Multivariate Assessment of Marginal Fit Concerning Two Types of Dental Inlays Obtained Through Two CAD-CAM Technology Scanning Methods-an in vitro Follow-up Study

MARIUS NEGUCIOIU1, DANIELA POPA1*, DANIELA CONDOR2, CRISTIAN CULCITCHI2, DOINA PRODAN3, MARIOARA MOLDOVAN3, SMARANDA BUDURU1 1 Iuliu Hatieganu University of Medicine and Pharmacy, Department of Prosthodontics, 32 Clinicilor Str., 400006, Cluj-Napoca, Romania 2 Iuliu Hatieganu University of Medicine and Pharmacy, Department of Periodontology, 15 Babes Str., 400006, Cluj-Napoca, Romania 3Babes Bolyai University, Raluca Ripan Chemistry Research Institute, Department of Polymer Composites, 30 Fantanele Str., 400294, Cluj-Napoca, Romania

Obtaining a model of good quality that true to its original is extremely important for the success of the treatment; different types of materials and impression techniques have been used over the years to achieve this desired accuracy [1]. Since the introduction of dental CAD/CAM systems, the adaptation of restorations may be improved by combining intraoral scanning and digital milling [2,3].
Marginal fit, or marginal adaptation as it is sometimes referred to, is also crucial for ceramic inlay restorations in order to avoid resin cement wear and plaque accumulation. Marginal gap formation at the toothrestoration interface exposes resin cement to the oral environment, easily leading to cement wear. Marginal discrepancies cause debris and food to act as potential irritants, which might induce gingival inflammation, secondary caries and later on even devitalization of the pulp [4].
The adaptation of a restoration is determined by measuring its marginal and internal gaps, which are important factors for the long-term clinical success [5][6][7][8][9][10].
Intraoral digital scanners have brought dentistry into a full digital era. Performing intraoral scans may increase efficiency due to several factors. Impression trays and materials, which have to be cleaned, disinfected and then shipped to laboratory, are no longer required. The electronic files can be digitally sent and stored, saving time, costs, and improving space management. Distortion as well as volumetric variations related to classic impression materials and cast stone properties are eliminated, simultaneously improving patient acceptance [11,12].
The performance of indirect composite and ceramic restorations, generated either chair-side (i.e. CAD/CAM) or in a dental laboratory, is well documented. There have been significant improvements in software management of function and anatomy, thus explaining their increasing current usage [13].
Despite this, however, studies regarding the accuracy digital systems are still only available in limited amounts [14].
The aim of our follow-up in vitro study was to continue the evaluation and comparison of the marginal adaptation of inlay restorations fabricated by CAD/CAM systems, scanned with two different scanners (the 3Shape TRIOS 3 intraoral scanner and the 3Shape E3 Lab scanner, both produced by 3Shape, Copenhagen, Denmark) and fabricated with two innovative materials (Enamic® hybrid ceramic, Vita, Germany, and IPS Empress CAD ceramic, Ivoclar Vivadent, Liechtenstein).

Experimental part Materials and methods
A preparation for a mesio-occluso-disto-ligual inlay was carried out on tooth 3.7 (according to the IDF classification) of a AG-3 Standard Restorative Typodont (Frasaco, USA) model. The preparation was extended on 4 distinct surfaces in order to obtain a more complex cavity shape, therefore raising the challenge for optimal scanning and milling. The preparation, along with the entire lower arch, the upper arch and the occlusion were first scanned (scan S1) with the TRIOS 3 intraoral scanner (3Shape, Copenhagen, Denmark). The scan files were then sent to the dental laboratory via the local network (Fig. 1). Later on, both arches of the Frasaco Typodont were sent to the dental laboratory in order to be scanned (scan S2) using the 3Shape E3 Lab Scanner (3Shape, Copenhagen, Denmark) (Fig. 2).
From the 2 scans we obtained 2 virtual models (M1 and M2) which were used by the same operator to design 3 virtual inlays for each material using the same design parameters. In the end, 12 inlays were obtained through milling with the Zenotec Select Hybrid milling machine (Wieland Dental + Technik GmbH & Co. KG, Pforzheim, Germany). The details of this stage of the study are summarized in the Study design section below (Flowchart 1). The software that was used for the design was DentalManager Dental System™ 2018 Premium (3Shape, Copenhagen, Denmark).
The Vita Enamic® hybrid ceramic and the IPS Empress CAD ceramic haven been particularly chosen due to their lack of subsequent sintering requirement after milling.
The following is a short technical description of the materials used: -Blocks of glass ceramic in a resin-interpenetrating matrix, shaded 1M1-HT (Enamic®, Vita, Germany). This is composed of a dual network: a feldspathic ceramic network (86% by weight/75% by volume) and a polymer network (14% by weight/25% by volume). The specific composition of the ceramic part is 58% to 63% SiO 2 , 20% to 23% Al 2 O 3 , 9% to 11% Na 2 O, 4% to 6% K 2 O, 0.5% to 2% B 2 O 3 , less than 1% of Zr 2 O and CaO. The polymer network is composed of urethane dimethacrylate (UDMA) and triethylene glycol dimethacr ylate (TEGDMA). The manufacturer refers to this as a hybrid ceramic [15].  . The Frasaco Typodont 3.7 tooth was removed from the phantom arch and 4 distinct areas were marked on each of the prepared surfaces (mesial, distal, occlusal and lingual) using a black permanent marker. Each of the 12 inlays was individually placed on the 3.7 prepared Frasaco phantom tooth and luted with a minimal amount of cement (Temp-Bond™ Clear, KaVo Kerr, California, USA) in order to prevent dislodging due to vacuum conditions. The tooth along with the luted inlay was then inserted and fixed into the FEI Quanta Inspect FP 2017/11 (FEI, Czech Republic) scanning electron microscope and the marginal fit was evaluated on each of the 4 marked points by measuring the gap between the cervical limit of the preparation on the Frasaco tooth and the luted inlay. The entire process was carried out at 1000x magnification whenever possible. After measuring the marginal fit on the pre-marked area, the entire tooth-preparation interface being scanned was examined at lower magnifications in order to detect whether or not other regions would display greater or lesser values of marginal adaptation (however, only discrepancies of more than 30 µm were taken into consideration). If such areas were found, they were also measured and recorded using the 1000x magnification whenever possible. As a general criterion, the commonly acceptable threshold of 120 µm was regarded as a reference for evaluating marginal fit values as either acceptable or not [12].

Results and discussions
The raw results of the measurements carried out on the marked areas and expressed in µm are shown in Table 1.
The raw values that were significantly greater/lesser (i.e. with a minimum of 30 µm discrepancy) than those obtained when measuring on the marked spot of the Frasaco tooth are displayed in Table 2. NF was noted when no such value was found.
Regarding the scanning methods, for intraoral scanning (i.e. when we used the TRIOS 3 intraoral scanner to scan the Typodont) the best marginal fit value was obtained in the case of the Enamic hybrid ceramic on the mesial surface (2.33 microns)( fig. 5), while the worst was in case of the Empress CAD ceramic on the mesial surface (475.89 microns), but not on the marked area ( fig. 6). For model scanning, the best marginal fit was obtained for the Empress ceramic also on the mesial surface (5.83 microns) - Fig. 7, and the worst fit was for the Enamic hybrid ceramic inlay on the occlusal surface but not on the marked area (554.00 microns) ( fig. 8), an extremely high value, the highest recorded in this study.   Regarding the materials, for the Enamic inlay series, the best value found was on the mesial surface at intraoral scanning (2.33 microns) - Fig. 5, and the worst value on the occlusal surface on model scanning (554.00 microns)figure 8. In the case of the Empress inlays, the best value found was with intraoral scanning on the mesial surface (5.83 microns)( fig. 7), and the worst at intraoral scaning on the distal surface (475.89 microns)( fig. 6).
Additionally, as can be seen in Table 1, almost all measurements showed scattered values.
Concerning the scanning method, model scanning seemed to be better than intraoral scanning, displaying a mean marginal fit value of 89.77 µm and a standard deviation of 79.62 µm, meanwhile intraoral scanning only reached a mean of 125.79 µm and a standard deviation of 82.43 µm.
The purpose of this study was to evaluate the effect of using two scanner types and two different materials (Vita Enamic®, Vita, Germany, and IPS Empress CAD, Ivoclar Vivadent, Liechtenstein) on the marginal fit of inlays, therefore obtaining the possibility to assess clinical implications.
Vita Enamic® is the first hybrid dental ceramic in the world with a dual-network structure, that combines the positive characteristics of a ceramic and a composite material. Vita Enamic®'s abrasion properties are similar to enamel, thus it has been reported that its protection of antagonist teeth is greater and a stronger adhesion with the self-adhesive dual-cure resin cement can be achieved due to its composite-like structure. The microstructural analyses showed a hybrid material composed of interconnected networks: a dominant ceramic and a polymer-based matrix [16][17][18].
IPS Empress CAD is a leucite-reinforced feldspathic ceramic indicated for the fabrication of inlays, onlays, veneers, endocrowns, anterior/posterior partial or full crowns. Monochrome IPS Empress CAD blocks have a bending resistance of 160 MPa; it shows outstanding esthetic properties, being available in a variety of shades from A to D, as well as in shades for whitened teeth. These blocks can be HT (high translucency), LT (low translucency) or MO (medium opacity), indicated for masking colored teeth. Polychromatic IPS Empress CAD multi-blocks have both a chameleon effect, as well as a fluorescence similar to neighboring dental structures [19].
Ceramic inlays have many advantages, including great esthetic properties, requirement of less tooth preparation and greater preservation of healthy tooth structures. The clinical process of placing inlays has become more efficient and convenient with the development of CAD/CAM systems [20].
In 2016, Kuhr et al. concluded that digital impression models are similar to conventional models in most anatomical areas, except in secondary areas such as grooves and pits, where gypsum casts can prove to be more accurate [15].
Marginal discrepancy values for direct and indirect inlays were found to be 91. 88 [12]. Our study, even if consisting of a smal lot (only 12 samples), showed an even greater discrepancy, namely a mean value of 108.17 µm and a standard deviation of 81.79 µm on the marked measuring area, thus exhibiting values that exceed the clinically acceptable threshold of 120 µm [12].
We have also found studies that have demonstrated a higher marginal accuracy of restorations derived from an intraoral scanner in comparison to conventional impressions [24].
Inlays generated from conventional wax and resin models tend to show higher marginal discrepancies than conventional digital and full digital patterns. Wax and resin materials yield similar marginal fit accuracies, regardless of impression/manufacturing technique. Better internal fit was shown in wax when compared to resin patterns, irrespective of technique [25].
The use of an intraoral scanner for digital impressions may be a viable alternative to analog techniques. Ferrini et al. concluded that even though PlanScan®, 3D PROGRESS Plus and True Definition Scanner may have showed the best performances, all intraoral scanners tested could provide clinically encouraging results, especially in terms of marginal accuracy, since mean marginal gap values were all within the clinically acceptable threshold of 120 µm [12]. However, regarding our study, marginal fit values above this threshold were 36 out of 71 (taking both marked and non-marked areas into account), which represented a 50.7% ratio that was considerably high.
Furthermore, our data was not very conclusive. For example, when analyzing the marginal fit of the materials, Empress CAD displayed the lowest mean value (96.20 µm), but it also presented the highest discrepancies recorded in this study (554.00 µm on the occlusal aspect and 475.89 µm on the mesial surface -however, both recorded on non-marked areas of the 3.7 Typodont tooth).