Deog-Gyu Seo

Background: Entering clinical training, dental students must learn to read tooth-centered electronic dental records, but limited teaching time in patient-centered clinics can leave gaps in chart-reading literacy. Multimodal large language models (LLMs) that process dental record images may offer scalable educational support, yet their performance has not been benchmarked. Objective: This study evaluated multimodal ChatGPT models on dental chart image interpretation and compared the best-performing model with dental students and residents. Methods: A retrospective, cross-sectional benchmark study used deidentified dental chart images from 15 patients at Seoul National University Dental Hospital (2017‐2025). Charts containing Korean and English text were captured as sequential screenshots (154 images). For each patient, 24 Korean-language questions (360 total) spanned four categories: Type A, general factual retrieval; Type B, tooth- or procedure-specific retrieval; Type C, interpretation requiring multientry synthesis; and Type D, absent-information questions assessing abstention. Nine multimodal ChatGPT models (available August 2, 2025‐August 9, 2025) were evaluated under standardized conditions. Outputs were scored against a gold standard using 7 metrics, with sentence bidirectional encoder representations from transformers (SBERT) similarity prespecified as the primary semantic measure. Human baselines included 2 third-year students and 2 first-year residents. Groups were compared with Kruskal-Wallis tests and Dunn post hoc analyses. Gold-standard reliability was assessed by independent senior-expert review and chance-corrected agreement (Gwet AC1) among clinical reference raters, and the model ranking was confirmed by content-based clinical accuracy analysis. Results: Across 360 items, GPT-5 Thinking achieved the highest median SBERT similarity (0.900, IQR 0.525‐1.000), followed by GPT-5 Pro (0.861, IQR 0.501‐1.000) and OpenAI o3 (0.831, IQR 0.489‐1.000), with significant overall group differences ( P <.001). Student 1 did not differ significantly from GPT-5 Thinking across Types A-D (all P ≥.16), whereas Student 2 differed only on Type D ( P =.002; Cliff δ=−0.27). Resident 1 scored higher than GPT-5 Thinking on Types A ( P =.01) and C ( P =.04) but lower on Type D ( P <.001; δ=−0.35), whereas Resident 2 scored higher on Types A ( P =.03), B ( P =.02), and C ( P =.04) and did not differ on Type D ( P =.23). Type C tasks showed compressed SBERT distributions and low exact-match rates, indicating persistent difficulty in synthesis. The clinical reference rater agreement was high (Gwet AC1=0.99), and the content-based clinical-accuracy ranking was closely aligned with the SBERT ranking (Spearman ρ=0.90; P =.001). Conclusions: Statistically nonsignificant differences were observed between GPT-5 Thinking and dental students for most question-type contrasts, with Student 2 differing only on absent-information items. Compared with first-year residents, GPT-5 Thinking remained lower on several Type A-C contrasts, particularly interpretive Type C and one tooth- or procedure-specific Type B comparison; Type D contrasts require cautious interpretation because all groups had ceiling medians. Within this single-center benchmark, multimodal LLMs may have potential as supervised educational tools for chart-reading practice and verification, rather than as replacements for clinical expertise, pending external validation across institutions, specialties, and record systems.

In this review, we introduce the latest discoveries in the interaction of lasers with various functional polymers, discussing energy delivery mechanisms and their influence on structure modifications and property changes.

Influence of adjacent restorative material and distance on the accuracy of inlay cavity impressions with intraoral scanner: an in vitro study

Extended reality (XR) is an emerging field that connects the physical and digital worlds, enabling communication that transcends time and space. Commercial XR devices have been developed to support such experiences, but they are limited to specific sensations, mainly vibrational cues. Furthermore, these devices are realized mainly in rigid form factors, requiring external controllers or equipment, which hinders intuitive interaction and causes a mismatch with natural body movements. In this regard, skin-integrated human-machine interfaces with wearable electronics have played an important role in intuitive and immersive interaction in the XR environment, facilitating highly authentic sensory reconstruction and perception. Novel innovations in materials and structural design have enabled a wider range of sensory modalities and miniaturization, overcoming the limitations of conventional rigid XR systems. In this article, we thoroughly review human perception mechanisms to replicate hyper-realistic sensations. Then, we deal with the design and functionality for sensory feedback and input, specifically tailored for XR applications. In addition, we discuss precise system-level integration for untethered XR devices, alongside the role of artificial intelligence in real-time processing and rapid sensation conversion through predictive algorithms. Finally, we introduce promising XR applications and conclude with the challenges and prospects of future XR technologies.

The purpose of this study was to evaluate the bactericidal effect of 270 nm UV-C light-emitting diode (LED) light delivered through a newly designed prototype device with thin optical fiber against Enterococcus faecalis (E. faecalis). The prototype device, developed to integrate UV-C light into a thin optic fiber (diameter 124 µm) connected to a UV-C LED (Luminous Device; Sunnyvale, CA, USA) via a specialized double-lens system that focuses divergent light to achieve a 65 mm working distance and a numerical aperture of 0.22. E. faecalis, was cultured at 37 °C under aerobic conditions for 24 h. The UV-C LED optical fiber was positioned 10 mm above the bacterial culture prepared in the wells of a 96-well plate. The E. faecalis cells were exposed to UV-C irradiation for 0, 10, 30, 60, 90, 120 and 180 s. Following irradiation, the OD600 values were measured after incubation at 37 °C for an additional 24 h. The data were statistically analyzed using one-way ANOVA, followed by Tukey’s honestly significant difference (HSD) test at a significance level of 0.05. UV irradiation at 270 nm significantly reduced E. faecalis growth in a time-dependent manner (p < 0.05). No significant changes were observed at 0 and 10 s, while peak reductions occurred at 120 and 180 s, with effects beginning at 30 s and increasing over time. The 270 nm UV-C wavelength was highly effective in bactericidal action against E. faecalis. The custom-designed UV-C delivery system effectively integrated the light source into a thin optical fiber, allowing for efficient UV-C light transmission and demonstrating its potential for application in narrow spaces such as root canals.

This study assessed the changes in the mechanical and surface properties of the transition zone in multilayered translucent monolithic zirconia subjected to long-term hydrothermal aging. A total of 360 disk-shaped specimens (diameter: 15.0 mm; thickness: 1.2 mm) were prepared using conventional (3Y-TZP in LT; ZL, 4Y-TZP in MT; ZM) and multilayered translucent zirconia (5Y-TZP in MT Multi; ZT, 3Y/5Y-TZP in Prime; ZP) among IPS e.max ZirCAD blocks. Specimens were divided into three groups (n = 30) and aged in the autoclave at 134 °C under 0.2 MPa for 0 h (control group), 5 h (first aged group), and 10 h (second aged group). The mechanical and surface properties of the transition zone in the multilayered translucent zirconia were investigated, followed by statistical analysis (α = 0.05). Before and after aging, ZL (1102.64 ± 41.37 MPa) and ZP (1014.71 ± 139.86 MPa) showed the highest biaxial flexural strength (BFS); ZL showed the highest Weibull modulus (31.46) and characteristic strength (1121.63 MPa); and ZT exhibited the highest nanoindentation hardness (20.40 ± 1.80 GPa) and Young’s modulus (284.90 ± 20.07 GPa). After aging, ZL (116.75 ± 9.80 nm) exhibited the highest surface roughness (Ra); the monoclinic phase contents in ZL and ZP increased; and surface uplifts, microcracks, and irregular defects caused by phase transformation appeared on ZL and ZP surfaces. The 3Y/5Y-TZP transition zone exhibited flexural strength, Vickers hardness, phase distribution changes, and surface microstructure changes similar to those of 3Y-TZP before and after aging; however, the surface roughness was lower than that of 3Y-TZP and higher than those of 4Y-TZP and 5Y-TZP after aging. The mechanical and surface characteristics, excluding BFS and Vickers hardness, were influenced by the yttrium oxide content in each zone and the aging process.