1
Comparative Accuracy of Facial Models Fabricated Using Traditional and 3D Imaging Techniques

Ketu P. Lincoln, DMD,1 Albert Y. T. Sun, PHD, Cswa,2 Thomas J. Prihoda, PHD,3 and Alan J. Sutton, DDS, MS, FACP 4

1Department of Graduate Prosthodontics, USAF, Joint Base San Antonio-Lackland, San Antonio, TX

2Department of Mechanical Engineering, National Taipei University of Technology, Taipei, Taiwan

3Department of Pathology, University of Texas Health Science Center, San Antonio, TX

4Department of Restorative Dentistry, University of Colorado School of Dental Medicine, Aurora, CO

Keywords
Moulage; facial prosthetics; 3D imaging; 3D models; dental materials; stereolithography; rapid prototyping.

Correspondence
Alan Sutton, 13045 E. 17th Ave Ste F845, Aurora, CO 80045. E-mail: Alan.sutton@ucdenver.edu

The authors deny any conflicts of interest.

Accepted June 14, 2015

doi: 10.1111/jopr.12358

Abstract

Purpose: The purpose of this investigation was to compare the accuracy of facial models fabricated using facial moulage impression methods to the three-dimensional printed (3DP) fabrication methods using soft tissue images obtained from cone beam computed tomography (CBCT) and 3D stereophotogrammetry (3D-SPG) scans.

Materials and Methods: A reference phantom model was fabricated using a 3D-SPG image of a human control form with ten fiducial markers placed on common anthropometric landmarks. This image was converted into the investigation control phantom model (CPM) using 3DP methods. The CPM was attached to a camera tripod for ease of image capture. Three CBCT and three 3D-SPG images of the CPM were captured. The DICOM and STL files from the three 3dMD and three CBCT were imported to the 3DP, and six testing models were made. Reversible hydrocolloid and dental stone were used to make three facial moulages of the CPM, and the impressions/casts were poured in type IV gypsum dental stone. A coordinate measuring machine (CMM) was used to measure the distances between each of the ten fiducial markers. Each measurement was made using one point as a static reference to the other nine points. The same measuring procedures were accomplished on all specimens. All measurements were compared between specimens and the control. The data were analyzed using ANOVA and Tukey pairwise comparison of the raters, methods, and fiducial markers.

Results: The ANOVA multiple comparisons showed significant difference among the three methods (p < 0.05). Further, the interaction of methods versus fiducial markers also showed significant difference (p < 0.05). The CBCT and facial moulage method showed the greatest accuracy.

Conclusions: 3DP models fabricated using 3D-SPG showed statistical difference in comparison to the models fabricated using the traditional method of facial moulage and 3DP models fabricated from CBCT imaging. 3DP models fabricated using 3D-SPG were less accurate than the CPM and models fabricated using facial moulage and CBCT imaging techniques.

Craniofacial dysmorphology (CD) is the study of structural defects caused by trauma, treatment of neoplasms, or congenital anomalies characterized by complex irregularities in the shape and configuration of facial soft tissue structures.1 Patients with CD may undergo extensive surgical procedures, including the fabrication of facial prostheses to restore an extraoral maxillofacial defect.2 The facial prostheses are not functional, but provide the patient with an esthetic result for psychological and social acceptance.3-6

Anthropometry is a way to assess changes in facial soft tissue over time through line measurements between two landmarks.7 The challenge has been to identify landmarks and plot them accurately in the three planes of space, in order to describe the dimensions of the face.8 Traditionally, direct anthropometry was done using calipers. This assessment was a reliable and inexpensive method for data collection of surface measurements.4 However, there were several limitations, including technician training, direct patient contact requiring extensive time to make multiple measurements, patient compliance to sit in one position, inability to archive information, difficulty attaining several measurements as tissue undergoes changes with time, and finally comparing tissue changes with accurate landmark location.9

Making a facial moulage impression was, and still is, another means for 3D facial structure capture, analysis, and documentation. This method has been used successfully for almost 100 years, dating back to World War I.10 Currently, various impression materials like alginate, poly(vinyl siloxane), and reversible hydrocolloid are used to create a facial moulage. The facial moulage method can be time consuming, and soft tissue deformation is a significant problem. Furthermore, it is difficult to obtain accurate impressions of certain defects involving the orbit where the periorbital tissue displaces easily.11 The casts made from the impressions are fragile and require large physical storage space, and it is extremely difficult to communicate physical data to other providers in distant locations.12 Also, archival preoperative casts may not be available for many patient treatments due to storage limitation.

Several types of 3D imaging systems have been created in the past three decades, including cone beam computed tomography (CBCT) and 3D stereophotogrammetry (3D-SPG). Both methods are noninvasive and allow for archival of data and virtual models that can subsequently be used for comparison purposes.

Computed tomography (CT), and more specifically CBCT, is currently used to capture soft tissue surface images because it is accurate and repeatable for anthropometric measurements.7 Collimating the X-ray beam decreases the radiation exposure dose, and the scan time is 10 to 70 seconds.13 The dose of radiation ranges between 60 and 1000 µSv versus medical grade CT of the mandible, which ranges from 1320 to 3324 µSv.13-15 More recent studies have generated 3D facial soft tissue surface computer models from image data captured by CBCT. Linear anthropometric measurements on computer models using CBCT software proved reliable and as accurate as the traditional direct method.7 The data and virtual models are easily archived without physical storage requirements and can provide pre- and postoperative information for skeletal or soft tissue comparisons.13

3D-SPG is a newer technique/method for craniofacial surface imaging that allows for the capture evaluation of the external surface of a subject. The method creates a 3D image reconstructed from multiple digital images taken at different angles simultaneously. The resultant image is a collection of points positioned along an x, y, and z coordinate system. These points can be identified as landmarks, then used for subsequent analysis.9 Reports indicate that 3D-SPG is reliable and accurate for determining the location of landmarks and interlandmark craniofacial distances.16,17 The advantages include minimal artifact production due to short image capture time (approximately 1.5 ms), ability to archive and compare subject images, three-point (x, y, z) coordinate format of locating tissue landmarks, high resolution, and no radiation. Software programs are available to identify landmarks and calculate anthropometric measurements.18 In addition, the error in the location of a landmark when using 3D-SPG is less than 1 mm.19

The use of 3D-SPG has a great potential for use in the military. During World War II, the Korean War, and the Vietnam War, the mean incidence of head, face, and neck injury (HNFI) was approximately 16%. A recent study looked at the characteristics and causes of HFNIs sustained by US military forces during the stability and support phase of Operation Iraqi Freedom (OIF-II). The number of HFNIs increased to 39%, and of these injuries, 65% were injuries to the face.20 A more recent study showed a comprehensive analysis of craniomaxillofacial battle injuries sustained by military members evacuated to level III-V military treatment facilities to be 42.2% HNFIs.21 The reason for the notable increase in the past decade is an increase in survival rate due to improvement in body armor, battlefield medicine, tactically placed medical units, and quick evacuation tactics.

Both CBCT and 3D-SPG use computer-aided design (CAD) software to facilitate the design of soft tissue surface images and virtual models. With rapid prototyping (RP), information from the CAD software can be used along with computer-aided manufacturing (CAM) to fabricate 3D physical models. Image data from CBCT and 3D-SPG scans translated into the digital imaging and communication in medicine (DICOM) file format, which are converted to a CAM file format to produce a 3D model using RP methods and equipment.

One RP method used in the medical and dental field is 3D printing (3DP). This process uses a polyjet selectively depositing fine powder polymer droplets evenly along a piston and liquid binder. Additional layers are added as the piston powder bed and cured model is lowered layer by layer. The resolution accuracy is 100 µm for one-dimensional features and 300 µm for 3D features.11 The 3D printed models are accurate to 0.016 mm (Objet Eden 260V; Stratasys Ltd., Minneapolis, MN), and the build time is at a rate of 1 cm of height per hour.22

The 3D models are useful for surgical planning, creation of surgical templates, and fabrication of craniofacial prostheses and custom implants used in...