Deep learning model could diagnose temporomandibular joint disorders – Recent dental research findings (1)

Deep learning model may diagnose TMD

A multinational study conducted in Germany, the United States, and Greece showed that deep learning models can be used to detect temporomandibular joint (TMJ) disorders with high sensitivity and specificity. The model is based on a rigorous evidence-based diagnostic framework, objective measurements, and advanced analytical techniques to improve diagnostic accuracy. This method can be used by oral healthcare professionals who are not specialized in TMJ disorders. The study was published online on May 17, 2024, in the Journal of Oral Rehabilitation.

To review the application of deep learning models in the diagnosis of TMJ disorders, researchers conducted a systematic review and searched relevant literature in multiple databases up to June 2023. Deep learning models based on TMJ or TMD were used to predict the effectiveness and classification of TMD compared to the reference standards. The Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool was used for critical analysis of the included studies to assess the risk of bias and calculate the diagnostic odds ratio (DOR).

Results:

A total of 21 studies met the inclusion criteria and were included in the systematic review. Four studies were assessed as having a low risk of bias according to QUADAS-2. The accuracy of all included studies ranged from 74% to 100%, sensitivity ranged from 54% to 100%, specificity ranged from 85% to 100%, Dice coefficient ranged from 85% to 98%, and the area under the curve (AUC) ranged from 77% to 99%. A meta-analysis of seven studies that met the criteria for pooling sensitivity, specificity, and dataset size showed a combined sensitivity of 95% (85% to 99%), specificity of 92% (86% to 96%), and AUC of 97% (96% to 98%). The DOR was 232 (74 to 729), indicating no significant bias in the studies.

Title: Deep learning for temporomandibular joint arthropathies: A systematic review and meta-analysis

Good efficacy of large periapical lesions – Microscopic apical surgery combined with PRP

A study conducted in India showed that large periapical lesions improved after treatment even after 5 years, with no deterioration in relevant parameters. The use of platelet-rich plasma (PRP) additionally enhanced the healing effect. Cone-beam computed tomography (CBCT) indices were used to better assess the healing of the apical resection, apical area, and cortical bone. The study was published online on May 17, 2024, in the International Endodontic Journal.

To evaluate the clinical and radiographic efficacy of microscopic apical surgery combined with PRP in penetrating periapical lesions at 1 year and 5 years, researchers selected 32 patients with large periapical lesions and randomly divided them into the PRP group and the control group. The healing of the lesions was assessed in both two-dimensional and three-dimensional dimensions. CBCT scans were used to evaluate the healing of the resection plane (R), apical area (A), buccal cortical bone (BC), palatal cortical bone (PC), and overall alveolar bone (B) and measure the lesion volume. The analysis included comparisons within and between groups at 1 to 5 years, and regression analysis was performed to determine the impact of various factors on the 5-year healing of the apical region.

Results:

At 1 year, a total of 32 patients (59 teeth) participated in the follow-up; at 5 years, 24 patients (44 teeth) participated. The overall success rates at 1 year were 66.7%, and at 5 years, they were 83.3%, with no cases deteriorating. The three-dimensional healing effect in the PRP group was significantly superior to that in the control group, with success rates at 1 year of 84.6% and 45.5%, respectively, and at 5 years of 100% and 63.6%, respectively. In terms of the R, BC, and B parameters, the PRP group had a significantly higher number of teeth showing complete healing at 5 years compared to the control group. The overall reduction in lesion volume was 88% at 1 year (91.4% in the PRP group and 84% in the control group) and 94% at 5 years (97.1% in the PRP group and 91.1% in the control group). Age, gender, lesion size, preoperative swelling and maxillary sinus condition, use of PRP, tooth position, preoperative buccal bone volume did not significantly affect the three-dimensional healing at 5 years.

Title: Two- and three-dimensional healing assessment after endodontic microsurgery in through-and-through periapical lesions: 5-year follow-up from a randomized controlled trial

Impacted lower third molar extraction – Significant correlation with these three factors

A study conducted by Professor Chi Yang and their team at the Ninth People’s Hospital affiliated with Shanghai Jiao Tong University School of Medicine showed that the position of the roots of impacted lower third molars (LM3) close to the inferior alveolar nerve (IAN) canal and the use of bone chisels during extraction may increase the risk of temporary IAN injury, but the use of corticosteroids treatment may promote nerve recovery. Compressive contact between LM3 and the IAN canal was the only risk factor for permanent IAN injury. The study was published online on May 15, in the Journal of Oral and Maxillofacial Surgery.

The study recruited patients who underwent bilateral LM3 extractions in the same clinic from May 2021 to December 2021, excluding patients with systemic diseases, previous maxillofacial surgery, or sensory abnormalities. Predictive variables were categorized into four groups: demographic, radiographic, surgery-related factors, and surgeon’s clinical experience. The outcome variable was the presence or absence of sensory disturbance after surgery, measured at one month (temporary) and one year (permanent).

Results:

A total of 705 patients (37.0% male) with a mean age of 28.51±6.51 years were recruited. Temporary and permanent IAN injuries occurred in 17 cases (2.4%) and 4 cases (0.57%), respectively. The following factors were significantly associated with a higher incidence of temporary nerve injury: the use of bone chisels during surgery (incidence rate of 6.4%, 95% confidence interval [CI] 2.7-12.3, P=0.02, risk ratio [RR] 11.4); LM3 located below the IAN canal (incidence rate of 8.7%, 95% CI 4.3-15.7, P<0.01, RR 7.3); compressive contact between LM3 and the IAN canal (incidence rate of 36.4%, 95% CI 12.3-78.2, P<0.001, RR 25.4); and no use of corticosteroids after surgery (incidence rate of 3.8%, 95% CI 1.9-6.5, P=0.03, RR 3.1). The only factor significantly associated with permanent nerve injury was compressive contact between LM3 and the IAN canal (incidence rate of 18.2%, 95% CI 2.2-62.3, P<0.001, RR 48.2).

Title: Risk factors associated with inferior alveolar nerve injury after extraction of impacted lower mandibular third molars: A prospective cohort study