CAD/CAM and rapid prototyped scaffold construction for bone regenerative medicine and surgical transfer of virtual planning: A pilot study

1. Introduction 

Scaffolds for stem cell seeding, ingrowth, and new tissue formation are in great demand in bone reconstructive orthopedic surgery. The scaffold structure must be biocompatible, biomimetic, and biodegradable. Moreover, the scaffolds employed must fit into the anatomical defect, possess mechanical properties capable of bearing the loads encountered in vivo, and produce biocompatible degradation by-products [1], [2], [3], [4]. The relationship between scaffold structure and tissue compatibility has been widely discussed over the past decade. Gauthier et al. defined the minimum pore size for osteoconduction in the range of 80–100μm [5]. Other studies discussed the biomimetic properties of diverse materials in terms of mechanical strength and loading force direction related to the trabecular organization of the scaffold. Landi et al. presented a novel nano-sized, pure, carbonate apatite powder with morphological and compositional features mimicking natural apatite with improved thermal properties. The samples showed high compressive strength along with biomimetic morphology [6]. Williams et al. evaluated the solid free-form fabrication (SFF) and the use of selective laser sintering (SLS) for scaffold construction using polycaprolactone bioresorbable polymer. The authors concluded that the integration of SFF and SLS techniques proved highly useful for construction of scaffolds with anatomy-specific exterior architecture [7]. Other studies investigated the biocompatible degradation of the scaffolds [8], [9]. The solid free-form fabrication of scaffolds enables use of three-dimensional computed tomography (3D CT) data to design anatomically shaped scaffolds with varying internal architectures, thereby allowing precise control over pore size, porosity, permeability, and stiffness [10], [11], [12], [13]. The construction of a biphasic scaffold to reproduce the bony anatomy of cortical-medullar bone or the anatomy of complex articular surfaces, such as the temporomandibular joint (TMJ), can be obtained using SFF and SLS. A recent report by Kong et al. [14] described a porous nano-hydroxyapatite–chitosan composite scaffold created with a multilayered structure. The results of this study demonstrated enhanced mechanical strength and more suitable ingrowth of stem cells. After 12 weeks post-insertion in a rabbit fibula defect, the center of the scaffold was found to be rich in blood vessels and bone formation.

This report describes a pilot study to develop a new protocol for animal experimentation based on computer-aided design/computer-aided manufacturing (CAD/CAM) and rapid prototyping (RP) in order to design hydroxyapatite (HA) scaffolds for bone marrow stem cells to reconstruct bony defects of a functional stress loaded area (i.e., the TMJ). The protocol is the first stage of a study of the construction of anatomical spare bone elements for the human body that will be used for bony reconstruction in orthopedics. The study was designed according to two main research guidelines: obtain an anatomically designed scaffold and reproduce the virtually planned sectioning of the condyle during experimental surgery in an animal model.

2. Materials and methods 

In this pilot study, the mandible extracted from a dead pig was chosen for the surgical experiment. A bone defect (condyle resection) was created virtually in a 3D digital model of the right mandibular ramus using computed tomography and computer-aided design elaboration. A model of the resected condyle was produced by rapid prototyping to evaluate the fit of the bone substitute scaffold according to this protocol. A surgical guide system was also developed to reproduce the virtually planned bone sectioning procedures during surgery accurately. Finally, in the surgical experiment, the condyle of the right mandibular bone was removed and replaced with the rapid prototyped plastic scaffold.

2.1. CAD and rapid prototyping of the pig mandible 

The pig mandible was examined using CT performed with a GE Healthcare BrightSpeed instrument (GE Healthcare, Fairfield, CT). The entire data acquisition procedure was carried out at the Radiology Department of Istituti Ortopedici Rizzoli of Bologna, Italy. CT scanning was performed with a slice thickness of 0.625mm and an image resolution of 512×512 pixels. This CT acquisition yielded 243 slices, which were stored in DICOM (digital imaging and communications in medicine) files.

Beginning with the stack of CT slices, a 3D digital model of the mandible (without soft tissue) was reconstructed by setting a suitable threshold value using an advanced software tool for 3D visualization, data analysis, and geometry reconstruction (Amira 3.1.1; Mercury Computer Systems, Chelmsford, MA).

This model was achieved semiautomatically by threshold-based segmentation, contour extraction, and surface reconstruction. These features are particularly useful for distinguishing between soft tissues and skeletal structures. Moreover, the right mandibular ramus was isolated from the whole reconstructed model of the mandible. Finally, manual operations, such as filling holes or deletion of erroneous polygon data, were needed to obtain a watertight mesh (130,426 points and 261,562 triangular faces) in STL format (Fig. 1a) suitable for reproduction by rapid prototyping. These refinement editing operations were performed using Rapidform XOS2 software for 3D scanners (INUS Technology, Seoul, Korea).

View Large Image Download to PowerPoint

Fig. 1. (a and b) CAD virtual planning of the bone section.

A fragment of the right mandibular ramus was then virtually cut away from the model to artificially create a bone defect in the mandible (Fig. 1b). This condyle resection was performed after planning the line cut with the surgeon, and led to two separate STL watertight meshes: the defective mandible (115,291 points and 231,292 triangular faces) and the resected condyle (16,427 points and 32,850 triangular faces). These two models fit together perfectly due to the virtual cut. Two holes 2mm in diameter for titanium fixing screws (Gebrüder Martin, Tuttlingen, Germany) were also included in both the intact and defective models of the right mandibular ramus.

The designed models were fabricated using a Stratasys Dimension SST (Soluble Support Technology) 3D printer (Stratasys, Eden Prairie, MN), which is a rapid prototyping system based on fused deposition modeling (FDM). This system performs layer-by-layer deposition of both fused acrylonitrile butadiene styrene (ABS) plastic and soluble support material, which supports the overhanging parts of the models under construction with a layer thickness of 0.254mm. The process is then completed by washing the models to remove the support material.

2.2. CAD and rapid prototyping of surgical guides 

Two surgical guides were constructed to reproduce virtual sectioning of the mandible in the surgical environment. These templates were custom-designed to allow the surgeon to section the condyle in the same line as the virtually planned section.

The surgical excision was designed in two steps: the first for resection of the condyle, and the second for refinement of the surface of the cut. This procedure permits positioning of the designed scaffold into the defect at the time of surgery. The surgical guides for condyle sectioning of the pig mandible were developed in two parts for use in sequence. Both templates also had two holes 2mm in diameter for titanium fixing screws in the same positions. Moreover, to position the surgical guides on the mandible, the reference contact surface between the templates and the bone was designed around the two holes for titanium fixing screws and at the extremities of the guides. The absence of contact along the main surface of the surgical templates allows fixation to the bone even in the presence of any residual soft tissues.

The planned condyle resection procedure using the designed surgical templates was first shown in a virtual environment. Template 1 (Fig. 2a) was designed to be positioned by the surgeon when the mandible is still intact and the condyle is inside: two holes for the titanium fixing screws must be made in the lateral part of the mandibular ramus to fix the template before sectioning. After condyle removal, the surgeon removes template 1 and positions template 2 to refine the surface of the cut plane (Fig. 2b) using the same holes in the bone and fixing it in the same position as template 1. In this way, the area to fit the scaffold is compatible with the planned volume of the scaffold.

View Large Image Download to PowerPoint

Fig. 2. (a and b) Design of surgical guides used to reproduce the virtual planned bone sectioning during surgery; (c and d) rapid prototypes used to test and validate the planned surgery; (e and f) initial cut during the surgical experiment; (g and h) refinement cut during the surgical experiment.

The two surgical guides were then directly fabricated in ABS using the Stratasys Dimension SST 3D printer. To gain a better understanding of the surgical intervention, the prototyped models were also used by the surgeon to test and validate the planned surgery. The first template was applied to the physical model of the intact mandibular ramus (Fig. 2c) and then the second surgical guide was positioned on the model of the defective mandible (Fig. 2d), allowing physical checking of the designed procedure.

2.3. Surgical experimentation 

The prototyped surgical guides and the designed volume of the scaffold were actually tested during surgical experiments using the same pig mandible that had been previously scanned by CT. Two bone screws were employed to achieve rigid fixation in the planned holes of the templates. Precise sectioning was achieved using the designed surgical guides in sequence: the initial cutting was performed with an oscillating saw (Fig. 2e and f), and the surface was then refined as planned in the protocol (Fig. 2g and h).

Surgical templates fit on the mandible surface were facilitated by the elasticity of the material. Furthermore, the precise (Fig. 5a and b) (Table 1) sectioning of the bone allowed very precise location of the model scaffold in the artificial defect (Fig. 3a). It was fixed to the bone using two bone plates held in place by bone screws passed through the same holes made to fix the surgical guides (Fig. 3b).

Table 1. Comparison of the main measures of the cut.

		Planned cut in the virtual environment	Actual cut made by the surgeon	Comparison (%)
	Height of the cut (mm)	49.12	49.00	0.24
	Depth of the cut (mm)	13.14	12.89	1.90
	Angle of the cut (°)	111.61	110.53	0.97

View Large Image Download to PowerPoint

Fig. 3. (a and b) Final result after fixing the model scaffold to the bone.

Fig. 4 shows a flow chart of the protocol.

View Large Image Download to PowerPoint

Fig. 4. Flow chart of the planned protocol with reference to the previous figures and indication of the technology used: CT (computed tomography), CAD (computer-aided design), and RP (rapid prototyping).

3. Results 

The accuracy of the proposed method was evaluated quantitatively by comparing the planned cut in the virtual environment with the actual cut made by the surgeon to remove the condyle. After the surgical experiment, the rapid prototyped condyle with the two holding plates was disconnected from the bone and a reverse engineering (RE) process was used to digitalize the cut bone of the pig mandible. A 3D laser scanner Vivid 9i (Konica Minolta Holdings, Tokyo, Japan) was used to scan the cut mandible and the corresponding digital model was created using Rapidform XOS2 (INUS Technology, Seoul, Korea).

The accuracy of the method was evaluated using the shell–shell deviation analysis technique provided by Rapidform. Consequently, the digital model of the defective mandible and the scanned model of the test cut-bone were superimposed to provide a color map of the distance deviation between the two models, limiting this analysis to the region around the cut (Fig. 5a). As result, the mean value was 0.35mm and the maximum value was 1.18mm. Moreover, the main measures of the cut (Fig. 5b) were compared and are listed in Table 1.

View Large Image Download to PowerPoint

Fig. 5. Evaluation of the accuracy of the planned procedure by comparison between the designed and actual surgical cut: (a) digital model of the defective mandible (left), scanned model of the actually cut bone (center), and the distance deviation color map between the two superimposed models (right); (b) main measures of the planned cut (left) and actual surgical cut (right).

Finally, the efficiency of the entire process was evaluated in terms of both time and cost. As shown in Table 2, the total time needed to process the virtual models of the prostheses and the surgical guides does not exceed 8h. The cost of ABS materials needed to produce such models is also very low, as usually required for industrial prototyping systems.

Table 2. Time and cost of the procedure.

		CAD time (h)	ABS model material (cm3)	Support material (cm3)	Build time	Cost of material (€)
	Intact mandibular ramus	2	46.75	23.81	12h, 31min	16.98
	Defect mandibular ramus	–	40.68	15.84	10h, 27min	13.71
	Condyle	–	7.33	4.42	2h, 14min	3.67
	Surgical guide no. 1	3	5.20	2.10	45min	2.64
	Surgical guide no. 2	3	6.03	3.12	1h, 1min	3.07
	Total	8	105.99	49.29	26h, 58min	40.07

4. Discussion 

Several studies over the last decade have focused on CAD/CAM scaffold construction using various scaffold materials. Bone regenerative medicine and histological studies on stem cell seeding inside the scaffolds indicated that pure hydroxyapatite is one of the best materials for regeneration of new bone. Bone marrow stem cells show a major difference in bone regeneration when seeded in a trabecular scaffold. However, the problem of planning the external volume of the scaffold and strategies for its rapid prototyping using HA are challenging. Here, we described a novel protocol that can be used to produce custom-made scaffolds for bone regenerative medicine.

Cooperative and collaborative work among engineers, surgeons, and prosthetic designers has resulted in the efficient exploitation of available instruments and technologies. Several multidisciplinary studies involving both mechanical engineers and medical departments have provided the opportunity to share innovative technologies and opened the way toward new approaches in prosthetic design and surgical planning [15], [16]. Many of these technologies, such as reverse engineering, CAD/CAM, virtual prototyping, and rapid prototyping of free-form shapes, have been conceived and developed in the industrial design sector [17]. Nevertheless, surgeons are always providing new requirements, which sometimes lead to the development of customized CAD/CAM solutions in the framework of specific technologies [18], [19]. In the present study, virtual and rapid prototyping was implemented to test several aspects of the designed procedure and to gather early user feedback on physical replicas of bones, prostheses, and guides. Total production time and cost were evaluated to determine the economic impact of such procedures. As described in Table 2, the total times and material costs in this pilot study and consumed for the models were affordable.

5. Conclusions 

This study represents the first step of a wider experimental protocol regarding scaffold creation for bone regenerative medicine, and tested the CAD/CAM elaboration of the scaffold external complex surface. The inner part of the scaffold, which reproduces the trabecular part of the bone, may be obtained in future, providing the scaffold with a reticular internal structure with well-defined reticules. Further studies are necessary to develop and test new 3D printing methods for scaffold material.

The main advantages of the procedure presented in this paper are the following: