P3Dental –Three-dimensional virtual preoperative implant planning using cone beam computed tomography images
A.M. Marques da Silva¹, E.C. Hoffmann², E.G. Link², A.B. Trombetta², F. Bacim² and J.A. Borges², ¹Pontifical Catholic University of Rio Grande do Sul, School of Physics, Porto Alegre, RS, Brazil ²Protótipos 3D, TECNOPUC, Porto Alegre, RS, Brazil.
Abstract—This article presents the development of a freeware software system for 3D virtual preoperative implant planning, called P3Dental, based on cone beam computed tomography (CBCT) or conventional CT images.The methodology requires two CT acquisitions: one image of the patient’s dental splint with gutta-percha fiducial markers; and another typical CT of the patient using the dental splint with the fiducial markers. A segmentation process is followed by a iso-surface volume rendering to create a STL 3D mesh. A semi-automatic threshold is used to extract the fiducial markers.The generalized Karhunen-Loève transform is used for the volumetric rigidregistration. The methodology was tested with five CBCT patient images. The linear and angular deviations between fiducial markers were between 0.5 and 2.0 mm in the x axis, and between 0.5 and 5 degrees, respectively. These deviation values are compatible with other in vitro studies,showing the reliability of the software for virtual surgical planning for implantodonty.
Contemporary preoperative planning techniques rely on three-dimensional (3D) surface model representations of the jaw bone obtained from image data acquired with computed tomography (CT), using a segmentation algorithm for extracting usinga segmentation algorithm for extracting the bony tissues. Computer systems with volumetric navigation enable the surgeon to locate the position of the implants.
Cone beam computed tomography (CBCT) allows a lower effective radiation dose,with out significant loss of image quality, in a short scanning time, with reduced costs. Compared with multislice CT, CBCT holds promising potential for oral and craniofacial imaging applications, playing a vital role in the diagnosis of hard tissue structures of the dentomaxillo facial region .
Artifacts inherent to CBCT image acquisition related to detector sensitivity, X-ray beam inhomogeneity and reconstruction technique result in a noticeably higher noise level, which influence the accuracy of the 3D surface models .These drawbacks may influence image quality, lowering the bone segmentation accuracy. Currently there are few software systems using CBCT images to aid in oral planning surgery and produce surgical drilling guides.These guides are manufactured in such a way that they match the location, trajectory, and depth of the planned implant with a high degree of precision.These guides stabilize the drilling by restricting the degrees of freedom of the drill trajectory and depth.
This article presents the development of a new independent software system for 3D virtual preoperative implant planning (P3Dental),based on dual acquisition of cone beam computed tomography (CBCT) images.
MATERIALS AND METHODS
A.Image AcquisitionThe system requires two CBCT datasets: •CBCT images of the patient’s dental splint (removable dental appliance reproducing a“bite”, made by acrylic) with gutta-percha fiducial markers •CBCT images of the patient using the dental splint with the gutta-percha fiducial markers.
B – Image processing
•Segmentation and Volume Rendering for both datasets – OsiriX.
CT images of 5 edentulous anonymous patients were used for P3Dental software testing. Images were acquired using two commercial CBCT systems: •i-CAT (Imaging Sciences International Inc.,USA) •Kodak 9500 Cone Beam 3D System (Kodak Dental Systems,Care stream Health,Rochester, NY ,USA).
RESULTS AND DISCUSSION
P3Dental software was developed in a user-friendly interface for dental professionals, offering interactive 3D image tools,zoom function,contrast settings,distance and angle measurement tools, 2D orthogonal and other arbitrary plane images,and panoramic view. A 3D and 2D conventional surface rendering view of the dental splint and the surface line in each slice plane are also available (Figure1).
Figure 2 shows an example of one patient CBCT image acquisition,where the gutta-percha fiducial markers in the patient dental splint are identified as the six bright points(Fig 2a). Since the gutta-percha spheres are positioned at the external surface of the dental splint, the radiological markers do not usually overlap with dental filling artifacts from metal licelements,such as tooth fillings, brackets, or surgical plates,allowing easy identification in the image dataset in most cases (Fig 2b). The preliminary methodology tests showed that at least three fiducial markers visible in the segmented CBCT images are required for good registration.The dental templates were adapted to all patients in a satisfactory and stable manner.
Fig. 2.CBCT images from: (a) the dental splint, where gutta-percha markers can be viewed as radiopaque (bright) points; (b) the patient with the dental splint attached.
Fig.1. P3Dental interface for dental professionals with orthogonal reconstructions, arbitrary angle blended slices, CT slices, panoramic view, surface rendering view of the dental splint and the surface line in each slice plane.
The linear and angular deviations between fiducial markers in the dental splint and the CBCT patient images were between 0.5 and 2.0mm in the x axis,and between 0.5 and 5 degrees,respectively. These values are compatible with other studies with in vitro jaws and CBCT conventional images .
Comparison of our results with those of other studies is limited because of the rarity of in vivo studies involving this type of double CBCT imaging technique. Usually the accuracy comparison studies are made with in vitros pecimens  or with conventional and MSCT images in vivo studies . Our results suggest that P3Dental preoperative implant virtual planning methodology using double imagining with cone-beam CT data coupled with radiological guides is viable.Positioning and angle deviations showed errors comparable to other in vitro studies.More CBCT in vivo studies must be done to assure the reliability of the methodology proposed. A new module of virtual planning for mini-screws positioning is under development and test. An ABS prototyping machine using the principle of stereolithography is being used to fabricate the SLA surgical guides. New CT scans will be made for all subjects after implant insertion to assess the magnitude of error in transferring the planned position of implants from CT scans to a surgical guide.
REFERENCES Liang X,et al.(2010a).A comparative evaluation of cone beam computed tomography (CBCT) and multi-slice CT (MSCT): PartI.On subjective image quality. European Journal of Radiology 75:265-269.  Endo M,Tsunoo T, Nakamori N,Yoshida K.(2001) Effect of scattered radiation on image noise in cone beam CT. Med Phys 28 (4):469-474.  Brooks RA,DiChiroG.(1976) Beam hardening in reconstructive to-mography. Phys Med Biol 21:390-398.  Hsieh J, MolthenRC,Dawson CA,Johnson RH.(2000) An iterative approach to the beam hard ening correction in cone beam CT. Med Phys 27:23-29.  Liang X, et al.(2010b) A comparative evaluation of cone beam computed tomography (CBCT) and multi-slice CT( MSCT).PartII:On 3D model accuracy.European Journal of Radiology 75:270-274.  HuaY,LiuW(1998)Generalized Karhunen-Loève transform,IEEE Signal Processing Letters, 5(6):141-142,doi:10.1109/97.6814301-469.  Loubele M,et al.(2008)Comparative localized linear accuracy of small-field cone-beam CT and multislice CT for alveolar bone measurements. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 105:512-518.  Ozan O et al.(2009)Clinical Accuracy of 3 Different Types of Computed Tomography-Derived Stereolithographic Surgical Guides in Implant Placement, Journal of Ora land Maxillofacial Surgery,67(2):394-401.
Download Software http://www.p3dental.com.br