-based registration to CT image data was investigated. The present study was performed to evaluate the clinical accuracy in finding defined target points within the surgical site after markerless patient registration in image-guided oral and maxillofacial surgery.

Preoperatively, the spatial position of 20 patients was registered to CT image data using a 3D laser surface scanner. Indications for surgery were tumours, foreign bodies and skeletal malformations. The accuracy of this surface-based registration was verified using additionally placed fiducial markers.

The study showed, that markerless surface-based registration was sufficiently accurate for clinical use when the surface used for matching was the upper jaw, the face, or — with reservations — the auricle. Surface-based registration using the mandible did not yield satisfying results. To conclude, image-to-patient registration based on laser surface scanning is a valuable method for surgery of the head. Multiple sites of the head were identified as appropriate for the method. Hence, dependent on the individual case and the intended surgery, the registration area can be selected with the necessary flexibility.

Image-guided surgery and medical robotics in the cranial area

Surgery in the cranial area includes complex anatomic situations with high-risk structures and high demands for functional and aesthetic results. Conventional surgery requires that the surgeon transfers complex anatomic and surgical planning information, using spatial sense and experience. The surgical procedure depends entirely on the manual skills of the operator. The development of image-guided surgery provides new revolutionary opportunities by integrating presurgical 3D imaging and intraoperative manipulation. Augmented reality, mechatronic surgical tools, and medical robotics may continue to progress in surgical instrumentation, and ultimately, surgical care. The aim of this article is to review and discuss state-of-the-art surgical navigation and medical robotics, image-to-patient registration, aspects of accuracy, and clinical applications for surgery in the cranial area.

Laser-scan-based navigation in cranio-maxillofacial surgery

BACKGROUND

In computer-assisted surgery, a correlation between a volume data set and the surgical site is required in order to localize the patient's head on the operating table. Registration markers are commonly used for this procedure. However, the marker registration is associated with high logistics, since the markers have to be placed prior to data set acquisition and have to be kept in their position until the patient enters the operating room. This study deals with a new markerless registration method in cranio-maxillofacial surgery that is based on a high-resolution laser-scan of the patient's (relaxed) skin surface.

PATIENTS

20 patients with tumours, bone malformations or foreign bodies, scheduled for computer-assisted surgery, were involved in the study.

STUDY DESIGN

The clinically applied accuracy of the laser-scan-based registration was measured through additionally placed registration markers. The inherent precision of the laser-scan registration system was controlled in phantom studies.

RESULTS

The clinically applied accuracy of the new laser-scan-based registration technique ranged between 0.2 and 1.8 mm with a mean deviation of 1.1mm and a standard deviation of 0.3 mm.

CONCLUSION

The facial skin surface can serve as a sufficiently stable and invariable reference base in order to register patients for computer-assisted cranio-maxillofacial surgery.

Neue Konzepte in der bildgestützten Chirurgie: automatische Patientenregistrierung anhand von Kiefer und Ohrmuschel

OBJECTIVE

Automatic and marker less patient registration based on natural anatomical interfaces may considerably reduce the radiation load and logistical input prior to computer-assisted surgical interventions, as it is not necessary to place and measure reference markers. The present study was to find out if, apart from the facial skin, also auricles as well as the upper and lower jaw can be used as anatomical interfaces for the intraoperative registration of the patient's position.

MATERIAL AND METHOD

Prior to surgical intervention the positions of 20 patients were registered by a high-resolution 3D laser scan and correlated with the preoperative CT data set. Tumors, skeletal malformations, and foreign bodies were indications for surgical intervention. Auricles as well as the upper and lower jaw were used to register the patient's positions. The accuracy of this basically marker less method was clinically evaluated through the additionally placed conventional registration markers.

RESULTS AND CONCLUSION

The marker less patient registration based on natural anatomical interfaces was successful in the upper jaw (deviation: 0.8+/-0.3 mm). The tongue and mobile floor of the mouth led to geometric incongruence and inadequate laser registration in the lower jaw. As far as the auricles were concerned, high accuracy could only be achieved as long as the auricles had not been deformed during CT imaging (deviation: 1.9+/-0.9 mm). The usual CT acquisition with a conventional head support, however, led to temporary auricular deformations in more than half of the patients, which made an exact laser scan registration impossible.