CRANIOMAXILLOFACIAL DEFORMITIES/COSMETIC SURGERY
Patient-Specific Polyetheretherketone Facial Implants in a Computer-Aided Planning Workflow Godoberto Guevara-Rojas, MSc,* Michael Figl, MSc, DSc,y Kurt Schicho, MD, DSc, PhD,z Rudolf Seemann, MSc, MD, DMD, MBA, PhD,x Hannes Traxler, MD,k Apostolos Vacariu, MD,{ Claus-Christian Carbon, MA, PhD,# Rolf Ewers, MD, DMD, PhD,** and Franz Watzinger, MD, DMD, PhDyy Purpose:
In the present study, we report an innovative workflow using polyetheretherketone (PEEK) patient-specific implants for esthetic corrections in the facial region through onlay grafting. The planning includes implant design according to virtual osteotomy and generation of a subtraction volume. The implant design was refined by stepwise changing the implant geometry according to soft tissue simulations.
Materials and Methods:
One patient was scanned using computed tomography. PEEK implants were interactively designed and manufactured using rapid prototyping techniques. Positioning intraoperatively was assisted by computer-aided navigation. Two months after surgery, a 3-dimensional surface model of the patient’s face was generated using photogrammetry. Finally, the Hausdorff distance calculation was used to quantify the overall error, encompassing the failures in soft tissue simulation and implantation.
Results:
The implant positioning process during surgery was satisfactory. The simulated soft tissue surface and the photogrammetry scan of the patient showed a high correspondence, especially where the skin covered the implants. The mean total error (Hausdorff distance) was 0.81 1.00 mm (median 0.48, interquartile range 1.11). The spatial deviation remained less than 0.7 mm for the vast majority of points. Conclusions: The proposed workflow provides a complete computer-aided design, computer-aided manufacturing, and computer-aided surgery chain for implant design, allowing for soft tissue simulation,
yyHead of the Department of Cranio-Maxillofacial and Oral
*PhD Student, Facial Esthetics Engineering Group, University Hospital of Cranio-Maxillofacial and Oral Surgery, Medical
Surgery, St. P€ olten, Austria.
University of Vienna, Vienna, Austria.
The present study was supported by the Austrian Research Pro-
yAssistant Professor, Center of Physics and Biomedical
motion Agency FFG (grant 818041B1), proposed and managed by
Engineering, Medical University of Vienna, Vienna, Austria.
Prof DDr Schicho.
zAssociate Professor, Head of Facial Esthetics Engineering Group,
Guevara-Rojas and Dr Watzinger contributed equally to the pre-
University Hospital of Cranio-Maxillofacial and Oral Surgery, Medical
sent study, with Guevara-Rojas responsible for the technical part
University of Vienna, Vienna, Austria. xAssistant Professor, Facial Esthetics Engineering Group,
and Dr Watzinger for the medical aspects. Address correspondence and reprint requests to Dr Figl: Center
University Hospital of Cranio-Maxillofacial and Oral Surgery,
of Physics and Biomedical Engineering, Medical University of
Medical University of Vienna, Vienna, Austria.
Vienna, Waehringer Guertel 18-20, Vienna 1090, Austria; e-mail:
kAssistant Professor, Center of Anatomy and Cell Biology, Medical
[email protected]
University of Vienna, Vienna, Austria.
Received October 28 2013
{Teaching Assistant, Center of Anatomy and Cell Biology, Medical
Accepted February 9 2014
University of Vienna, Vienna, Austria.
Ó 2014 American Association of Oral and Maxillofacial Surgeons
#Professor, Otto Friedrich Universit€at Bamberg, Bamberg, Germany.
0278-2391/14/00175-X$36.00/0 http://dx.doi.org/10.1016/j.joms.2014.02.013
**Professor Emeritus, Facial Esthetics Engineering Group, University Hospital of Cranio-Maxillofacial and Oral Surgery, Medical University of Vienna, Vienna, Austria.
1
2
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
fabrication of patient-specific implants, and image-guided surgery to position the implants. Much of the surgical complexity resulting from osteotomies of the zygoma, chin, or mandibular angle might be transferred into the planning phase of patient-specific implants. Ó 2014 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg -:1-12, 2014 Medical rapid prototyping is gaining significance in different areas of preoperative planning such as maxillofacial surgery, orthopedics, neurosurgery, and or-
thognathic surgery. These 3-dimensional (3D) models allow the surgeon to become acquainted with the local anatomy and support the surgeon’s intraoperative ‘‘3D
FIGURE 1. A-C, Preoperative 3-dimensional photograms of the 27-year-old female patient. The midface hypoplasia is clearly visible. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
3
GUEVARA-ROJAS ET AL
FIGURE 2. A-D, Concept of our planning approach. Virtual Le Fort III osteotomy for achievement of the desired outcome according to soft tissue prediction and subtraction of the original (ie, actual preoperative) situation provides an initial definition of the polyetheretherketone patientspecific implant design. (Fig 2 continued on next page.) Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
imagination.’’1,2 The surgeon has the possibility to plan the osteotomies and, if necessary, to bend the fixation plates onto the model before surgery. Thus, it is possible to reduce the operative time and increase the accuracy. Modern software such as Mimics (Materialise, Leuven, Belgium) allows the planning of osteotomies or distractions using the computed tomography (CT) data from the patients. In many cases, it will be necessary to repair defects caused by trauma or tumor resection. For these purposes, standardized implants are available that can be individually adapted to the defect during surgery. This can, in some cases, increase the operative time. Also, the results will also strongly depend on the surgeon’s experience and skills. Recent developments such as patient-specific implants (PSIs)3 do not require adaptation of the implant’s shape and geometry to the patient’s anatomy during surgery. Instead, computer-
aided implant design using the CT data of the patient is performed. The PSIs are designed to fit precisely in the patient’s defects or malformations. After having finalized the computer-aided implant design, the implant shape can be controlled visually and, if needed, modified in the course of an iterative process, using a rapid prototyping model of the implant combined with the patient’s anatomy to control and optimize the shape and fit of the implant on the bone. The material polyetheretherketone (PEEK) has been used in several medical applications such as cranial vault reconstruction.4,5 Patient-specific alloplastic implants have reduced both the need for major manipulation during surgery and the actual operative time. In preparation for calvarial defect reconstruction, a preoperative 3D-CT scan should be performed and the images sent to the implant manufacturer. An anatomically correct skull
4
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 2 (cont’d). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
model and an implant are built using rapid prototyping and sent to the physician for review and approval. The company then delivers the definitive PEEK-PSI to the physician who will perform the implant procedure.6 Reports have been published on the reconstruction of complex orbital frontotemporal reconstruction using PEEK-PSIs.6 PEEK polymers have been used in spinal surgery and orthopedic surgery and have been shown to be a highly reliable material with advantageous characteristics.7-9 On follow-up radiographs and CT imaging, translucency without artifacts has been observed.6,10 The advantageous radiolucent material characteristics of PEEK have made it a widely accepted alternative to metallic materials for spinal implants.7 Furthermore, it is compatible with magnetic resonance imaging, because it causes no magnetic interference. In addition, PEEK polymers have excellent chemical resistance, and the degree of elasticity or stiffness can be modified to fit the situ-
ation. The main drawback has been the possibility of postoperative infection.6 In the present study, we report on a new workflow using PEEK-PSIs for esthetic corrections in the facial region. The planning includes implant design according to the specifications derived from 3D simulation of soft tissue behavior.
Materials and Methods The present study followed the Declaration of Helsinki on medical protocol, and the ethics and ethical committee of the Medical University of Vienna (approval no. EK 665/2008) approved the study. PATIENT’S CASE
The present patient was female, 27 years old, and had severe congenital midface hypoplasia. Figure 1 shows preoperative 3D photograms of the patient.
5
GUEVARA-ROJAS ET AL
FIGURE 3. Iterations of osteotomy movements and simulation of soft tissue changes. A,B, Axial images with 0-mm bone segment reposition. C,D, Axial images showing bone segment reposition (3 mm anterior) and respective soft tissue changes. E,F, Axial images showing bone segment reposition (5 mm anterior) and respective soft tissue changes. (Fig 3 continued on next page.) Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
We decided to offer her treatment using PSIs for augmentation of the zygomatic prominence. The complete workflow included the following steps:
1. Computer-assisted treatment planning with soft tissue prediction and manufacturing of the PEEK implants using an interactive, iterative teleplanning process;
6
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 3 (cont’d). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
2. Surgery, with navigation-assisted insertion of the PEEK implants; 3. Evaluation of the concept (quantitative comparison of the soft tissue prediction with postoperative 3D photogrammetry).
COMPUTER-ASSISTED PLANNING
The planning began with importing the CT data of the skull (Philips Brilliance 64, Amsterdam, Holland) into the planning software, Mimics, version
7
GUEVARA-ROJAS ET AL
FIGURE 4. ‘‘Expressiveness’’ is an important measure of the quality of a simulation, especially in esthetic surgery. A, Reconstruction from computed tomography data provided precise information on the shape and geometry of the facial surface for the surgeon, but were obviously not sufficiently illustrative to inform the patient, primarily because the hair, eyebrows, and eyelashes were not visualized. B, Merging of a (conventional) digital photograph with the 3-dimensional reconstruction from the computed tomography data clearly enhanced the ‘‘natural and realistic’’ look, but lacked reliability and precision owing to distortions of the surface geometry from the matching method. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
14.0 (Materialise). The CT data were accessed using a secure ftp server for the rapid prototyping company (Synthes, Mesocco, Switzerland). Because no occlusal disturbance was present, maxillary advancement was
subtracted from the virtual Le Fort III advancement procedure. Consequently, isolated advancement of the zygoma resulted, which was the goal of the planning procedure (Fig 2). The subtraction volume of
FIGURE 5. Intraoperative view. Left, Electromagnetic patient reference frame fixed on the calvaria using microscrews, and Right, patientspecific implant in situ. Reliable and precise intraoperative positioning of the implant can be achieved with navigation. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
8
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 6. Screenshot of the soft tissue simulation model with color coding for the Hausdorff distance, evaluated using the software MESH. Left, Color coded (‘‘cold to warm’’) point to surface distances are shown. Far left, Perceptible within the histogram, almost all the distances remained less than 0.7 mm. Right, a rendering of the simulation planning CT is shown. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
the primary anatomy and the planned final situation according to the virtual Le Fort osteotomy gave the initial geometry of the implant. The realization of this concept has been illustrated in Figure 2. Fur-
ther refinement, including edge removal and contour smoothing, was achieved with mesh distortion tools and vertex removal. Different simulations of the osteotomies and bone repositioning were used for soft tissue simulation with Mimics (Materialise) (Fig 3), and evaluated by us using an interactive teleplanning process. Finally, a small implant was manufactured by the rapid prototyping department of Synthes. The ‘‘expressiveness’’ of the simulation is a crucial goal. Pure 3D reconstruction from CT data cannot sufficiently provide a realistic impression, especially for the patient (Fig 4). NAVIGATION-ASSISTED SURGERY
FIGURE 7. Histogram showing the point to surface distances of the points from the first to the second surface. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
A coronal approach was used to avoid visible scars. Two PEEK onlay implants were used to augment both zygomatic prominences. The microscrew positions were predefined by a hole through the PSIs during the rapid prototyping process. Drilling in the bone for fixation screws was navigated using Fusion ENT Navigation (Medtronic, Minneapolis, MN), a customary computer-aided surgery (CAS) system using electromagnetic tracking technology. The patient reference frame
9
GUEVARA-ROJAS ET AL
FIGURE 8. Preoperative and postoperative photogrammetry images of the patient, showing the esthetic benefit. The postoperative photographs were taken 2 months after surgery. A,C,E, Preoperative photogrammetry images of the patient. B,D,F, Postoperative photogrammetry images of the patient. (Fig 8 continued on next page.) Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
was fixed to the calvaria using microscrews (Fig 5). A conventional registration procedure was performed on the basis of anatomic landmarks. POSTOPERATIVE FOLLOW-UP AND PHOTOGRAMMETRY
The patient’s postoperative face was scanned using a 3D photogrammetry system (Dimensional Imaging,
Glasgow, Scotland, UK) 2 months after surgery. To evaluate the simulation error using MESH, version 1.13 (open source11), the photogrammetric surface data set was first aligned with the simulated soft tissue data by surface registration using the planning software Mimics, version 14.0 (Materialise) on a defined region of interest outside the augmented area. In addition, another region of interest was defined in the
10
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 8 (cont’d). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
zygomatic area, and the Hausdorff distances were computed to ascertain the quantitative measure of planning error.
Results Our experience has shown the feasibility for routine application using standard instruments. The duration of the surgical intervention was 4 hours, including the approach and wound closure. No complications were observed during or after surgery. The patient expressed high satisfaction. Three weeks postoperatively the soft tissue swelling had resolved. At 2 months after surgery, a photogrammetry scan was performed. The simulated soft tissue surface and the photogrammetry scan of the patient showed a high correspondence, especially of the skin covering the implants. The mean total error (Hausdorff distance) was 0.81 1.00 mm (median 0.48, interquartile range 1.11). Figure 6 shows the soft tissue simulation model with color-coding for the Hausdorff distance. A histogram of the latter (Fig 7) indicated that the spatial deviation (ie, surface distance) remained at less than 0.7 mm for the vast majority of points. Outliers occurred only in the eyeball region; these were closed in the soft tissue simulation and opened during acquisition of the photogrammetric data (ie, red areas). The patient’s pre- and postoperative photogrammetry images are shown in Figure 8.
Discussion The present study evaluated a workflow for computer-aided design and manufacturing (CAD-
CAM) fabrication and CAS insertion of the patientspecific implants. Although this workflow included different software packages and several steps, the evaluation of the soft tissue simulation revealed good correspondence with the postoperative appearance. The accuracies observed in the present study remained reliably within the range of accuracy known from published studies of computer-assisted surgery.12-14 A high correspondence between the predicted facial surface geometry and the actual surgical outcome was achieved within the competence of the Mimics software (Materialise). The interactive, iterative teleplanning workflow proved feasible for efficient preparation of the surgical intervention. In particular, the integration of 3D photogrammetry contributed significantly to the ‘‘expressiveness’’ of the simulation (Figs 1, 4). However, the present workflow has one important shortcoming: the Mimics package allows for only soft tissue simulations associated with osteotomies. In the concept we have presented, the PSI design was based on the subtraction of the original data set from the simulated optimal surgical plan. Consequently, some deviations between simulation and outcome were inevitably caused by the method itself and could not be eliminated (ie, systematic error). Differences between subtraction and the final PSI implant can be seen in Figure 9. These images show that the geometries of the bone segment and implant differed mainly at the edges. Nevertheless, the remaining shape of the PSI and the bone were very similar. Therefore, we should always expect minor deviations in the soft tissue simulation based on osteotomies, although these deviations will be confined to the edges.
11
GUEVARA-ROJAS ET AL
FIGURE 9. A, Axial and B, coronal images showing that the geometries of the bone segment and implant differed mainly at the edges (green contour indicates the bone segment; red contour, the implant). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
Computer-designed alloplastic implants could represent a milestone in the evolution of planning and realization of 3D reconstruction during surgery. They have the potential to become a reliable and irreplaceable part of the surgical armamentarium.6 The present report covers the first clinical case of maxillofacial reconstruction using computer-designed PEEKPSI bilateral augmentation. In conclusion, the present study has shown that the proposed workflow provides a complete CAD-CAMCAS chain for implant design that allows for soft tissue simulation, fabrication of patient-specific implants, and image-guided surgery to position the implants.
Much of the surgical complexity resulting from osteotomies of the zygoma, chin, or mandibular angle might be transferred to the planning phase of patient-specific implants. Furthermore, our method was resource intensive because it was a process development. In routine practice, with support of telemedicine (interactive teleplanning), the planning process could be transferred to a specialized center (commercial company), and the physician’s time would be freed. Acknowledgments We acknowledge the active participation of Synthes, who provided engineering knowledge and implant material.
12
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
References 1. Mavili ME, Canter HI, Saglam-Aydinatay B, et al: Use of threedimensional medical modeling methods for precise planning of orthognathic surgery. J Craniofac Surg 18:740, 2007 2. Papadopoulos MA, Christou PK, Christou PK, et al: Threedimensional craniofacial reconstruction imaging. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 93:382, 2002 3. Patient Specific Implants. 2012. Available at: http://www. synthes.com/sites/intl/Products/CMF/Cranial/Pages/PatientSpecific-Implants.aspx. Accessed April 11, 2012 4. Chacon-Moya E, Gallegos-Hernandez JF, Pina-Cabrales S, et al: Cranial vault reconstruction using computer-designed polyetheretherketone (PEEK) implant: Case report. Cir Cir 77:437, 2009 5. Hanasono MM, Goel N, DeMonte F: Calvarial reconstruction with polyetheretherketone implants. Ann Plast Surg 62:653, 2009 6. Scolozzi P, Martinez A, Jaques B: Complex orbito-frontotemporal reconstruction using computer-designed PEEK implant. J Craniofac Surg 18:224, 2007 7. Kurtz SM, Devine JN: PEEK biomaterials in trauma, orthopedic, and spinal implants. Biomaterials 28:4845, 2007 8. Moon SM, Ingalhalikar A, Highsmith JM, Vaccaro AR: Biomechanical rigidity of an all-polyetheretherketone anterior thoracolum-
9.
10.
11.
12.
13.
14.
bar spinal reconstruction construct: An in vitro corpectomy model. Spine J 9:330, 2009 Walter J, Kuhn SA, Reichart R, et al: PEEK cages as a potential alternative in the treatment of cervical spondylodiscitis: A preliminary report on a patient series. Eur Spine J 19:1004, 2010 Citak M, Kendoff D, Wanich T, et al: The influence of metal artifacts on navigation and the reduction of artifacts by the use of polyether-ether-ketone. Comput Aided Surg 13:233, 2008 Aspert N, Santa-Cruz D, Ebrahimi T: MESH: Measuring errors between surfaces using the Hausdorff distance. Multimedia and Expo, 2002 ICME ’02 Proceedings 2002. IEEE International Conference 1:705, 2002 Klug C, Schicho K, Ploder O, et al: Point-to-point computerassisted navigation for precise transfer of planned zygoma osteotomies from the stereolithographic model into reality. J Oral Maxillofac Surg 64:550, 2006 Wagner A, Schicho K, Birkfellner W, et al: Quantitative analysis of factors affecting intraoperative precision and stability of optoelectronic and electromagnetic tracking systems. Med Phys 29:905, 2002 West JB, Fitzpatrick JM, Toms SA, et al: Fiducial point placement and the accuracy of point-based, rigid body registration. Neurosurgery 48:810, 2001
Patient-Specific Polyetheretherketone Facial Implants in a Computer-Aided Planning Workflow Godoberto Guevara-Rojas, MSc,* Michael Figl, MSc, DSc,y Kurt Schicho, MD, DSc, PhD,z Rudolf Seemann, MSc, MD, DMD, MBA, PhD,x Hannes Traxler, MD,k Apostolos Vacariu, MD,{ Claus-Christian Carbon, MA, PhD,# Rolf Ewers, MD, DMD, PhD,** and Franz Watzinger, MD, DMD, PhDyy Purpose:
In the present study, we report an innovative workflow using polyetheretherketone (PEEK) patient-specific implants for esthetic corrections in the facial region through onlay grafting. The planning includes implant design according to virtual osteotomy and generation of a subtraction volume. The implant design was refined by stepwise changing the implant geometry according to soft tissue simulations.
Materials and Methods:
One patient was scanned using computed tomography. PEEK implants were interactively designed and manufactured using rapid prototyping techniques. Positioning intraoperatively was assisted by computer-aided navigation. Two months after surgery, a 3-dimensional surface model of the patient’s face was generated using photogrammetry. Finally, the Hausdorff distance calculation was used to quantify the overall error, encompassing the failures in soft tissue simulation and implantation.
Results:
The implant positioning process during surgery was satisfactory. The simulated soft tissue surface and the photogrammetry scan of the patient showed a high correspondence, especially where the skin covered the implants. The mean total error (Hausdorff distance) was 0.81 1.00 mm (median 0.48, interquartile range 1.11). The spatial deviation remained less than 0.7 mm for the vast majority of points. Conclusions: The proposed workflow provides a complete computer-aided design, computer-aided manufacturing, and computer-aided surgery chain for implant design, allowing for soft tissue simulation,
yyHead of the Department of Cranio-Maxillofacial and Oral
*PhD Student, Facial Esthetics Engineering Group, University Hospital of Cranio-Maxillofacial and Oral Surgery, Medical
Surgery, St. P€ olten, Austria.
University of Vienna, Vienna, Austria.
The present study was supported by the Austrian Research Pro-
yAssistant Professor, Center of Physics and Biomedical
motion Agency FFG (grant 818041B1), proposed and managed by
Engineering, Medical University of Vienna, Vienna, Austria.
Prof DDr Schicho.
zAssociate Professor, Head of Facial Esthetics Engineering Group,
Guevara-Rojas and Dr Watzinger contributed equally to the pre-
University Hospital of Cranio-Maxillofacial and Oral Surgery, Medical
sent study, with Guevara-Rojas responsible for the technical part
University of Vienna, Vienna, Austria. xAssistant Professor, Facial Esthetics Engineering Group,
and Dr Watzinger for the medical aspects. Address correspondence and reprint requests to Dr Figl: Center
University Hospital of Cranio-Maxillofacial and Oral Surgery,
of Physics and Biomedical Engineering, Medical University of
Medical University of Vienna, Vienna, Austria.
Vienna, Waehringer Guertel 18-20, Vienna 1090, Austria; e-mail:
kAssistant Professor, Center of Anatomy and Cell Biology, Medical
[email protected]
University of Vienna, Vienna, Austria.
Received October 28 2013
{Teaching Assistant, Center of Anatomy and Cell Biology, Medical
Accepted February 9 2014
University of Vienna, Vienna, Austria.
Ó 2014 American Association of Oral and Maxillofacial Surgeons
#Professor, Otto Friedrich Universit€at Bamberg, Bamberg, Germany.
0278-2391/14/00175-X$36.00/0 http://dx.doi.org/10.1016/j.joms.2014.02.013
**Professor Emeritus, Facial Esthetics Engineering Group, University Hospital of Cranio-Maxillofacial and Oral Surgery, Medical University of Vienna, Vienna, Austria.
1
2
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
fabrication of patient-specific implants, and image-guided surgery to position the implants. Much of the surgical complexity resulting from osteotomies of the zygoma, chin, or mandibular angle might be transferred into the planning phase of patient-specific implants. Ó 2014 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg -:1-12, 2014 Medical rapid prototyping is gaining significance in different areas of preoperative planning such as maxillofacial surgery, orthopedics, neurosurgery, and or-
thognathic surgery. These 3-dimensional (3D) models allow the surgeon to become acquainted with the local anatomy and support the surgeon’s intraoperative ‘‘3D
FIGURE 1. A-C, Preoperative 3-dimensional photograms of the 27-year-old female patient. The midface hypoplasia is clearly visible. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
3
GUEVARA-ROJAS ET AL
FIGURE 2. A-D, Concept of our planning approach. Virtual Le Fort III osteotomy for achievement of the desired outcome according to soft tissue prediction and subtraction of the original (ie, actual preoperative) situation provides an initial definition of the polyetheretherketone patientspecific implant design. (Fig 2 continued on next page.) Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
imagination.’’1,2 The surgeon has the possibility to plan the osteotomies and, if necessary, to bend the fixation plates onto the model before surgery. Thus, it is possible to reduce the operative time and increase the accuracy. Modern software such as Mimics (Materialise, Leuven, Belgium) allows the planning of osteotomies or distractions using the computed tomography (CT) data from the patients. In many cases, it will be necessary to repair defects caused by trauma or tumor resection. For these purposes, standardized implants are available that can be individually adapted to the defect during surgery. This can, in some cases, increase the operative time. Also, the results will also strongly depend on the surgeon’s experience and skills. Recent developments such as patient-specific implants (PSIs)3 do not require adaptation of the implant’s shape and geometry to the patient’s anatomy during surgery. Instead, computer-
aided implant design using the CT data of the patient is performed. The PSIs are designed to fit precisely in the patient’s defects or malformations. After having finalized the computer-aided implant design, the implant shape can be controlled visually and, if needed, modified in the course of an iterative process, using a rapid prototyping model of the implant combined with the patient’s anatomy to control and optimize the shape and fit of the implant on the bone. The material polyetheretherketone (PEEK) has been used in several medical applications such as cranial vault reconstruction.4,5 Patient-specific alloplastic implants have reduced both the need for major manipulation during surgery and the actual operative time. In preparation for calvarial defect reconstruction, a preoperative 3D-CT scan should be performed and the images sent to the implant manufacturer. An anatomically correct skull
4
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 2 (cont’d). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
model and an implant are built using rapid prototyping and sent to the physician for review and approval. The company then delivers the definitive PEEK-PSI to the physician who will perform the implant procedure.6 Reports have been published on the reconstruction of complex orbital frontotemporal reconstruction using PEEK-PSIs.6 PEEK polymers have been used in spinal surgery and orthopedic surgery and have been shown to be a highly reliable material with advantageous characteristics.7-9 On follow-up radiographs and CT imaging, translucency without artifacts has been observed.6,10 The advantageous radiolucent material characteristics of PEEK have made it a widely accepted alternative to metallic materials for spinal implants.7 Furthermore, it is compatible with magnetic resonance imaging, because it causes no magnetic interference. In addition, PEEK polymers have excellent chemical resistance, and the degree of elasticity or stiffness can be modified to fit the situ-
ation. The main drawback has been the possibility of postoperative infection.6 In the present study, we report on a new workflow using PEEK-PSIs for esthetic corrections in the facial region. The planning includes implant design according to the specifications derived from 3D simulation of soft tissue behavior.
Materials and Methods The present study followed the Declaration of Helsinki on medical protocol, and the ethics and ethical committee of the Medical University of Vienna (approval no. EK 665/2008) approved the study. PATIENT’S CASE
The present patient was female, 27 years old, and had severe congenital midface hypoplasia. Figure 1 shows preoperative 3D photograms of the patient.
5
GUEVARA-ROJAS ET AL
FIGURE 3. Iterations of osteotomy movements and simulation of soft tissue changes. A,B, Axial images with 0-mm bone segment reposition. C,D, Axial images showing bone segment reposition (3 mm anterior) and respective soft tissue changes. E,F, Axial images showing bone segment reposition (5 mm anterior) and respective soft tissue changes. (Fig 3 continued on next page.) Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
We decided to offer her treatment using PSIs for augmentation of the zygomatic prominence. The complete workflow included the following steps:
1. Computer-assisted treatment planning with soft tissue prediction and manufacturing of the PEEK implants using an interactive, iterative teleplanning process;
6
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 3 (cont’d). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
2. Surgery, with navigation-assisted insertion of the PEEK implants; 3. Evaluation of the concept (quantitative comparison of the soft tissue prediction with postoperative 3D photogrammetry).
COMPUTER-ASSISTED PLANNING
The planning began with importing the CT data of the skull (Philips Brilliance 64, Amsterdam, Holland) into the planning software, Mimics, version
7
GUEVARA-ROJAS ET AL
FIGURE 4. ‘‘Expressiveness’’ is an important measure of the quality of a simulation, especially in esthetic surgery. A, Reconstruction from computed tomography data provided precise information on the shape and geometry of the facial surface for the surgeon, but were obviously not sufficiently illustrative to inform the patient, primarily because the hair, eyebrows, and eyelashes were not visualized. B, Merging of a (conventional) digital photograph with the 3-dimensional reconstruction from the computed tomography data clearly enhanced the ‘‘natural and realistic’’ look, but lacked reliability and precision owing to distortions of the surface geometry from the matching method. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
14.0 (Materialise). The CT data were accessed using a secure ftp server for the rapid prototyping company (Synthes, Mesocco, Switzerland). Because no occlusal disturbance was present, maxillary advancement was
subtracted from the virtual Le Fort III advancement procedure. Consequently, isolated advancement of the zygoma resulted, which was the goal of the planning procedure (Fig 2). The subtraction volume of
FIGURE 5. Intraoperative view. Left, Electromagnetic patient reference frame fixed on the calvaria using microscrews, and Right, patientspecific implant in situ. Reliable and precise intraoperative positioning of the implant can be achieved with navigation. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
8
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 6. Screenshot of the soft tissue simulation model with color coding for the Hausdorff distance, evaluated using the software MESH. Left, Color coded (‘‘cold to warm’’) point to surface distances are shown. Far left, Perceptible within the histogram, almost all the distances remained less than 0.7 mm. Right, a rendering of the simulation planning CT is shown. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
the primary anatomy and the planned final situation according to the virtual Le Fort osteotomy gave the initial geometry of the implant. The realization of this concept has been illustrated in Figure 2. Fur-
ther refinement, including edge removal and contour smoothing, was achieved with mesh distortion tools and vertex removal. Different simulations of the osteotomies and bone repositioning were used for soft tissue simulation with Mimics (Materialise) (Fig 3), and evaluated by us using an interactive teleplanning process. Finally, a small implant was manufactured by the rapid prototyping department of Synthes. The ‘‘expressiveness’’ of the simulation is a crucial goal. Pure 3D reconstruction from CT data cannot sufficiently provide a realistic impression, especially for the patient (Fig 4). NAVIGATION-ASSISTED SURGERY
FIGURE 7. Histogram showing the point to surface distances of the points from the first to the second surface. Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
A coronal approach was used to avoid visible scars. Two PEEK onlay implants were used to augment both zygomatic prominences. The microscrew positions were predefined by a hole through the PSIs during the rapid prototyping process. Drilling in the bone for fixation screws was navigated using Fusion ENT Navigation (Medtronic, Minneapolis, MN), a customary computer-aided surgery (CAS) system using electromagnetic tracking technology. The patient reference frame
9
GUEVARA-ROJAS ET AL
FIGURE 8. Preoperative and postoperative photogrammetry images of the patient, showing the esthetic benefit. The postoperative photographs were taken 2 months after surgery. A,C,E, Preoperative photogrammetry images of the patient. B,D,F, Postoperative photogrammetry images of the patient. (Fig 8 continued on next page.) Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
was fixed to the calvaria using microscrews (Fig 5). A conventional registration procedure was performed on the basis of anatomic landmarks. POSTOPERATIVE FOLLOW-UP AND PHOTOGRAMMETRY
The patient’s postoperative face was scanned using a 3D photogrammetry system (Dimensional Imaging,
Glasgow, Scotland, UK) 2 months after surgery. To evaluate the simulation error using MESH, version 1.13 (open source11), the photogrammetric surface data set was first aligned with the simulated soft tissue data by surface registration using the planning software Mimics, version 14.0 (Materialise) on a defined region of interest outside the augmented area. In addition, another region of interest was defined in the
10
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
FIGURE 8 (cont’d). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
zygomatic area, and the Hausdorff distances were computed to ascertain the quantitative measure of planning error.
Results Our experience has shown the feasibility for routine application using standard instruments. The duration of the surgical intervention was 4 hours, including the approach and wound closure. No complications were observed during or after surgery. The patient expressed high satisfaction. Three weeks postoperatively the soft tissue swelling had resolved. At 2 months after surgery, a photogrammetry scan was performed. The simulated soft tissue surface and the photogrammetry scan of the patient showed a high correspondence, especially of the skin covering the implants. The mean total error (Hausdorff distance) was 0.81 1.00 mm (median 0.48, interquartile range 1.11). Figure 6 shows the soft tissue simulation model with color-coding for the Hausdorff distance. A histogram of the latter (Fig 7) indicated that the spatial deviation (ie, surface distance) remained at less than 0.7 mm for the vast majority of points. Outliers occurred only in the eyeball region; these were closed in the soft tissue simulation and opened during acquisition of the photogrammetric data (ie, red areas). The patient’s pre- and postoperative photogrammetry images are shown in Figure 8.
Discussion The present study evaluated a workflow for computer-aided design and manufacturing (CAD-
CAM) fabrication and CAS insertion of the patientspecific implants. Although this workflow included different software packages and several steps, the evaluation of the soft tissue simulation revealed good correspondence with the postoperative appearance. The accuracies observed in the present study remained reliably within the range of accuracy known from published studies of computer-assisted surgery.12-14 A high correspondence between the predicted facial surface geometry and the actual surgical outcome was achieved within the competence of the Mimics software (Materialise). The interactive, iterative teleplanning workflow proved feasible for efficient preparation of the surgical intervention. In particular, the integration of 3D photogrammetry contributed significantly to the ‘‘expressiveness’’ of the simulation (Figs 1, 4). However, the present workflow has one important shortcoming: the Mimics package allows for only soft tissue simulations associated with osteotomies. In the concept we have presented, the PSI design was based on the subtraction of the original data set from the simulated optimal surgical plan. Consequently, some deviations between simulation and outcome were inevitably caused by the method itself and could not be eliminated (ie, systematic error). Differences between subtraction and the final PSI implant can be seen in Figure 9. These images show that the geometries of the bone segment and implant differed mainly at the edges. Nevertheless, the remaining shape of the PSI and the bone were very similar. Therefore, we should always expect minor deviations in the soft tissue simulation based on osteotomies, although these deviations will be confined to the edges.
11
GUEVARA-ROJAS ET AL
FIGURE 9. A, Axial and B, coronal images showing that the geometries of the bone segment and implant differed mainly at the edges (green contour indicates the bone segment; red contour, the implant). Guevara-Rojas et al. Computer Planning of Patient-Specific Facial Implants. J Oral Maxillofac Surg 2014.
Computer-designed alloplastic implants could represent a milestone in the evolution of planning and realization of 3D reconstruction during surgery. They have the potential to become a reliable and irreplaceable part of the surgical armamentarium.6 The present report covers the first clinical case of maxillofacial reconstruction using computer-designed PEEKPSI bilateral augmentation. In conclusion, the present study has shown that the proposed workflow provides a complete CAD-CAMCAS chain for implant design that allows for soft tissue simulation, fabrication of patient-specific implants, and image-guided surgery to position the implants.
Much of the surgical complexity resulting from osteotomies of the zygoma, chin, or mandibular angle might be transferred to the planning phase of patient-specific implants. Furthermore, our method was resource intensive because it was a process development. In routine practice, with support of telemedicine (interactive teleplanning), the planning process could be transferred to a specialized center (commercial company), and the physician’s time would be freed. Acknowledgments We acknowledge the active participation of Synthes, who provided engineering knowledge and implant material.
12
COMPUTER PLANNING OF PATIENT-SPECIFIC FACIAL IMPLANTS
References 1. Mavili ME, Canter HI, Saglam-Aydinatay B, et al: Use of threedimensional medical modeling methods for precise planning of orthognathic surgery. J Craniofac Surg 18:740, 2007 2. Papadopoulos MA, Christou PK, Christou PK, et al: Threedimensional craniofacial reconstruction imaging. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 93:382, 2002 3. Patient Specific Implants. 2012. Available at: http://www. synthes.com/sites/intl/Products/CMF/Cranial/Pages/PatientSpecific-Implants.aspx. Accessed April 11, 2012 4. Chacon-Moya E, Gallegos-Hernandez JF, Pina-Cabrales S, et al: Cranial vault reconstruction using computer-designed polyetheretherketone (PEEK) implant: Case report. Cir Cir 77:437, 2009 5. Hanasono MM, Goel N, DeMonte F: Calvarial reconstruction with polyetheretherketone implants. Ann Plast Surg 62:653, 2009 6. Scolozzi P, Martinez A, Jaques B: Complex orbito-frontotemporal reconstruction using computer-designed PEEK implant. J Craniofac Surg 18:224, 2007 7. Kurtz SM, Devine JN: PEEK biomaterials in trauma, orthopedic, and spinal implants. Biomaterials 28:4845, 2007 8. Moon SM, Ingalhalikar A, Highsmith JM, Vaccaro AR: Biomechanical rigidity of an all-polyetheretherketone anterior thoracolum-
9.
10.
11.
12.
13.
14.
bar spinal reconstruction construct: An in vitro corpectomy model. Spine J 9:330, 2009 Walter J, Kuhn SA, Reichart R, et al: PEEK cages as a potential alternative in the treatment of cervical spondylodiscitis: A preliminary report on a patient series. Eur Spine J 19:1004, 2010 Citak M, Kendoff D, Wanich T, et al: The influence of metal artifacts on navigation and the reduction of artifacts by the use of polyether-ether-ketone. Comput Aided Surg 13:233, 2008 Aspert N, Santa-Cruz D, Ebrahimi T: MESH: Measuring errors between surfaces using the Hausdorff distance. Multimedia and Expo, 2002 ICME ’02 Proceedings 2002. IEEE International Conference 1:705, 2002 Klug C, Schicho K, Ploder O, et al: Point-to-point computerassisted navigation for precise transfer of planned zygoma osteotomies from the stereolithographic model into reality. J Oral Maxillofac Surg 64:550, 2006 Wagner A, Schicho K, Birkfellner W, et al: Quantitative analysis of factors affecting intraoperative precision and stability of optoelectronic and electromagnetic tracking systems. Med Phys 29:905, 2002 West JB, Fitzpatrick JM, Toms SA, et al: Fiducial point placement and the accuracy of point-based, rigid body registration. Neurosurgery 48:810, 2001
