Journal of Prosthetic Dentistry最新文献

Statement of problem: Cone beam computed tomography (CBCT) scans acquired with interocclusal separation improve stability and reduce motion artifacts but lack reliable occlusal contacts. Accurately reconstructing occlusion is critical for diagnosis, restorative planning, and implant workflows. While integrating intraoral scans (IOSs) can capture occlusion accurately, the validity of artificial intelligence (AI)-driven alignment algorithms for reconstructing occlusion in separated CBCT datasets has not been established.

Purpose: The purpose of this retrospective study was to validate an AI-based tool for reconstructing occlusion by aligning separated maxillary and mandibular CBCT segmentations using IOS-derived occlusal data, with IOS occlusion as the reference standard.

Material and methods: Forty paired CBCT scans acquired with interocclusal separation and corresponding IOS datasets were uploaded into an AI-driven platform, which automatically segmented and registered the CBCT and IOS scans. The AI tool used IOS‑derived occlusal relationships to align the segmented CBCT models, generating 3 occluded models: IOS‑only, AI-fused CBCT‑IOS, and CBCT‑only. Occlusal contacts, occlusal intersections, occlusal contact surface area, and 3-dimensional (3D) occlusal surface deviations were analyzed. Statistical analyses were performed using the Friedman test, repeated-measures ANOVA, and the Mann-Whitney U test. Intra-operator reliability was assessed using weighted kappa (κ) (α=.05).

Results: No significant differences were observed between IOS and AI-fused CBCT-IOS models for occlusal contacts or contact surface area (median=46, mean ±standard deviation=256 ±170 mm²) and (median=48, mean=255 ±166 mm²), respectively. AI-driven CBCT models showed significantly fewer contacts (median=22, mean ±standard deviation=184 ±163 mm²). Median surface deviation between IOS and fused models was 0 µm, whereas CBCT-only models showed deviations of 70 to 80 µm. Dentition status did not influence outcomes. Intra-operator reliability was excellent (κ=0.86).

Conclusions: The AI-driven fusion of IOS occlusal data with CBCT scans acquired with interocclusal separation accurately reconstructs occlusion, enabling reliable treatment planning without the need for CBCT acquisition in occlusion.

Statement of problem: Although cracked teeth have been a prevalent clinical concern, a consensus on appropriate treatment strategies-particularly for early-stage cracks without clinical symptoms-remains lacking. Current guidelines do not specify the critical crack dimensions, in terms of depth and width, at which tooth structural integrity becomes significantly compromised, leading to uncertainty in clinical decision-making between monitoring and restorative intervention.

Purpose: The purpose of this study was to investigate the effects of crack depth and width on the ultimate strength and stress distribution of human first premolars to guide clinical treatment decisions.

Material and methods: Forty extracted, sound human first premolars were divided into 5 groups: a control group (no crack) and 4 groups with experimental mesio-occluso-distal cracks of varying depth (D:2 to 4 mm) and width (W:0.5 to 1 mm). Compression tests were conducted to assess ultimate strength, fracture origin, and propagation direction. Statistical analyses were performed using 1-way ANOVA and multiple linear regression (α=.05). Finite element analysis (FEA) was used to simulate stress distribution based on both principal stress and energy-based failure theories. The computational results were validated with experimental findings. Smaller cracks (D:1 to 3 mm; W:0.1 to 0.5 mm) were modeled via FEA to identify subclinical critical thresholds.

Results: Crack depth dominated tooth ultimate strength reduction (β=-.803, P<.001), while width became insignificant at a depth of 4 mm (β=-.059, P=.480). Fractures in cracked teeth originated at crack tips (91% of cases), whereas fractures in sound teeth occurred at the palatal cusp. The von Mises stress criterion accurately predicted the behavior of cracked teeth, unlike the principal stress theory. Critical crack size thresholds were identified: a 15% strength reduction in the D1W0.5 and D2W0.1 cases and a 67% strength reduction (relative to sound teeth) when cracks extended to the pulp chamber in the D3W1 case.

Conclusions: Increasing crack size significantly weakens tooth strength, with depth being the dominant factor. The von Mises stress theory is recommended for analyzing cracked teeth. The identified critical crack sizes provide actionable thresholds for restorative intervention.

Statement of problem: The metal-ceramic bond strength of selective laser melting (SLM) Ti-6Al-4V alloys is critical to the longevity of dental porcelain restorations. However, achieving a reliably strong metal-ceramic bond strength of SLM Ti-6Al-4V alloys remains challenging.

Purpose: The purpose of this in vitro study was to investigate the effect of an as-built surface on the metal-ceramic bond properties of SLM Ti-6Al-4V alloys.

Material and methods: Plate specimens (25×3×0.5 mm) were fabricated via SLM and divided into 2 groups (n=8): one without airborne-particle abrasion (WAPA) and the other with airborne-particle abrasion (APA) using 110-µm Al₂O₃ particles. Surface morphology and roughness were characterized using scanning electron microscopy (SEM) and confocal laser scanning microscopy. Veneering ceramic was applied by following the manufacturer's instructions. After specimen preparation, the metal-ceramic bond strength was assessed via 3-point bend tests. Elemental composition and interfacial microstructure were analyzed using energy-dispersive X-ray spectroscopy (EDS). X-ray diffraction (XRD) was used to analyze the phase compositions of 2 groups. All data were analyzed using Student t tests (α=.05).

Results: The as-built surface exhibited a high roughness along with numerous molten hemispherical structures. In contrast, APA reduced the surface roughness and eliminated the molten features. The 3-point bend tests results revealed that the WAPA group exhibited significantly higher bond strength (36.25 ±2.34 MPa) than the APA group (31.07 ±2.24 MPa) (P<.001). Debonded surface analysis revealed a mixed failure mode in the WAPA group with approximately 62% ceramic retention, while the APA group exhibited cohesive failure with 93% retention. Interfacial characterization confirmed the presence of a continuous TiO₂ oxide layer at the metal-ceramic interface, with a thinner native oxide layer observed in the WAPA group (1.1 µm) compared with that of the APA group (1.9 µm). Metallographic analysis and XRD results verified that both groups shared similar phase compositions, predominantly consisting of α/α' phases and β phases. Additionally, EDS showed a narrower elemental diffusion zone in the WAPA group (2.5 µm) relative to the APA group (3.3 µm).

Conclusions: The as-built surface of SLM Ti-6Al-4V alloys exhibited higher metal-ceramic bonding performance than the APA surface. This enhancement was attributed to the better mechanical locking provided by the molten hemispherical structures, the optimized chemical bonding facilitated by a stable oxide layer, and an increased effective interface contact area.

Statement of problem: Limited studies have assessed the accuracy of noncalibrated implant scan body (ISB) systems for recording implant scans, reporting a reliable method for capturing implant positions. A recently introduced commercial noncalibrated ISB method is available, yet its accuracy remains unknown.

Purpose: The purpose of this in vitro study was to evaluate the accuracy of complete arch implant scans obtained using a noncalibrated ISB system and 5 intraoral scanners (IOSs).

Material and methods: A stone cast with 6 implant abutment analogs was selected. A laboratory scan was recorded (T710) to obtain the control file. Five groups were described depending on the IOS: Lumina, TRIOS5, Primescan, i700, and i900. Two subgroups were created based on the technique tested: standard (ISB subgroup) and noncalibrated (NC-ISB subgroup) ISBs (n=15). In the ISB subgroup, an ISB (AB-SR-01 IPD) was tightened into each implant abutment, and scans were captured with the IOSs tested. In the NC-ISB subgroup, a noncalibrated ISB (ScanlogiQ) was hand tightened into each implant abutment toward the center of the arch. Implant scans were recorded involving the pyramid geometry of the noncalibrated ISBs using the IOSs tested. Euclidean linear and angular calculations were obtained for the 6 implants. The measurements acquired in the control scan were used to measure scanning discrepancies with each experimental scan. Two-way ANOVA and the Tukey tests were used to analyze trueness. The Levene test was used to analyze precision (α=.05).

Results: Linear trueness discrepancies were found among the groups (P<.001) and subgroups (P<.001), with a significant group×subgroup interaction (P<.001). The Lumina and i900 obtained the best linear trueness. The i700 and TRIOS5 had the highest mean linear discrepancy. The NC-ISB group had better linear trueness than the ISB group (P=.05). The Levene test revealed significant linear precision discrepancies among the groups (P<.001) and subgroups tested (P<.001). The Lumina and i900 had the best linear precision, while the i700 had the worst linear precision. The NC-ISB group showed better linear precision than the ISB group. Additionally, angular trueness discrepancies were revealed among the groups (P<.001), with a significant group×subgroup interaction (P<.001). The Lumina had the best angular trueness, while the i900 presented the worst angular trueness. The Levene test revealed significant angular precision discrepancies among the groups (P<.001) and subgroups tested (P<.001). The Primescan and Lumina had the best angular precision, while the i900 had the worst angular precision. Lastly, the NC-ISB group showed better angular precision than the ISB group.

Conclusions: The implant scanning technique and IOS influenced the trueness and precision of the complete arch implant scans.

Statement of problem: The accuracy of complete arch implant scans acquired using intraoral scanners (IOSs) has been extensively examined; however, these evaluations often disregard the influence of the scanning technique used to capture implant positions.

Purpose: The purpose of this systematic review was to assess the trueness and precision of complete arch implant scans recorded using nonsplinting techniques.

Material and methods: A literature search was completed in 5 databases: PubMed/Medline, Scopus, Embase, Web of Science, and Cochrane. A manual search was also conducted. Studies analyzing the accuracy of complete arch scans using commercially available nonsplinting techniques recorded with IOSs were included. Two investigators evaluated the studies independently by applying the Joanna Briggs Institute critical appraisal. A third examiner was consulted to resolve any lack of consensus. The mean and standard deviation of the accuracy values reported were extracted, and a meta-analysis was performed. The mean difference between the analog and nonsplinting groups was calculated using the random effect model (a=.05). The I-squared (I²) statistic and its associated P value were used to assess the heterogeneity between studies. Publication bias was assessed using a funnel plot. The Egger test was used to determine the significance of the funnel plots.

Results: A total of 100 articles were included. The linear discrepancy ranged from 22 to 1050 µm, with a mean linear discrepancy of 117 µm. The angular discrepancy values ranged from 0.01 to 1.75 degrees, with a mean angular discrepancy of 0.48 degrees. The reported linear trueness ranged from 41 to 557 µm, with a mean linear trueness of 153 µm. The linear precision ranged from 7 to 166 µm, with a mean linear precision of 57 µm. The angular trueness ranged from 0.20 to 1.69 degrees, with a mean angular trueness of 0.65 degrees. The angular precision ranged from 0.05 to 1.69 degrees, with a mean angular precision of 0.41 degrees. The mean RMS discrepancy ranged from 9 to 408 µm, with a mean RMS error of 101 µm. The RMS trueness ranged from 27 to 366 µm, with a mean RMS error of 60 µm. Lastly, the RMS precision ranged from 5 to 251 µm, with a mean RMS error of 35 µm. The funnel plots and Egger regression asymmetry tests revealed significant publication bias (P<.001).

Conclusions: Nonsplinting implant scanning techniques demonstrated high variability in the scanning accuracy outcomes. Further clinical studies are needed to evaluate the accuracy of implant scans, as substantial heterogeneity was observed among the included studies.

Statement of problem: The conventional workflow of computer-assisted implant surgery necessitates the placement of mucosa-embedded registration pins; these have limitations in patients receiving vascularized bone grafts (VBGs) for jaw reconstruction.

Purpose: The purpose of this clinical study was to evaluate and compare implant accuracy using registration pin (RP) and titanium screw (TS) registration in dynamic navigated implant placement after reconstruction of the jawbone with VBGs.

Material and methods: Patients who underwent the vascularized bone flap repair of jaw defects followed by application of dynamic navigation for implant placement in VBGs from March 2022 to May 2025 at Peking University school and hospital of stomatology were included. Accuracy was assessed by comparing preoperative design and postoperative cone beam computed tomography (CBCT) scans. Measurements were made to compare angulation deviation, overall deviation of platform and apex, and bidirectional deviation between groups. Group comparisons were performed using appropriate parametric or nonparametric tests and generalized estimating equations (GEEs) were used to assess the effects of registration method, jaw, implant position, and deviation direction on positional errors.

Results: A total of 17 participants were included, 9 (34 implants) in the RP group and 8 (28 implants) in the TS group. Compared with the RP group, no significant difference in angular deviation and overall deviation was found in the TS group (RP versus TS: 2.88 ±2.14 degree versus 3.02 ±3.71 degree, P=.771; 1.13 ±0.68 mm versus 0.93 ±0.44 mm at platform, P=.656; 1.04 ±0.59 mm versus 1.14 ±0.76 mm at apex, P=.493). In addition, for maxillary implants, the overall deviation at the platform was smaller in the TS group than in the RP group (P=.028).

Conclusions: Dynamic navigation for implant placement in VBG reconstructed jaws using TSs alignment achieved implant accuracy comparable with that of RPs.

Statement of problem: The mechanical properties and biocompatibility of 3-dimensionally (3D) printed dental ceramics require further investigation to provide a theoretical basis for clinical application.

Purpose: The purpose of this in vitro study was to evaluate and compare the mechanical properties and biocompatibility of zirconia fabricated by 3D printing and conventional computer numerical control (CNC) milling.

Material and methods: Zirconia specimens were fabricated by 3D printing and CNC milling. The crystalline phase (CP) and surface roughness (SR) were assessed by X-ray diffraction (XRD) and scanning electron microscopy (SEM) (n=5). Flexural strength (FS), fracture toughness (FT), and Vickers hardness (HV) were examined with a universal testing machine and Vickers microhardness tester (n=5). Cell viability was assessed by using the CCK-8 assay and bone marrow stromal cells (BMSCs), with 5 replicate wells per group. The mechanical property tests and CCK-8 assay were independently repeated 3 times. In vivo biocompatibility was assessed by acute oral toxicity testing in Sprague-Dawley rats (n=5) and mucosal irritation evaluation in golden hamsters (n=10). Data are presented as mean ±standard deviation (t test or 1-way analysis of variance) or median (nonparametric tests) (α=.05).

Results: Three-dimensionally printed zirconia possessed a rougher surface with a microporous structure compared with CNC milled zirconia. Both ceramics were primarily in the tetragonal phase, with the 3D printing zirconia exhibiting broader and lower-intensity peaks. The CNC milled zirconia demonstrated significantly higher FS than the 3D printed zirconia (P<.001). The 3D printed ceramics showed better mechanical properties in FT (P=.029) and Vickers hardness (P=.002). No statistically significant differences in cell viability were observed among groups at 3 time points (P>.05). No meaningful systemic toxicity or mucosal irritation was observed in short-term animal models.

Conclusions: Three-dimensionally printed zirconia ceramics exhibit better toughness, hardness, and short-time biosafety. The flexural strength preliminarily meets the clinical requirements for low-stress restorations, providing a theoretical foundation for the future clinical application of 3D printed zirconia restorations. Further research and optimization of manufacturing processes are needed to meet the demands of high stress and long-term clinical application.

Statement of problem: Systematic reviews (SRs) are time-consuming and resource-intensive processes. Whether large language models (LLMs) can improve the process is unclear.

Purpose: The purpose of this study was to evaluate the accuracy and reliability of 4 LLMs (GPT-4, Gemini, Claude, and Elicit) in performing SR tasks (full-text screening, data extraction, and risk of bias assessment) at 3 different sequential periods of time (0, 15, and 30 days).

Material and methods: A comprehensive systematic search was conducted across 5 databases in December 2024, with 59 articles evaluated for screening (2 used for pilot) and 31 for data extraction (2 used for pilot). A 3-pronged prompting strategy was used, including persona-based initialization, few-shot learning, and structured population, intervention, control, outcome (PICO) criteria. Performance was assessed through 3 repeated evaluations at 2-week intervals by measuring accuracy and reliability using standard metrics (accuracy, precision, F1-score, sensitivity, and specificity) against expert assessments, data extraction quality on a 0 to 5 scale, and risk of bias agreement via the Cohen kappa, with statistical analysis using Kruskal-Wallis and Dunn post-hoc tests (α=.05).

Results: In full-text screening, Claude achieved the highest sensitivity at 97%, while Claude and Elicit both showed strong overall performance with 86% accuracy and 87% F1-scores. All models maintained sensitivity above 90%. For data extraction, GPT-4 consistently performed best with median scores of 5.0, while Claude and Gemini showed similar capabilities. Significant differences only appeared in labeling and modeling tasks during Week 1 (P=.04). Risk of bias assessment agreement with experts varied from 55% to 90% across different criteria.

Conclusions: LLMs show potential for SR efficiency (especially for data extraction) but require human oversight because of variable performance across models and tasks.