Journal of Oral and Maxillofacial Surgery最新文献

Background: Large language models (LLMs) such as Chat Generative Pre-Trained Transformer (ChatGPT; OpenAI, San Francisco, CA) and Claude (Anthropic, San Francisco, CA) are increasingly used by patients seeking information about surgical procedures, including external sinus lifting. However, the accuracy, quality, and readability of these artificial intelligence (AI)-generated explanations remain uncertain.

Purpose: The study purpose was to measure and compare 2 AI language models regarding the reliability, quality, usefulness, and readability of their responses to frequently asked patient questions about external sinus lifting.

Study design, setting, and sample: This cross-sectional study assessed computer-generated responses provided by LLMs, referred to as decoder-only-based LLM (DO-LLM) and transformer-based LLM (TB-LLM) to standardized patient questions.

Predictor variable: The predictor variable was AI model type (DO-LLM vs TB-LLM).

Main outcome variables: Outcome measures included reliability, quality, usefulness, and readability. These were assessed using the modified DISCERN instrument, Global Quality Score, a 4-point usefulness scale, and 2 readability indices (Flesch Reading Ease and Flesch-Kincaid Grade Level). Seventy-two standardized questions across 10 clinical domains were submitted to both models. Responses were independently evaluated by one oral and maxillofacial surgeon and 2 periodontists, followed by consensus scoring.

Covariates: Not applicable.

Analyses: Descriptive statistics summarized outcomes. Depending on normality, comparisons used the independent samples t-test or Mann-Whitney U test. Associations between categorical variables were analyzed using Pearson's χ² or Fisher's Exact test.

Results: For modified DISCERN, DO-LLM scored 21.88 (3.09), 22.14 (2.04), and 22.63 (2.56) in preoperative preparation, graft materials, and risks/complications, whereas TB-LLM scored 13.88 (4.52), 17.29 (4.27), and 19 (1.77), respectively (P < .05). For Global Quality Score in lifestyle and behavioral recommendations, TB-LLM scored 4 (0) compared with 3.29 (0.49) for DO-LLM (P < .05). Moderate-quality responses were more common with DO-LLM (56.9%), while TB-LLM produced a higher proportion of good quality scores (29.2%) (P < .05).

Conclusion and relevance: Both AI models demonstrated potential value for patient education on external sinus lifting, though their strengths differed by content domain. DO-LLM provided stronger procedural and risk-related explanations, whereas TB-LLM offered more comprehensive lifestyle-related guidance. Continued refinement of dental-specific AI tools and integration of patient-centered considerations remain essential.

In the traditional jaw in a day technique, discrepancies between the virtual surgical plan and final implant positions are accommodated by creating oversized access openings in the prefabricated prosthesis, allowing passive adaptation around implant cylinders before registering definitive implant positions with autopolymerizing or light-curing resin. Utilizing digital scanning and 3-dimensional printing, a direct-to-multiunit abutment dental prosthesis can be fabricated as a fully digital alternative to this analog method. Although direct-to-multiunit abutment connections have been reported in head and neck reconstruction using the IPS Preprosthetic System, their application in jaw in a day has not been previously described. This digital workflow eliminates restorative materials unsuitable for the operative environment, maintains sterility, simplifies intraoperative handling, and enhances mechanical integrity through minimized access openings. Moreover, it transfers prosthetic fabrication to the dental laboratory, allowing concurrent production during surgical reconstruction and enabling immediate intraoperative delivery.

Background: Surface electromyography provides an objective, noninvasive method for assessing masseter muscle function and postoperative recovery following open reduction and internal fixation (ORIF) of mandibular condylar fractures.

Purpose: The purpose of the study was to compare masseter muscle power and contractility following ORIF of unilateral mandibular condylar fractures using the transparotid (TP) and transmasseteric anterior parotid (TMAP) approaches.

Study design, setting, sample: This retrospective cohort study included adult subjects who underwent ORIF for unilateral condylar fractures using TP or TMAP approaches between January 2021 and December 2022. Subjects with pretraumatic temporomandibular joint disorders, additional surgical approaches, or incomplete surface electromyography data were excluded. Twenty-two subjects were analyzed (11 per group).

Predictor variable: The predictor variable was surgical approach (TP vs TMAP).

Main outcome variables: The primary outcome variables were maximum amplitude and mean power. The amplitude is the maximum peak of the negative phase to the maximum peak of positive phase in electromyography. Power is defined as the sum of action potentials produced in the muscle.

Covariates: The covariates were age, sex, and fracture patterns.

Analyses: Longitudinal changes were analyzed using repeated measures analysis of variance. Operated and contralateral sides were compared using paired t-tests. Kaplan-Meier survival analysis assessed time to ≥80% recovery (P < .05).

Results: The mean age was 32.8 ± 11.2 years in the TP group and 27.8 ± 8.5 years in the TMAP group (P = .3), with male predominance (86.3%, n = 19). Repeated-measures analysis of variance demonstrated a significant effect of time on both amplitude and power within group (P < .001), with no significant time × approach interaction. Kaplan-Meier analysis showed that all subjects achieved ≥80% recovery by the earliest postoperative interval (7-10 days), with no difference between approaches (P = .4). At 3 months, a small but significant reduction in mean power persisted on the operated side in the TMAP group (P = .024).

Conclusion and relevance: Regardless of the approach used, the masseter muscle achieved ≥80% functional recovery within 7-10 days. Recovery is primarily time-dependent, allowing the choice of surgical approach to be guided by fracture characteristics and surgeon expertise.

Background: Osteoradionecrosis (ORN) is a severe adverse effect of radiation therapy for head and neck cancers. Despite advances in radiation techniques and surgical management, reliable predictors of ORN prognosis remain lacking.

Purpose: The purpose of this study was to measure the association between systemic inflammatory biomarkers and ORN prognosis.

Study design, setting, sample: This retrospective cohort study included patients with ORN treated between January 2010 and March 2022 at the Department of Oral and Maxillofacial Surgery, Kobe University Graduate School of Medicine. Patients with unavailable blood test data were excluded.

Predictor variable: The predictor variables were pretreatment systemic inflammatory biomarkers, including neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio, and lymphocyte-to-monocyte ratio (LMR). The cutoff values for NLR, platelet-to-lymphocyte ratio, and LMR were determined using receiver operating characteristic curve analysis. Each marker was calculated at the time of diagnosis of ORN.

Main outcome variable: The outcome variable was prognosis. Subjects had a good prognosis if the ORN stage improved. The prognosis was poor if the ORN stage was stable or worse over time.

Covariates: Sex, age, stage, onset region, primary tumor site, radiation dose, smoking history, medical history, steroid therapy, clinical symptoms, treatment methods, and computed tomography image findings were covariates.

Analyses: Statistical analysis was performed using Mann-Whitney U nonparametric test, Fisher's exact test, χ² test, and Cox regression analysis. Statistical significance was set at P < .05.

Results: The sample consisted of 75 subjects, of whom 16 (21.3%) had a poor prognosis. The median follow-up period was 88 (53 to 122) months. Bivariate analysis revealed that high NLR (P < .01) and low LMR (P = .02) were significantly associated with poor prognosis. In multivariate Cox regression analysis, poor prognosis was significantly associated with advanced age (hazard ratio [HR], 5.33; 95% CI, 1.61 to 17.65; P < .01), ORN stage 3, 4 (HR, 0.18; 95% CI, 0.06 to 0.56; P < .01), high NLR (HR, 14.78; 95% CI, 3.50 to 62.35; P < .01), and periosteal reaction (HR, 19.15; 95% CI, 4.65 to 78.87; P < .01).

Conclusions and relevance: We identified high NLR as a risk factor for poor prognosis in ORN. High NLR is considered to reflect worse local tissue and systemic health.

This technological innovation introduces a mixed reality-based system using HoloLens 2 for surgical removal of impacted supernumerary teeth. The technology superimposes 3-dimensional virtual models reconstructed from cone-beam computed tomography and intraoral scan data onto the actual surgical field, providing real-time visual guidance for precise localization. This innovative approach enables minimally invasive procedures while eliminating the expenses and limitations associated with traditional surgical guides. Although implementation faces challenges, including hardware costs, operational learning curve, and environmental lighting requirements, the method demonstrates significant potential for reducing surgical trauma and improving outcomes. With anticipated advancements in hardware accessibility and protocol standardization, clinical adoption is projected within 12 to 24 months. The technique also shows promising applications in localizing deep-seated tumors and foreign bodies.

Background: Severe ocular injuries (SOI) can occur in orbital wall fractures and may be overlooked in acute care. Recognition of factors associated with SOI may improve patient outcomes.

Purpose: This study aimed to create and validate a predictive model for SOI in orbital wall fractures.

Study design, setting, sample: This is a retrospective cohort study. To create the training predictive model, the investigators screened a study population composed of all patients who presented to Farabi Eye Hospital between 2016 and 2023 for the evaluation and management of orbital wall fractures. The exclusion criteria were the absence of an ophthalmologic examination or an orbital computed tomography report. The validation model was composed of an external dataset that included 97 subjects.

Predictor variables: The predictor variables were trauma details, uncorrected visual acuity (UCVA), lid laceration, conjunctival hemorrhage, computed tomography scan features, including the number of herniated extraocular muscles (EOMs).

Main outcome variable(s): The main outcome variable was the presence of SOI, including open globe injuries, optic nerve-related injuries, and severe closed globe injuries.

Covariates: The covariates were age and sex.

Analyses: Multivariable logistic regression was performed to identify factors associated with SOI. Model performance was assessed using the area under the receiver operating characteristic curve, specificity, and negative predictive value (NPV) in an external validation dataset. Statistical significance was set at P < .05.

Results: The sample consisted of 537 subjects, with a mean age of 29.8 ± 13.91 years, and 22.7% (124) were female. The multivariable regression analysis revealed that the 4 main variables associated with SOI were: trauma setting (odds ratio [OR] = 2.61, 95% CI 1.1 to 6.21), UCVA (OR = 4.45, 95% CI 2.65 to 6.27), conjunctival hemorrhage (OR = 4.07, 95% CI 1.21 to 3.49), and herniated EOMs (OR = 1.77, 95% CI 1.2 to 2.61). In the training set, the specificity and NPV were 76.8 and 85.5%, respectively. This model achieved an area under the receiver operating characteristic curve of 0.77 for the subset of closed-globe injuries. When applied to the validation set, the model achieved a specificity of 76.6% and an NPV of 86%.

Conclusions and relevance: We identified that trauma setting, reduced UCVA, conjunctival hemorrhage, and herniated EOMs are associated with SOI in orbital fractures.