Accuracy of 3-dimensional Printing of Orthodontic Study Models by the Objet30 OrthoDesk Printer
MetadataShow full item record
INTRODUCTION 3D printing capabilities are rapidly becoming commonplace among orthodontic laboratories as an increasing number of practitioners begin to invest in intraoral and desktop scanning equipment in their practices. Similarly, digital .STL computer software has been advancing with the demand for 3D printing. Software programs are now available that are capable of detailed in-depth analyses of .STL models. This study aimed to examine printed rapid prototype orthodontic models and digitally compare them to their plaster counterparts. The rapid prototypes were printed by one of the most common 3D printers in the orthodontic field-the Objet30 OrthoDesk printer. MATERIALS AND METHODS In this study, 30 maxillary plaster models were randomly selected from the finished case archives at the University at Buffalo, School of Dental Medicine, Orthodontic Department and then scanned using the Maestro 3D Desktop Scanner to produce digital .STL files. Pyramidal landmarks were digitally added to the models to provide stable and reproducible reference points for measurement during the study. The 30 models were then printed by the Objet30 OrthoDesk. The rapid prototypes produced were re-scanned and digitally compared to the .STL files of the original plaster model counterparts. Linear measurements were taken using Netfabb software and digital superimposition and cross-section analysis was conducted using Ortho Analyzer software. RESULTS The first portion of the study included 13 types of linear measurements (8 pyramidal landmark measurements and 5 dental anatomical measurements). A Bland-Altman analysis was performed where a comparison of the difference between the 2 methods of measurement (plaster and rapid prototype) was made to the average of the 2 measurements. The differences between plaster and rapid prototype were generally smaller for pyramid measurements than for non-pyramid measurements. 8 measurement types were found where the plaster and rapid prototype are significantly different: Pyramid3 - Pyramid14 (p < 0.001), Pyramid3 - Pyramid6 (p<0.001), Pyramid11 - Pyramid14 (p<0.006), Pyramid3 - PyramidP (p<0.036), Pyramid6 - PyramidP (p<0.007), MD3 (p<0.008), MD8 (p < 0.001) and MD13 (p < 0.001). For the non-pyramid measurement types, 3 of the plaster measurements are statistically significantly smaller on the average than for the rapid prototype measurements. For Pyramid3 - PyramidP and Pyramid6 - PyramidP the plaster tends to be less than for rapid prototype while for Pyramid3 - Pyramid14, Pyramid3 - Pyramid6 and Pyramid11 - Pyramid14 the plaster measurements on the average are larger than the rapid prototype measurements. The second portion of the study involved digital superimposition of plaster and rapid prototype counterparts using Ortho Analyzer software. Overlays were judged to be acceptable or unacceptable by examination of cross-sections of teeth #4 and #9 as well as cross-sections of the palatal pyramid in both the transverse and anteroposterior dimensions. Out of the 30 models, 13 pairs had produced acceptable overlays through the Ortho Analyzer superimposition. Overlays deemed unacceptable were excluded from this portion of the study since the software failed to produce an accurate superimposition that would yield relevant data. The greatest discrepancy between the plaster and rapid prototype overlay was measured at several locations in the cross-sections of tooth #4, tooth #9, and the palatal pyramid. The facial surface and palatal surface of tooth #4 on the rapid prototype model was thicker than the plaster model 100% of the time, measuring on average 0.162mm thicker facially and 0.130mm palatally. The rapid prototype buccal cusp was on average 0.076mm deficient and the palatal cusp was on average 0.108mm deficient compared to the plaster model counterpart. The three measurements recorded within the center region of the occlusal surface (lingual incline of buccal cusp, central groove, buccal incline of palatal cusp) of tooth #4 were variable. All 3 regions for tooth #9 displayed statistical significance. The mean value for the discrepancy in the area of the facial surface was a 0.097mm excess in the rapid prototype. The lingual surface showed an even greater mean excess of 0.222mm for the rapid prototype model. The incisal edge showed a mean deficiency for the rapid prototype of 0.053mm compared to the plaster model. Finally, the palatal pyramid was cross-sectioned in the transverse dimension and the discrepancy between the apex of the plaster model and the rapid prototype model was recorded. The mean value for the discrepancy between the palatal pyramid apices was a deficiency of 0.555mm for the rapid prototype. CONCLUSIONS The results of the study indicate slight dimensional discrepancies between rapid prototypes and their original plaster models. Patient use of appliances fabricated from rapid prototype models should be the aim of future studies in order to evaluate clinical significance. Furthermore, the American Board of Orthodontics should maintain their current guidelines and continue to prohibit printed models for the Cast/Radiograph Evaluation at this time.