Introduction

Nowadays, digital technologies are increasingly used in dentistry1, especially in the fabrication of single crowns or short-span fixed partial dentures. Intraoral scanners combined with computer‐aided design and computer‐aided manufacturing (CAD-CAM) techniques offer an alternative pathway for the fabrication of oral prostheses, which will improve the communication with technicians, facilitate the prostheses design, make storage of clinical data easier, reduce time consumption, and increase comfort level of patients2,3,4. The accuracy (precision and trueness) of intraoral digital impression in short-span scan has been proved to be similar or even better than conventional impression5,6,7,8,9, while the accuracy tends to decrease with expanded scanning range10,11,12. In terms of RPDs, it is more complicated because digital impressions should cover a wider range of dentitions as well as soft tissues that touch with major or minor connectors. However, few studies have been carried out to validate the accuracy of soft tissues in vivo13. In our previous study, the feasibility for scanning the entire maxillae of volunteers with intact dentitions by intraoral scanner had been proved14, and the adaptation of major connectors manufactured subsequently by digital technique were clinically acceptable15. To date, some studies have preliminarily verified the accuracy of intraoral scanning on partially edentulous models through testing multiple scanners by using the method of best-fit alignment between intraoral and reference impressions. Most of them turned out that the overall accuracy of intraoral scanners was less than 100 μm in vitro16,17,18,19,20, and the scanning trueness of residual ridges is satisfactory21. However, the existing in vitro studies still did not fully simulate the oral environment, and did not consider whether different types of dentition defects, maxillary anatomy such as arch width and palatal height would affect the trueness of intraoral scanning.

The purpose of this in vivo study was to compare the trueness of intraoral digital impression for dentitions, palatal soft tissues, and edentulous areas, as well as to determine the effect of arch width, palatal vault height, and different types of dentition defects and numbers of missing teeth on the trueness of intraoral digital impressions in the oral environment. The null hypotheses stated that: (1) It was feasible to use the intraoral scanner to obtain digital impressions of partially edentulous maxillae; (2) There were no significant differences between the intraoral scanning trueness of dentitions, edentulous areas, and the palatal regions. (3) The arch width, palatal vault height or different types and numbers of teeth loss did not influence the trueness of intraoral digital impressions.

Materials and methods

Participants

Patients who were maxillary partially edentulous and chose RPDs as treatment were recruited in this study. The study was registered in the Chinese Clinical Trial Registry (Registration No. ChiCTR-ONC-16009899; Registration name: The clinical evaluation of digital maxillary removable partial denture fabricated by intra-oral scanning and 3D printed rapid prototyping; Registration date: 2016/11/17). The study was approved by the Independent Ethics Committee of Shanghai Ninth People’s Hospital affiliated to Shanghai Jiao Tong University, School of Medicine (Application No.2016–120-T69; Approval date: 2016/10/26). It was conducted in accordance with the Declaration of Helsinki. Informed consents were obtained from all participants.

Inclusion criteria: (1) Aged 18 to 75 years old, no restrictions for gender; (2) Good oral hygiene; (3) No teeth to be further extracted; (4) No maxillary hard and soft tissue defects; (5) Good cooperation to finish all data acquisition. Exclusion criteria: (1) Undergoing orthodontic treatment; (2) Being allergic to denture materials; (3) Uncontrolled periodontitis, mucosal diseases, and cariology; (4) Having severe systematic diseases; (5) Obvious teeth mobility (mobility degree higher than I°), residual roots, and untreated decayed teeth; (6) Patients who were unable to take care of themselves and have risks to swallow the denture by mistake.

The acquisition of digital and conventional impressions

The whole data acquisition procedure is shown in Fig. 1. Trios 3 (3Shape, Denmark) was used to acquire intraoral digital impressions. Saliva was removed from the surface of the teeth and soft tissues with cotton rolls and air syringe before the scan began. The scanning process started from the molars on the side with more remaining posterior teeth, and then move to the other side. The images of teeth were captured following the sequence of occlusal surface, palatal side, and then buccal side. Palatal scanning was performed by continuously narrowing down the scope in an inverted U manner which was recommended in Yang’s study (Fig. 2)22. The data were converted into STL file format at last, defined as DI. All intraoral scans were performed by one experienced researcher once for each patient. The scanning time (excluding connecting and calibrating time) was recorded.

Fig. 1
figure 1

Flow diagram of the study.

Full size image
Fig. 2
figure 2

Scanning sequence of intraoral scanner.

Full size image

At the first appointment, alginate impression (Orthoprint, Zhermack, Italy) was made and poured into plaster cast (uni-base 300, Dentona, Germany) for the fabrication of individual tray (light-curing resin, HUGE Dental, China). After the digital impression was completed during the second visit, the individual tray border refining was finished with a compound material (Shanghai Dental Material Factory, China). Tray adhesive (DMG, Germany) was applied on individual tray and conventional impression procedure was performed with polyether rubber (Impregum Penta Soft, 3 M ESPE, USA). The final impression was poured with type IV gypsum and stored at room temperature. The edge border molding plus polyether rubber impression time was recorded. After at least 96 h until the expansion of gypsum was completed, the cast was digitized by a laboratory scanner (d-Station 3D scanner, Breuckmann, Denmark). Finally, these data were converted into STL formats and defined as CI.

Measurement of arch width and palatal vault height

An alginate impression of upper jaw was taken after the patient wore the definitive RPD. So it was possible to measure the arch width and palatal vault height in existence of dentition defect by using “Distance Measurement” tool in Geomagic Studio 2013 (V.2013.0.1.1206, Geomagic, USA). Arch width was defined as the distance in a straight line between the mesiobuccal cusp tips of both maxillary first molars23. Palatal vault height was denoted as the vertical distance from the connection line between mesiolingual cusps of both maxillary first molars to the median line of the palate14.

Grouping

The patients were classified according to: (1) Arch width (Group Narrow, Medium, and Wide); (2) Palatal vault height (Group Low, Medium, and High); (3) Types of dentition defects. The patients were divided into Kennedy class I, II, III, and IV according to Kennedy classification and were divided into Group Tooth Borne and Group Extension according to Kratochvil classification; (4) Numbers of missing teeth. The patients who had less than five (including five) missing teeth were divided into Group A and more than six (including six) were divided into Group B; (5) Distribution of edentulous areas was classified by the number of separated edentulous areas (Group 1, 2, 3 and 4); (6) Numbers of remaining molars (Group 0, 1, 2, 3, and 4).

Data processing

All STL files were imported into Geomagic Studio 2013. “Three-point Registration” function was used for preliminary alignment, and “Boundary from Spline” was applied to further trim the model into dentition, edentulous areas, and palate. The main steps were as follows: (1) Dentition was trimmed along the spline boundary creating by the gingival margin of remaining teeth; (2) The distal end was set at the connection line of distal gingival margin of both second molars (non-free-end dentition defect), or it was estimated by the mirror image of the contralateral second molar (unilateral free-end dentition defect) or 5 mm before hamular notch (free-end dentition defect); (3) Edentulous area is determined within the range of 6 mm-width on each side of the alveolar ridge crest line; (4) The above boundaries around palates constitute the range of the palatal soft tissues. After that, “Best-fit Alignment” was activated for each part of the model as the final superimposition and registration. Finally, the deviation information and color-coded deviation maps was generated. The average absolute deviation, on behalf of the trueness, was calculated as (|average positive deviation| +|average negative deviation|)/214.

Statistical analysis

The sample size of this study was calculated based on the experimental result of Lee et al.17. According to the formula \(\text{n}=\frac{{\Psi }^{2}(\sum {S}_{i}^{2}/k)}{\sum ({\overline{X} }_{i}-\overline{X })(k-1)}\), the sample size met the demand of ANOVA under the circumstances of α = 0.05, power = 0.90, and k = 4, which required at least 36 participants. The sample size was calculated as 48 participants at least when taking account of the dropout rate. The sample size was estimated by a sample size calculation software (PASS 11, NCSS, USA).

Statistical analysis was performed by SPSS V.23.0 (IBM, USA). Normal distribution analysis was carried out using the Kolmogorov–Smirnov test and homogeneity of variance was analyzed by Levene’s test. It turned out that the absolute deviations were not normally distributed. For each group classification, the median and interquartile range were calculated. Pearson correlation test was used to analyze the correlation between the intraoral scanning time and the trueness of entire scanning. Friedman’s test was performed to analyze the differences between the scanning trueness of palatal soft tissues, dentitions, and edentulous areas. Kruskal–Wallis test was used to analyze the effect of different arch width, palatal vault height, and different types of dentition defects and numbers of missing teeth on the trueness of intraoral digital impressions. LSD test was applied for further comparison in pairs. The statistical significance was set at 0.05.

Results

Patient information

The collection of patient data in this clinical trial began in August 2018 and ended in September 2022. A total of 66 patients finally signed the informed consent forms to participate in this experiment, in which 2 patients were eliminated because of errors occurred in the molding process. 39 patients were male (60.9%), and 25 patients were female (39.1%). The age distribution ranged from 44 to 75 years old with the average age of (63.1 ± 9.0) years old. The classification of arch width and palatal vault height were shown in Tables 1 and 2.

Table 1 Classification of arch width.
Full size table
Table 2 Classification of palatal vault height.
Full size table

Scanning time

The average time for intraoral scanning of 64 patients was (6.38 ± 2.77) min, while it was (12.67 ± 4.83) min for the edge border molding plus polyether rubber impression time. The intraoral scanning time was much shorter than that of conventional impression with a significant difference (P < 0.001). No significant correlation was showed between the intraoral scanning time and the trueness of entire maxilla scanning (P = 0.386).

Scanning trueness of palatal soft tissues, edentulous areas, and dentitions

The scanning trueness of palatal soft tissues, alveolar ridges of edentulous areas, and dentitions was shown in Table 3 and Fig. 3. The results presented significant differences in the average absolute deviations among teeth, palate, edentulous areas, and entire maxilla scans (P < 0.001). Post-hoc analysis showed significant statistical differences in the scanning trueness between dentition and palate, dentition and entire maxilla, edentulous area and palate, and edentulous area and entire maxilla, respectively (P < 0.001).

Table 3 Trueness of intraoral digital impression (μm).
Full size table
Fig.3
figure 3

Trueness of intraoral digital impression of dentition, palate, edentulous area, and entire maxilla. The black dots represent outliers, and “ + ” denotes the mean value. ***P < 0.001.

Full size image

Deviation chromatogram was shown in Fig. 4. The deviation of dentition was relatively high in the interproximal space, the adjacent surfaces of the teeth close to the edentulous areas, deep fossae of the posterior teeth, and the incisal edge of the anterior teeth. The deviation distribution of palatal region had the following characteristics: (1) Positive deviations were observed on both sides of palatal rugae and palatal vault, while negative deviations were detected at the middle palatal suture; (2) The maximum error usually appeared at both sides of the palatal vault, corresponding to the position between the maxillary second premolar and the second molar. The maximum deviation was about 500 μm in most patients, while it reached 800 ~ 1000 μm among the other seven patients; (3) There were also positive deviations at the incisive papilla and palatal rugae, but the deviations were often less than 500 μm. Positive deviations of the edentulous areas mainly distributed at the buccal alveolar mucosa, while the deviations at palatal sides were relatively small.

Fig.4
figure 4

Color-coded maps of whole maxilla.

Full size image

Effect of arch width, palatal vault height, and dentition defects on digital impression

Scanning trueness of Group Narrow, Middle, and Wide in arch width was shown in Table 4. There were no significant differences between average absolute deviations of dentitions, palates, and edentulous areas of patients with different arch width (P > 0.05).

Table 4 Effect of arch width on scanning trueness (median ± interquartile range) (μm).
Full size table

Scanning trueness of Group Low, Middle, and High in palatal vault height was shown in Table 5. There was no significant difference in average absolute deviations of dentitions and palates of patients with different palatal vault height. The scanning trueness of edentulous areas in Group Low was significantly better than that in Group Middle (P = 0.021) and Group High (P = 0.006), while no significant difference was noted between Group Middle and High (P = 1.000).

Table 5 Effect of palatal vault height on scanning trueness (median ± interquartile range) (μm).
Full size table

Scanning trueness of each Kennedy classification and Kratochvil classification were shown in Tables 6 and 7, respectively. Kennedy classification and Kratochvil classification had no significant effect on the average absolute deviations of dentitions, palates, and edentulous areas (P > 0.05).

Table 6 Effect of Kennedy classification on scanning trueness (median ± interquartile range) (μm).
Full size table
Table 7 Effect of Kratochvil classification on the scanning trueness (median ± interquartile range) (μm).
Full size table

The average absolute deviations of different number and distribution of missing teeth were shown in Tables 8 and 9. The number and distribution of missing teeth had no significant effects on the average absolute deviations of dentitions, palates, and edentulous areas. The average absolute deviations of edentulous areas became larger with the increased number of missing teeth, however, the result was not statistically significant (P > 0.05). In addition, the average absolute deviations of patients with different number of remaining molars were shown in Table 10.

Table 8 Effect of number of missing teeth on scanning trueness (median ± interquartile range) (μm).
Full size table
Table 9 Effect of distribution of missing teeth on scanning trueness (median ± interquartile range) (μm).
Full size table
Table 10 Effect of number of remaining molars on scanning trueness (mean ± standard deviations) (μm).
Full size table

Discussion

The null hypothesis was partially accepted. It was feasible to use the intraoral scanner to obtain digital impressions of partially edentulous upper jaws, although there were significant differences between the scanning trueness of dentitions, edentulous areas, and palates. Furthermore, the arch width, as well as the number and the type of missing teeth had no significant effects on scanning trueness, while there was a significant effect of palatal vault height on the scanning trueness of edentulous areas.

In this study, deviation chromatogram showed the differences between two aligned models (intraoral digital impression and conventional impression). In terms of the palatal region, positive deviations were observed at palatal rugae, both sides of the palatal vault, and the posterior margin of the palate. Mucosae at these regions are flexible and will be oppressed when polyether rubber are pressed onto, whereas intraoral digital scan is a kind of anatomical impression and has no pressure on the oral mucosae. So, the intraoral digital impression was above the conventional impression, showing positive deviations at palatal rugae, two sides of the palatal vault, and the posterior margin of the palate. On the other hand, obvious negative deviations were presented around the mid-palatal suture. Mucosae at the mid-palatal suture are thin and lack of flexibility. And there were more difficulties for intraoral scanning head to parallel to the mucosa surface and capture the accurate images of the smooth surface around the mid-palatal suture14. These might be the reason that the intraoral digital impression was below the conventional impression with negative deviations in these areas. When considering the deviation chromatogram of dentitions, increased deviations were presented at the adjacent surfaces of the teeth which close to the edentulous areas. Although tooth preparation for proximal plate was carried out, scattered deviations still remained in small undercuts. Moreover, the buccal parts of edentulous areas had larger mucosal mobility and tended to be oppressed when making conventional impressions, thus resulting in more positive deviations. This also explained that relatively smaller deviations were seen at the palatal mucosae of edentulous areas because of their less mobility. The phenomenon of the positive and negative deviation distribution was basically in agreement with the in vivo results revealed by Gan et al.14 and Yang et al.22.

In terms of the intraoral scanning trueness of intact dentition, Yang et al.22, Lim et al.24, and Gan et al.14 reported the results of (64.92 ± 19.50) μm, (79.8 ± 13.6) μm, and (80.01 ± 17.78) μm in vivo, respectively. When it comes to the scanning trueness of the dentition in maxillary partially edentulous models, however, it was revealed less than 50 μm in most previous studies17,19,20,25. So, oral environment did greatly affect the scanning trueness. The scanning trueness of palatal areas ranged from (108.25 ± 33.83) μm to (130.54 ± 33.95) μm14,22,26 in other researches. It’s worth noting that misfits (stratification) were prone to occur in palatal soft tissues when scanning patients with multiple posterior dentition defects. Thus, the effect of the number of remaining molars on the scanning trueness of palatal regions was further analyzed. The results showed that the edentulous areas and palatal scanning trueness of the patients with more remaining molars tended to be better. So multiple molars loss might influence the trueness of maxillary scanning to some extent, but the difference was not statistically significant. Lack of clear anatomical landmarks and differentiated structures led to difficulties and errors in image stitching during our primary scanning, so we chose the side with more remaining posterior teeth and more clinical landmarks to start the scanning. Hayama et al. pointed out that scanning trueness of the edentulous areas was 54 ~ 107 μm in partially edentulous models18. Lee et al. found the scanning trueness of the partially edentulous model with only two missing teeth was close to that of complete dentition, while the scanning trueness of the model with more than five missing teeth was significantly lower17. However, no significant difference was observed between the scanning trueness in patients with less than five (including five) missing teeth and with more than five missing teeth in this study, which might be due to the interference of oral environment.

There is a flexibility diversity of mucosae at different regions of palate. Thereby, it is quite complicated to analyze the effect of palatal scanning trueness on RPDs fabrication for the fact that RPDs frameworks may cover a wider range of palatal soft tissues in partially edentulous upper jaws. As we know, the posterior palatal bar is placed between the hardened zone and vibrating line of the upper jaw27, so the trueness of this area should be taken into account. However, there has not been a universal standard for determining the distal end of the maxilla. The distal end was set at the distal gingival margin of second molars14,26 or at the mid-gingival margin of second molars22. In Yang’s study, the researchers believed that the movement at the junction of the soft and hard palate was obvious, and the subsidence at this area in conventional impression would lead to larger deviations22. The relatively larger scanning deviations in this area might have an impact on the accuracy of major connector of the definitive RPD. The affordable inaccuracy is critical for further application of intraoral scanner in RPD. It turned out that the overall accuracy of intraoral scanners was less than 100 μm in most vitro studies16,17,18,19,20. For all we know, the scanning accuracy in vivo is usually slightly worse than in vitro result due to the intraoral environment, including the impact of resilient mucosae. Lytle et al. proposed the mucosal subsidence in the distal edentulous areas of removable denture could reach 300 μm28, while Miyashita et al. reported the deformation of mucosal tissues of 0.7 ~ 1.0 mm29. The maximum average absolute deviation was 230 μm for palatal scan in this study, which was within above range. However, in recent in vitro studies, the Co-Cr framework adaptation was about 300 μm30, with discrepancies within clinically acceptable values (50 to 426.3 μm) regardless of the technique (conventional or digital removable partial denture) 31. It was also reported that a gap (between major connector and palatal tissues) ranging from 0 to 50 μm was considered close contact (no gap), and a gap ranging from 50 to 311 μm was defined as a clinically acceptable fit32,33. Therefore, the clinically affordable inaccuracy of intraoral scanning still needed to be checked by additional clinical studies.

In clinical practice, altered cast impression is usually used to reduce torque on abutment teeth by solving the difference between the mobility of abutment teeth and alveolar ridge mucosae under occlusal function in free-end partially edentulous patients. Intraoral scanning is a static impression, so there is no oppression on mucosae during impression acquisition. Mohammed et al. pointed out no significant advantage in altered cast impression technique over conventional single-impression techniques was found in existing studies34. In our study, there was also no significant difference between the trueness of Group Tooth Borne and Group Extension in edentulous areas according to Kratochvil classification. Therefore, altered cast impression might not be necessary for distal extension partially edentulous patients, which needed to be further explored in subsequent experiments focusing on the adaptation of accomplished RPDs.

Limitations of the present study included the number of patients in different Kennedy classifications was uneven in this clinical trial, because most patients of Kennedy class III and IV tended to choose fixed partial dentures or implant fixed partial dentures. Therefore, we added Kratochvil classification, i.e., distal extension and tooth borne, to minimize the possible errors caused by sample bias when considering the clinical acceptance. In addition, further studies focusing on definitive RPDs are also needed to validate whether the trueness of intraoral scanner is clinically acceptable by checking the adaptation of prostheses intraorally.

Conclusions

Despite the limitations of this study, it was basically feasible to use intraoral digital impressions for the fabrication of RPDs. The median trueness for complete-arch intraoral scanning of partially edentulous upper jaws was less than 120 μm in vivo. There were significant differences between scanning trueness of dentitions, edentulous areas, and palate regions. Larger palatal vault height increased the deviations of intraoral scanning, while arch width and type of missing teeth had no significant effect.