Detection of facial landmarks by a convolutional neural network in patients with oral and maxillofacial disease

Empowering surgeons: will artificial intelligence change oral and maxillofacial surgery?

Artificial Intelligence (AI) can enhance the precision and efficiency of diagnostics and treatments in oral and maxillofacial surgery (OMS), leveraging advanced computational technologies to mimic intelligent human behaviors. The study aimed to examine the current state of AI in the OMS literature and highlight the urgent need for further research to optimize AI integration in clinical practice and enhance patient outcomes. A scoping review of journals related to OMS focused on OMS-related applications. PubMed was searched using terms "artificial intelligence", "convolutional networks", "neural networks", "machine learning", "deep learning", and "automation". Ninety articles were analyzed and classified into the following subcategories: pathology, orthognathic surgery, facial trauma, temporomandibular joint disorders, dentoalveolar surgery, dental implants, craniofacial deformities, reconstructive surgery, aesthetic surgery, and complications. There was a significant increase in AI-related studies published after 2019, 95.6% of the total reviewed. This surge in research reflects growing interest in AI and its potential in OMS. Among the studies, the primary uses of AI in OMS were in pathology (e.g., lesion detection, lymph node metastasis detection) and orthognathic surgery (e.g., surgical planning through facial bone segmentation). The studies predominantly employed convolutional neural networks (CNNs) and artificial neural networks (ANNs) for classification tasks, potentially improving clinical outcomes.

Deep learning-assisted diagnosis of large vessel occlusion in acute ischemic stroke based on four-dimensional computed tomography angiography

Purpose

To develop deep learning models based on four-dimensional computed tomography angiography (4D-CTA) images for automatic detection of large vessel occlusion (LVO) in the anterior circulation that cause acute ischemic stroke.

Methods

This retrospective study included 104 LVO patients and 105 non-LVO patients for deep learning models development. Another 30 LVO patients and 31 non-LVO patients formed the time-independent validation set. Four phases of 4D-CTA (arterial phase P1, arterial-venous phase P2, venous phase P3 and late venous phase P4) were arranged and combined and two input methods was used: combined input and superimposed input. Totally 26 models were constructed using a modified HRNet network. Assessment metrics included the areas under the curve (AUC), accuracy, sensitivity, specificity and F1 score. Kappa analysis was performed to assess inter-rater agreement between the best model and radiologists of different seniority.

Results

The P1 + P2 model (combined input) had the best diagnostic performance. In the internal validation set, the AUC was 0.975 (95%CI: 0.878-0.999), accuracy was 0.911, sensitivity was 0.889, specificity was 0.944, and the F1 score was 0.909. In the time-independent validation set, the model demonstrated consistently high performance with an AUC of 0.942 (95%CI: 0.851-0.986), accuracy of 0.902, sensitivity of 0.867, specificity of 0.935, and an F1 score of 0.901. The best model showed strong consistency with the diagnostic efficacy of three radiologists of different seniority (k = 0.84, 0.80, 0.70, respectively).

Conclusion

The deep learning model, using combined arterial and arterial-venous phase, was highly effective in detecting LVO, alerting radiologists to speed up the diagnosis.

Electroacupuncture with intermittent wave stimulation as rehabilitation approach for chronic Bell's palsy: a randomized controlled trial

PURPOSE

To evaluate the effectiveness and safety of electroacupuncture (EA) using intermittent wave stimulation in enhancing facial symmetry and nerve function in chronic Bell's palsy patients.

METHODS

A 6-week assessor-blinded, randomized trial followed by an 18-week observational period was conducted. Sixty individuals with chronic Bell's palsy, showing no signs of recovery after 12 months, were equally divided to receive either 18 sessions of EA using intermittent wave stimulation or Transcutaneous Electrical Stimulation (TES), administered thrice weekly over 6 weeks. The primary outcome measure was the change in the total facial nerve index (TFNI) score from baseline to Week 6, with secondary outcomes including TFNI scores at Weeks 12 and 24, as well as the change in Sunnybrook Facial Grading System (SFG) score from baseline to Week 6, and SFG scores at Weeks 12 and 24.

RESULTS

The EA group showed a significant improvement, with a mean total facial nerve index score increase of 24.35 (4.77) by Week 6 compared with 14.21 (5.12) in the Transcutaneous Electrical Stimulation group (P<.001). This superiority persisted during the 24-week follow-up. While no significant difference was observed in the Sunnybrook Facial Grading System score change from baseline to Week 6, variations were noted at Weeks 12 and 24. No major adverse effects were reported.

CONCLUSION

EA with intermittent wave stimulation notably enhanced facial symmetry in chronic Bell's palsy patients over Transcutaneous Electrical Stimulation by Week 6, maintaining this edge throughout the follow-up.

Enhancing head and neck tumor management with artificial intelligence: Integration and perspectives

Head and neck tumors (HNTs) constitute a multifaceted ensemble of pathologies that primarily involve regions such as the oral cavity, pharynx, and nasal cavity. The intricate anatomical structure of these regions poses considerable challenges to efficacious treatment strategies. Despite the availability of myriad treatment modalities, the overall therapeutic efficacy for HNTs continues to remain subdued. In recent years, the deployment of artificial intelligence (AI) in healthcare practices has garnered noteworthy attention. AI modalities, inclusive of machine learning (ML), neural networks (NNs), and deep learning (DL), when amalgamated into the holistic management of HNTs, promise to augment the precision, safety, and efficacy of treatment regimens. The integration of AI within HNT management is intricately intertwined with domains such as medical imaging, bioinformatics, and medical robotics. This article intends to scrutinize the cutting-edge advancements and prospective applications of AI in the realm of HNTs, elucidating AI's indispensable role in prevention, diagnosis, treatment, prognostication, research, and inter-sectoral integration. The overarching objective is to stimulate scholarly discourse and invigorate insights among medical practitioners and researchers to propel further exploration, thereby facilitating superior therapeutic alternatives for patients.

Assessing facial weakness in myasthenia gravis with facial recognition software and deep learning

OBJECTIVE

Myasthenia gravis (MG) is an autoimmune disease leading to fatigable muscle weakness. Extra-ocular and bulbar muscles are most commonly affected. We aimed to investigate whether facial weakness can be quantified automatically and used for diagnosis and disease monitoring.

METHODS

In this cross-sectional study, we analyzed video recordings of 70 MG patients and 69 healthy controls (HC) with two different methods. Facial weakness was first quantified with facial expression recognition software. Subsequently, a deep learning (DL) computer model was trained for the classification of diagnosis and disease severity using multiple cross-validations on videos of 50 patients and 50 controls. Results were validated using unseen videos of 20 MG patients and 19 HC.

RESULTS

Expression of anger (p = 0.026), fear (p = 0.003), and happiness (p < 0.001) was significantly decreased in MG compared to HC. Specific patterns of decreased facial movement were detectable in each emotion. Results of the DL model for diagnosis were as follows: area under the curve (AUC) of the receiver operator curve 0.75 (95% CI 0.65-0.85), sensitivity 0.76, specificity 0.76, and accuracy 76%. For disease severity: AUC 0.75 (95% CI 0.60-0.90), sensitivity 0.93, specificity 0.63, and accuracy 80%. Results of validation, diagnosis: AUC 0.82 (95% CI: 0.67-0.97), sensitivity 1.0, specificity 0.74, and accuracy 87%. For disease severity: AUC 0.88 (95% CI: 0.67-1.0), sensitivity 1.0, specificity 0.86, and accuracy 94%.

INTERPRETATION

Patterns of facial weakness can be detected with facial recognition software. Second, this study delivers a 'proof of concept' for a DL model that can distinguish MG from HC and classifies disease severity.

Artificial Intelligence in Oral and Maxillofacial Surgery Education

Artificial intelligence has become ubiquitous with modern technology. Digital transformations are occurring in every field including medicine, surgery, and education. Computers and computer programs are getting sophisticated to form neural networks globally. These algorithms allow for sophisticated and complex pattern recognitions and make accurate predictions. This allows for both accurate diagnosis and prognostication in medicine and opens opportunities for medical and surgical education. Oral and Maxillofacial surgeons and OMS education like all of the surgery are adapting well to the world of AI, incorporating machine learning into simulation, and attaching sensors to master surgeons to understand motion economy.

Development of a Method for Quantitative Evaluation of Facial Swelling in a Rat Model of Cerebral Ischemia by Facial Image Processing

A quantitative method for the evaluation of facial swelling in rats with middle cerebral artery occlusion (MCAO) was established using a mathematical method for the first time. The rat model of MCAO was established via bilateral common carotid artery ligation. Three groups of rats with the same baseline were selected (model group, positive drug group, and control group) according to their behavioral score and body weight 24 h after surgery. Drug administration was initiated on post-MCAO day 8 and was continued for 28 days. Mobile phones were used to collect facial images at different time points after surgery. In facial image analysis, the outer canthi of both eyes were used as the facial dividing line, and the outer edge of the rat's face was framed using the marking method, and the framed part was regarded as the facial area (S) of the rats. The histogram created with Photoshop CS5 was used to measure the face area in pixels. The distance between the outer canthi of both eyes (Le) and vertical line from the tip of the nose to the line joining the eyes was recorded as H1, and the line from the tip of the nose to the midpoint of the line joining the eyes was recorded as H2. The facial area was calibrated based on the relationship between H1 and H2. The distance between the eyes was inversely proportional to the distance between the rats and mobile phone such that the face area was calibrated by unifying Le. The size of Le between the eyes was inversely proportional to the distance between the rats and mobile phone. This was used to calibrate the face area. When compared with the control group, the facial area of the model group gradually increased from postoperative day 1 to day 7, and there was a significant difference in the facial area of the model group on postoperative day 7. Hence, positive drugs exhibited the effect of improving facial swelling. H1 and H2 can reflect the state of turning the head and raising the head of the rats, respectively. Facial area was calibrated according to the relationship between H1 and H2, which had no obvious effect on the overall conclusion. Furthermore, mobile phone lens was used to capture the picture of rat face, and the distance between the eyes and H1 and H2 was used to calibrate the facial area. Hence, this method is convenient and can be used to evaluate subjective judgment of the human eyes via a quantitative method.