Artificial intelligence (AI)-assisted denture design is increasingly being incorporated into digital complete denture workflows, but its clinical performance in edentulous patients remains insufficiently documented. This preliminary within-patient case series evaluated the immediate clinical feasibility and morphological characteristics of AI-assisted complete dentures (ADs) compared with conventionally arranged complete dentures (CDs). Six completely edentulous participants each received one pair of ADs and CDs fabricated from the same definitive impressions and jaw-relation records obtained using the JB Tray and POP-Bow. After delivery and routine adjustment, left–right occlusal force distribution and bilateral balanced occlusal contacts were assessed using T-Scan III. Facial scans were used to evaluate facial maxillary midline relationships and nasolabial angle, and digitized denture data were analyzed to measure maxillary arch dimensions. All dentures were successfully delivered. Occlusal force imbalance did not differ significantly between CDs and ADs (3.90 ± 3.20% and 5.25 ± 2.75%, respectively; p = 0.219), and bilateral balanced occlusal contacts were observed in both denture types for all participants. ADs showed smaller facial-maxillary midline discrepancies, smaller nasolabial angles, and narrower intercanine and intermolar widths (p = 0.031 for each). These preliminary findings suggest that AI-assisted denture design can be incorporated into a clinician-controlled complete denture workflow, although clinical superiority over conventional tooth arrangement was not established.

최종원, et al. J Korean Acad Prosthodont. 2026 Jul;64(3):229-238. https://doi.org/10.4047/jkap.2026.64.3.229

Background/Objectives: This study aimed to compare the final impressions obtained using three different impression techniques in completely edentulous patients, focusing on the clinical applicability of the JB Tray^® for immediate custom tray fabrication with a polycaprolactone-based thermoplastic resin. Methods: Five edentulous patients were recruited for a preliminary clinical evaluation, and three impression techniques were tested: (1) JB Tray^® (JB), which allows simultaneous border molding and final impression taking; (2) the conventional modeling compound border-molded technique (MC, control); and (3) the one-step silicone border-molded technique (OS). All impressions were made by a single prosthodontist, and the internal surfaces were digitized using a 3D scanner for analysis of internal fit. Each arch was divided into reference regions for quantitative comparison. Statistical significance was set at p < 0.05. Results: Color map analysis (threshold of ±1.0 mm) revealed that the JB Tray^® group exhibited wider border extension in both the maxilla and mandible compared with the modeling compound and silicone groups, with a more pronounced effect in the mandible. However, no statistically significant differences in overall internal fit were observed among the three methods (p > 0.05). Conclusions: These preliminary results due to small sample size suggest that the JB Tray^® enables efficient impression taking in edentulous patients and can serve as a viable alternative to conventional custom tray techniques. Further research with a larger sample size is warranted to support widespread clinical application.