Comparing the accuracy of 3 different liquid crystal display printers for dental model printing

Introduction

This study aimed to evaluate the accuracy in terms of trueness and precision of 3 different liquid crystal display (LCD) printers with different cost levels.

Methods

Three LCD 3-dimensional (3D) printers were categorized into tiers 1-3 on the basis of cost level. The printers’ accuracies were assessed in terms of trueness and precision. For this research, 10 standard tessellation language (STL) reference files were used. For trueness, each STL file was printed once with each 3D printer. For precision, 1 randomly chosen STL file was printed 10 times with each 3D printer. After that, a model scanner was used to scan the models, and STL comparisons were performed using reverse engineering software. For the measurements regarding trueness and precision, the Friedman test was used.

Results

There were significant differences among the 3 printers ( P <0.05). The trueness and precision error were lower in models printed with a tier-1 printer than in the remaining 3D printers ( P <0.05). The tier-2 and -3 printers presented very similar performance.

Conclusions

LCD 3D printers can be accurately used in orthodontics for model printing depending on the specific orthodontic use. The cost of a printer is relevant to the results only for the higher expense of the 3D printer in this study.

Highlights

•

Liquid crystal display 3D printers can be accurately used in dentistry and orthodontics for model printing.
•

The cost of such a printer is relevant to the results only for the higher-expense 3D printers.
•

The most expensive out of the 3 liquid crystal display printers performed the best.
•

The entry-level 3D printer performed similarly to the 2 times more expensive 3D printer.

During the last decade, 3-dimensional (3D) printing technology has been used in the dental field, more specifically in orthodontics. The fabrication of dental casts was one of the first 3D printing uses in orthodontics, followed by the fabrication of various custom appliances and the fabrication of sterile aerosol aspirators. ^, Direct measurements were conducted on plaster models, and digital measurements obtained from model scanners were found to be identical. Furthermore, the usage of intraoral scanners in everyday practice in orthodontics gave doctors the ability to take impressions of dental arches without causing any discomfort to the patients. This resulted in the mandatory use of 3D printers to print dental casts that could be used for diagnosis or appliance fabrication. A very promising combination for 3D imaging appears to be laser scanning combined with stereolithographic biomodeling.

In 1986, Charles Hull introduced to the scientific world the first 3D printer. Stereolithography (SLA) was the first type of 3D printing technology presented by Hull. In 1990, Scott Crump introduced fused deposition modeling. Today, there are various 3D printing technologies. Some of these are laser-SLA, direct light processing (DLP), liquid crystal display (LCD), fused filament fabrication, and PolyJet photopolymer. SLA, DLP, and LCD are the most popular in the orthodontic field.

LCD printing technology is a type of VAT polymerization technology. These printers are characterized by a vat and a building platform. A liquid photopolymer resin is placed in the vat to create the printed model. The light source for these printers is an LCD. During the printing process, the light shines in parallel, coming through the LCD panels onto the build area. It is important to mention that the light isn’t expanded using any lens or other device. Therefore, pixel distortion is not an issue when working with an LCD 3D printer. According to the present literature, SLA and DLP printers can provide more accurate results on 3D-printed dental models than entry-level LCD printers. However, there are no studies that assessed the accuracy of high-end LCD printers in 3D dental model printing. The primary benefit of LCD printers over SLA printers is that they are faster in printing. The LCD printers are as fast as the DLP printers. The low-cost manufacturing materials used with LCD printers make them more inexpensive, and this is the key reason why these printers are growing in popularity nowadays.

The evaluation of the accuracy of all new technologies is mandatory to have a clinical application. The accuracy is characterized by precision and trueness. Precision corresponds to the evaluation of results compared with each other. Therefore, higher precision means that the machine can give more repeatable and consistent prints. Trueness corresponds to the deviation of each result when compared with the actual object dimensions. Therefore, a machine with great trueness can produce results that are comparable to or identical to the actual dimensions of a digital 3D object. More specifically, in orthodontics, we can now use 3D-printed dental models to fabricate orthodontic appliances. Various orthodontic appliances would be fabricated on the basis of 3D-printed dental models. We can fabricate metal appliances such as Hyrax or lingual arches, but we can also fabricate clear aligners, which are one of the most technique-sensitive orthodontic appliances. Each clear aligner contains small movements of teeth. Part of clear aligner therapy is the placement of the attachments. Attachments act as anchors for the aligners, enabling the teeth to move more efficiently. To transfer the designed attachments to the patient’s mouth, we have to use a template that is fabricated on the basis of 3D-printed models. Therefore, the printing accuracy of 3D-printed dental models will affect the accuracy of orthodontic appliance fabrication and the orthodontic treatment success.

This study aimed to evaluate the accuracy in terms of trueness and precision of 3 different LCD printers presenting different cost levels. The null hypothesis of this study was that there is no difference regarding the accuracy among 3 different in-cost LCD printers.

Material and methods

This study was approved by the ethical committee of the Dental Association of Larisa, Greece. Before the initiation of the study, a power analysis for sample size calculation was performed, which was based on the data from a previous study. ^, The results of the power test suggested that a minimum sample size of 10 subjects will give a confidence level of 95%. Finally, 10 subjects (5 maxillary and 5 mandibular dental arches) were decided to be evaluated for this study.

Scans of maxillary and mandibular arches were used in a standard tessellation language (STL) form. These files were selected retrospectively from a private office database. All scans were made by the same clinician.

According to the inclusion criteria, the selected maxillary and mandibular arches were well aligned to be consistent. In addition, the selected arches should have been free from any other appliances or any orthodontic attachments.

An intraoral scanner was used to scan a set of well-aligned dental arches, and the STL files were chosen from those scans (CS 3700; Carestream Dental LLC, Atlanta, Ga). These STL files were used as reference files and then they were printed using the following three 3D printers: Phrozen Sonic Mighty 4K (Phrozen 3D Tech Co, Ltd, Hsinchu City, Taiwan), Uniz Slash 2 (Uniz Technology LLC, San Diego, Calif), and Flashforge Focus 6K (Zhejiang Flashforge 3D Technology Co, Ltd, City of Industry, Calif) ( Fig 1 ). These 3 printers have different costs, which are defined by 3 different cost tiers in Table I . The cost tier was categorized into 3 categories. Tier 1 was the most expensive, tier 2 was 2 times less expensive than tier 1, and tier 3 was 4 times less expensive than tier 1. To compare 3 LCD technology-based 3D printers in which the brand is irrelevant, the geometric specifications of the object must always be the same. To accomplish this, a number of adjustments must take place. Such involves the light intensity regulation, uniformity of the projected light intensity structure all over the LCD’s surface (commonly known as a light mask), and an accurate z-offset calibration method (especially concerning the z dimension). The calibration method depends on 3 elements. The first concern is finding the correct exposure time for each layer during printing. The second concerns the offset compensation that is added through the corresponding program (software) of each printer, and last but not least, the scale that we also add through the corresponding program (software) of each printer. All 3 elements above are interrelated. This means that they affect each other mainly in terms of the final dimensions of the object but also terms of the purity and quality of the surface of the object in terms of exposure time. By adjusting these elements accordingly, one can eliminate or reduce to the minimum any dimensional difference that is obtained by printing the same object on different LCD-based printers. Thus, all LCD printers were calibrated in the same way. For the consistency of the study, a third-party resin (U-ORTHO almond HD, Unishape, Thessaloniki, Greece) was used. As the 3D printers were calibrated, the printing settings of each printer for the U-ORTHO resin were established. The exposure time for the Phrozen Mighty 4K printer was set at 3.2 seconds, the scale was 100.18%, and the offset compensation was 0.050 mm. The exposure time for Uniz Slash 2 was 2.5 seconds, the scale was 100.67%, and the offset compensation was −0.006 mm. Finally, the exposure time for Flashforge Focus 6K was 2.8 seconds; the scale was 100.59%, and the offset compensation was −0.012 mm. The layer of thickness for all prints was set at 100 μm, and the postcure time was 5 minutes.

Table I

Cost tiers of the 3D printers and manufacturers’ suggested retail price range used in this study

Printer	Flashforge	Uniz	Phrozen
Tier	1	2	3
	$$$$	$$	$
MSRP range	$4900-$7900	$2000-$4000	$580-$780

For the evaluation of trueness, all 10 STL files were printed 1 time with each printer. Each printing template contained 5 dental casts. Once all models were printed, they were scanned with a model scanner (D900; 3Shape, Copenhagen, Denmark). Afterward, each STL file that resulted from a print was compared with the corresponding reference STL file. For the evaluation of precision, one randomly selected STL file was printed 10 times with each of the 3D printers. When all models were printed, they were scanned with the model scanner (D900). Then, all STL files resulting from the prints were compared with the reference STL file. The comparisons of the STL files to evaluate trueness and precision were made using the reverse engineering software (Geomagic Control X 2020.1; 3D Systems, Inc, Rock Hill, SC). The best-fit 3D superimposition technique was applied for this purpose. This technique overlays the STL file onto the reference one using mathematical methods that find the greatest fit. The variation throughout the experimental file is measured compared with the reference cast file after the files have been superimposed. The difference between the reference STL file and the scanned STL file is represented by the root mean square value (RMS), which is the value used to compare the files ( Fig 2 ).

Statistical analysis

In an Excel spreadsheet, all measurements were recorded (Microsoft, Redmond, Wash). SPSS software (version 27; IBM, Armonk, NY) was used for the statistical analysis. Data distribution was examined using the Shapiro-Wilk test. The distribution of all data measurements by the different printers is described in relevant Tables and Figures. The Friedman test was used to evaluate the differences in RMS values between the printers. Statistically significant results were followed by post-hoc tests with Bonferroni corrections. Because the Friedman test was statistically significant, the Wilcoxon test for matched samples was applied as a post-hoc test to compare every printer with the others. Furthermore, the conservative Bonferroni correction method for multiple comparisons was used to adjust the P values.

The same statistical test was used to evaluate each printer’s precision. All tests were 2-sided at the 5% level of statistical significance.

Results

The total sample of this study was 10 dental arches, 5 maxillary, and 5 mandibular dental arches. To test the reliability of the operator with the Geomagic Control X software, intraclass correlation was performed on 5 randomly selected subjects. These subjects were superimposed once to extract the measurements, and then 2 weeks later, the same subjects were remeasured. The intraclass correlation coefficient tests showed no differences between the 2-time measurements, with an excellent correlation that ranged 0.984-0.998 for intraobserver reliability, meaning the operator could not affect the software calculations.

According to the measurements of trueness, the mean RMS value for Flashforge Focus 6K was 0.21 ± 0.04 mm, for Uniz Slash 2 was 0.28 ± 0.05 mm, and for Phrozen Sonic Mighty 4K was 0.30 ± 0.09 mm. Overall, there were significant differences among the 3 printers ( P = 0.007). More specifically, the post-hoc comparison between the Flashforge printer and the Uniz printer was statistically significant ( P = 0.039). The comparison between the Flashforge printer and the Phrozen printer was statistically significant ( P = 0.0015). In contrast, the comparison between the Uniz printer and the Phrozen printer was not statistically significant ( P >0.999) ( Fig 3 ; Table II ).