Advances in software technology have resulted in the ability to produce three-dimensional computer images of the facial skeleton and to manipulate these images to perform a “virtual” reconstruction of abnormal or missing parts. Physical models can be created from these data by special printers that use serial material deposition to construct a three-dimensional object as a stack of slices, a technique known as rapid prototype modeling.1 Rapid prototype modeling has mostly been used in reconstructive surgery to plan treatment for congenital and traumatic craniofacial deformities.2 Beyond surgical planning, however, rapid prototype models can serve as templates to bend titanium hardware used for bone flap fixation before surgery,3 and custom cutting guides can be manufactured using the same technology for making osteotomies to precisely match the virtual plan.4
Restoration of the midface is one of the most challenging and arduous reconstructive operations. Although prosthetic obturators and soft-tissue flaps are satisfactory for limited defects, extensive skeletal defects require vascularized bone flaps to adequately restore midfacial height, width, and projection, provide a firm surface for mastication, and accommodate osseointegrated implants for dental restoration. We have applied virtual planning and rapid prototype modeling to midfacial reconstruction using fibula free flaps in an effort to improve the accuracy and speed of the operation. In addition, we have used stereotactic navigation to ensure proper free flap positioning through limited incisions.5 We present our protocol for reconstruction of extensive midfacial skeletal defects using these technologies and report our outcomes.
MATERIALS AND METHODS
A three-dimensional computer model of the facial skeleton is first created using fine-cut (≤1.5 mm) computed tomographic data. Next, the reconstruction is planned virtually by osteotomizing a digitized fibula free flap to match the contours of the midfacial bones that are absent or will be resected. Care is taken to ensure that the virtually reconstructed maxillary occlusal plane aligns with the mandibular dentition.
Physical models based on the virtual reconstruction are then created using rapid prototype modeling technology (Figs. 1 and 2). Titanium plates are bent to match the contours of the fibula and adjoining facial bones on the model. Initially, models were created in-house, using a Z510 three-dimensional printer (Z Corp., Burlington, Mass.). After the first 15 cases, models were obtained from a third party (Medical Modeling, Inc., Golden, Colo.).
Custom cutting guides are also manufactured to assist in making the osteotomies needed to shape the fibula bone. During surgery, the cutting guide is temporarily fixed to the fibula bone using monocortical screws, and a reciprocating saw blade is inserted into slots in the cutting guide to make osteotomies at the precise angles and lengths required to replicate the virtually planned reconstruction (Figs. 3 and 4).
Commonly, the upper lip is split and a lateral rhinotomy is made to perform the resection (avoiding an infraorbital incision whenever possible); however, if the resection does not require these incisions, we perform the flap inset entirely transorally. Because insetting through relatively limited incisions can be challenging and proper positioning is critical, stereotactic navigation is used to confirm the position of the free flap.
An additional computed tomographic scan is obtained with the patient wearing adhesive fiducial markers on the forehead, which the patient is instructed to leave on until the operation. The computed tomographic data are imported into the navigation system workstation (VectorVision; BrainLab, Munich, Germany). Before the start of surgery, a registration process is performed, allowing the navigation system to integrate the position of the fiducial markers with spatial coordinates on the images loaded into the workstation. The fiducial markers are then replaced by a navigation array, which is fixed rigidly to the patient's calvaria and allows head movement during surgery without repeating the registration process.
Data files containing the virtual reconstruction are also loaded into the workstation. During surgery, the position of the actual fibula flap, indicated by the tip of a navigation probe, is compared with the intended position of the virtual fibula flap (Figs. 5 and 6). The flap is fixed rigidly once its position matches the virtual plan.
Midfacial reconstructions using the technique we describe were performed in 27 patients, including 12 performed immediately after resection and 15 that were delayed. Nine reconstructions were unilateral and 18 were bilateral. Vein grafts were needed in 12 cases. Nine patients received preoperative radiotherapy and six patients received postoperative radiotherapy.
There were no major complications, including no free flap losses. All patients regained 100 percent speech intelligibility after reconstruction as determined by a speech pathologist. Only one patient (who also had a soft palate resection) experienced hypernasal speech and occasional nasal regurgitation. Postoperatively, 19 patients consume a regular diet and eight patients consume a soft diet. All 16 patients who received osseointegrated implants consume a regular diet. Figure 7 depicts a representative result.
Our experience suggests that virtual planning and rapid prototype modeling are especially useful when the normal midfacial anatomy has been distorted by disease or when a prior resection has been performed. In such situations, plates cannot be bent along the contours of the patient's facial skeleton and it can be difficult to estimate the shape of the original midface. One drawback is the amount of time needed to plan the reconstruction and create the models. However, based on our experience, we feel that these efforts are worthwhile because they increase accuracy of the reconstruction and the efficiency of the surgery by helping to eliminate trial and error during the shaping of the bone flap. As mentioned, prebending plates and using cutting guides also save operative time. Similarly, we feel that the time needed to use stereotactic navigation is offset by the added confidence in the positioning of the free flap (which can be difficult to correct secondarily) and the ability to minimize making facial incisions for exposure.
We present a method based on virtual planning and creation of rapid prototype models to accurately reestablish midfacial anatomy using vascularized bone flaps while saving time by streamlining the operative procedure, allowing prebending of titanium plates needed for fixation, and simplifying osteotomies. Accurate flap positioning through limited incisions is made possible using stereotactic navigation.
The patient provided written consent for the use of his image.
This work was funded in part by a grant from the Kyte Foundation. The authors thank Andrew Christensen and Katie Weimer of Medical Modeling, Inc., for invaluable contributions to this study.
1.Saigal K, Winokur RS, Finden S, Taub D, Pribitkin E. Use of three-dimensional computerized tomography reconstruction in complex facial trauma. Facial Plast Surg
2.Chang PS, Parker TH, Patrick CW Jr, Miller MJ. The accuracy of stereolithography in planning craniofacial bone replacement. J Craniofac Surg
3.Kernan BT, Wimsatt JA III. Use of a stereolithography model for accurate, preoperative adaptation of a reconstruction plate. J Oral Maxillofac Surg
4.Hirsch DL, Garfein ES, Christensen AM, Weimer KA, Saddeh PB, Levine JP. Use of computer-aided design and computer-aided manufacturing to produce orthognathically ideal surgical outcomes: A paradigm shift in head and neck reconstruction. J Oral Maxillofac Surg
5.Strong EB, Rafii A, Holhweg-Majert B, Fuller SC, Metzger MC. Comparison of 3 optical navigation systems for computer-aided maxillofacial surgery. Arch Otolaryngol Head Neck Surg