The mandible has an important role in appearance and function. The aim of this study was to describe and evaluate surgical navigation-assisted mandibular reconstruction with the fibula flap. Patients recruited into the study had a custom dental splint fabricated to maintain the mandible in a fixed position. Later, the computed tomography (CT) scan, preoperative design, and operation on the mandible were done in the same position. At 1 week after surgery, a CT scan was done to evaluate the repeatability between the preoperative design and the postoperative result. Twenty patients were enrolled in this study. Good repeatability between the postoperative CT and the preoperative design was found. The repeatability between the preoperative plan and postoperative outcome was 79.1 ± 8.6% at within 1 mm, 87.1 ± 6.7% at within 2 mm, and 91.9 ± 5.4% at within 3 mm. From this study, it can be concluded that surgical navigation techniques can precisely transfer the preoperative design to the operation in mandible reconstruction with a fibula flap. This will assist the surgeon in achieving good cosmetic and functional outcomes.
The mandible has an important role in facial harmony, mastication, speech, swallowing, support of the tongue base, and airway function. Jaw defects caused by a tumour or injury usually result in compromise of these functions, as well as facial aesthetics. The utility of the free fibula osteocutaneous flap for mandibular reconstruction was recognized and subsequently popularized by Hidalgo in 1989. The fibula flap is now used widely in mandibular reconstruction. Advantages include the ability to incorporate skin, muscle, and components of bicortical bone that can accurately reproduce the mandible. Comprehensive long-term follow-up studies assessing the cosmetic effect, speech, and deglutition have demonstrated satisfactory outcomes. Morbidity at the donor site has been described as ‘mild’.
Even though mandibular reconstruction with the fibula flap is now a common oral and maxillofacial surgical procedure, it is technically challenging and time-consuming. In addition to microvascular anastomoses, contouring of the flap must be precise to re-establish the form-and-function occlusal relationships of the mandible.
To improve precision and simplify the surgical procedure, computer-assisted surgery (CAS) involving preoperative virtual planning and computer-assisted intraoperative navigation has evolved. Computerized navigation was first indicated for the resection of tumours in the craniomaxillofacial skeleton. Based on identification of the tumour on computed tomography (CT) scans and the planned resection margins, the navigation probe is used intraoperatively to provide real-time feedback to the surgeon. Several studies have demonstrated the value of this technology in improving the precision with which tumours can be resected with safe margins while reducing the morbidity of unaffected tissues.
The application of computerized navigation in surgery to the mandible is complicated due to its mobility. Unlike the rest of the craniomaxillofacial skeleton (which acts as one solid structure), the mandible is an independently movable body, so its synchronization with the pre-acquired CT image is more difficult.
The aim of this study was to describe and evaluate a straightforward method for navigation-assisted mandibular reconstruction with a fibula flap.
Materials and methods
Ethical approval of the study protocol
The study protocol was approved by the institutional ethics committee. All patients provided written informed consent to participate in the study.
From November 2011 to September 2014, 20 patients with a lesion or defect of the mandible attending the study hospital were enrolled prospectively. Inclusion criteria were the following: partial resection of the mandible was indicated; reconstruction with a fibula graft was possible; the patient could wait 3–5 days for a design to be created; the patient agreed to the surgical team using a computer-assisted navigation method. Exclusion criteria were the following: operation time had to be controlled because of the general status of the patient; the patient could not wait 3–5 days for a design to be created; the patient had an advanced malignant tumour with a poor prognosis; mouth opening was limited, so making a dental splint was difficult. A preoperative incisional biopsy was done in order to obtain a pathological diagnosis. Reconstruction was indicated for all patients.
All patients underwent preoperative spiral CT (helix with a slice thickness of 1.25 mm), which was repeated at 1 week after surgery (BrightSpeed 16; GE Healthcare, Buckinghamshire, UK). CT data were processed and transferred to SurgiCase CMF version 5.0 (Materialise, Leuven, Belgium) and iPlan CMF (Brainlab, Feldkirchen, Germany) via Digital Imaging and Communications in Medicine (DICOM) files for preoperative surgical planning and postoperative evaluation. The VectorVision system (Brainlab) was used for surgical navigation.
A custom-made dental splint was fabricated. Once fitted, the mandible could be maintained in a fixed position. A high-resolution CT of the skull and fibula region was taken with the dental splint in situ. Two-dimensional CT data were saved in DICOM format. DICOM data were processed in SurgiCase CMF version 5.0 to generate accurate three-dimensional rendering of the bone contours. Virtual segmentation of the skull was undertaken to separate the mandible from the remaining parts of the skull.
Following standard principles, tumour resection was planned and the precise defect defined. Virtual design of the fibula flap involved segmentation, length of segments, angulation, and orientation to achieve the ‘ideal’ contour of the ‘neo-mandible’. If distortion of the mandible due to the tumour was evident, the ‘mirror’ function was used to replicate the corresponding contralateral anatomical structure. If deformation was bilateral, the neo-mandible was guided by the position of maxillary teeth and the underlying skeletal base.
Surface data for all parts were exported into stereolithography (STL) files. Three-dimensional cylindrical-shaped objects in STL format were placed in a triangular pattern for use as reference points ( Fig. 1 ). These were placed on the unaffected mandible on the facial aspect on both sides of the anticipated ostectomy. Cylinders were 3 mm in height and 2 mm in diameter. The three cylindrical-shaped objects were separated ≥1.5 cm to facilitate identification.
STL data of segments with surface markers were imported into iPlan CMF, and merged into one object as the surface marker-assisted navigation plan.
Intraoperative navigation was undertaken using VectorVision (Brainlab). Preoperative preparation included matching the STL and CT in advance on iPlan (Brainlab) and exporting the file to the navigation workstation.
Under general anaesthesia, a reference frame with three light-reflecting spheres was fixed rigidly to the skull. Registration was completed through facial-surface imaging with the infrared ray (IR) emitter and IR receiver, as per the VectorVision protocol. Registration accuracies of the surgical area were verified automatically by the software, and registration errors were <0.7 mm in all cases. Also, surgeons could verify the actual surgical process on the virtual plan.
Intermaxillary fixation was applied with a custom-made dental splint in situ ( Fig. 2 a). Under the navigation plan, all surface markers were located by the tip of a navigation probe. Holes were drilled in the mandible before osteotomy ( Fig. 2 b). The line of the mandibulectomy was guided by the navigation system ( Fig. 3 ).
A second surgical team concomitantly harvested the fibula flap. Segmentation and bevelling of the segments was completed according to the preoperative design in SurgiCase CMF. The fibula flap was transferred to the mandible. Before fixation of the fibula, the mandibular position was checked through the drilled holes over the mandible using the tip of the navigation probe. While insetting the fibula bone into the defect, the tip of a navigation probe was used to identify the inferior border and posterior border of the fibula and residual mandible ( Fig. 4 ). After appropriate placement of the fibula bone in the intended position, rigid fixation with miniplates was carried out.
CT was done 1 week after surgery to assess the accuracy of the reconstruction. Three-dimensional rendering of bone contours was reconstructed from the postoperative CT and exported into STL files. Preoperative and postoperative data were imported into Geomagic Studio 2012 (3D Systems, Valencia, CA, USA). Manual and global registration functions were used to match the non-surgical parts of the two models. ‘3D comparison’ was applied to illustrate the deviation from the preoperative planning in a deviation spectrum. The resulting error grade colour map provided a direct impression of the match between the preoperative plan and postoperative results ( Fig. 5 ).
Demographic data and clinical characteristics of the 20 patients (13 women and seven men, mean age 33 years (range 22–54 years)) are shown in Table 1 . Among these patients, there were five with secondary mandibular defects, 10 with benign disease, and five with malignant disease.
|No.||Age, years||Sex||Diagnosis||Defect||Fibula segments||Repeatability within 1 mm (%) a||Repeatability within 2 mm (%) a||Repeatability within 3 mm (%) a|
|1||48||F||Ossifying fibroma||Left body||1||83.6||94.7||97.9|
|2||54||F||Carcinoma of floor of mouth||Left body and mental area||3||66.8||83.2||92|
|3||34||F||Sarcoma||Right ramus and body||2||86||93||96.5|
|4||25||M||Secondary defect||Right ramus and body||2||86.3||92||94.1|
|5||29||F||Sarcoma||Right ramus and body||3||91.9||95.9||98|
|6||26||F||Ameloblastoma||Right ramus and body||2 + 1 FBG||88.3||93.1||96.3|
|7||53||F||Keratocyst||Right body and mental area||2||54.7||66.4||74.4|
|8||48||F||Ameloblastoma||Left ramus and body||3 with double barrel||86.2||89.9||93|
|9||26||M||Ameloblastoma||Right body and mental area||2||80||90.6||95|
|10||50||F||Ameloblastoma||Right ramus and body||2||84.6||92||93.8|
|11||53||M||Secondary defect||Left ramus and body||2 + 1 FBG||78.6||87.1||91.6|
|12||22||M||Secondary defect||Right ramus and body||2 + 2 FBG||83.1||88.6||92.8|
|13||24||M||Secondary defect||Left ramus and body||3 + 2 FBG||73.6||80.8||87.1|
|14||50||F||Ameloblastoma||Left ramus and body||2 + 1 FBG||77.2||84.1||89.7|
|15||26||F||Ameloblastoma||Right ramus and body||3 + 1 FBG||76.8||83.7||91.2|
|16||23||M||Secondary defect||Right ramus and body||2 + 1 FBG||83||90.5||94.2|
|17||38||F||Ameloblastoma||Left ramus, body, and mental area||3||77.9||88||94.5|
|18||35||F||Sarcoma||Left body and mental area||2 + 2 FBG||68.6||85.8||93.9|
|19||43||M||Mucoepidermoid carcinoma||Right ramus and body||3||75.1||79.6||84.5|
|20||26||F||Ameloblastoma||Left ramus and body||3||78||83.9||87.5|