Creation of an artificial intelligence model for intubation difficulty classification by deep learning (convolutional neural network) using face images: an observational study

Global trends in artificial intelligence research in anesthesia from 2000 to 2023: a bibliometric analysis

BACKGROUND

Interest in artificial intelligence (AI) research in anesthesia is growing rapidly. However, there is a lack of bibliometric analysis to measure and analyze global scientific publications in this field. The aim of this study was to identify the hotspots and trends in AI research in anesthesia through bibliometric analysis.

METHODS

English articles and reviews published from 2000 to 2023 were retrieved from the Web of Science Core Collection (WoSCC) database. The extracted data were summarized and analyzed using Microsoft Excel, and bibliometric analysis were conducted with VOSviewer software.

RESULTS

AI research literature in anesthesia has exhibited rapid growth in recent years. The United States leads in the number of publications and citations, with Stanford University as the most prolific institution. Hyung-Chul Lee is the author with the highest number of publications. The journal Anesthesiology is highly recognized and authoritative in this field. Recent keywords include "musculoskeletal pain", "precision medicine", "stratification", "images", "mean arterial pressure", " enhanced recovery after surgery", "frailty", "telehealth", "postoperative delirium" and "postoperative mortality" indicating hot topics in AI research in anesthesia.

CONCLUSIONS

Publications on AI research in the field of anesthesia have experienced rapid growth over the past two decades and are likely to continue increasing. Research areas such as depth of anesthesia (DOA) and drug infusion (including electroencephalography and deep learning), perioperative risk assessment and prediction (covering mean arterial pressure, frailty, postoperative delirium, and mortality), image classification and recognition (for applications such as ultrasound-guided nerve blocks, vascular access, and difficult airway assessment), and perioperative pain management (particularly musculoskeletal pain) have garnered significant attention. Additionally, topics such as precision medicine, enhanced recovery after surgery, and telehealth are emerging as new hotspots and future directions in this field.

Artificial intelligence in regional anesthesia

PURPOSE OF REVIEW

Artificial intelligence (AI) is having an increasing impact on healthcare. In ultrasound-guided regional anesthesia (UGRA), commercially available devices exist that augment traditional grayscale ultrasound imaging by highlighting key sono-anatomical structures in real-time. We review the latest evidence supporting this emerging technology and consider the opportunities and challenges to its widespread deployment.

RECENT FINDINGS

The existing literature is limited and heterogenous, which impedes full appraisal of systems, comparison between devices, and informed adoption. AI-based devices promise to improve clinical practice and training in UGRA, though their impact on patient outcomes and provision of UGRA techniques is unclear at this early stage. Calls for standardization across both UGRA and AI are increasing, with greater clinical leadership required.

SUMMARY

Emerging AI applications in UGRA warrant further study due to an opaque and fragmented evidence base. Robust and consistent evaluation and reporting of algorithm performance, in a representative clinical context, will expedite discovery and appropriate deployment of AI in UGRA. A clinician-focused approach to the development, evaluation, and implementation of this exciting branch of AI has huge potential to advance the human art of regional anesthesia.

Use of Artificial Intelligence in Difficult Airway Assessment: The Current State of Knowledge

Artificial Intelligence (AI) has become one of the most transformative technologies of the 21st century. It is poised to reshape medicine, as almost every field of hospital treatment has seen an increase in AI's presence. In this article, we focus on its impact in the field of anesthesia. We discuss its possible influence on difficult airway management, as it remains one of the most critical and potentially hazardous aspects of anesthesia, often leading to life-threatening complications. The accurate prediction of difficult airways can significantly improve patient safety. We covered the available literature on AI-based models for difficult airway prediction in comparison to traditional forms of airway assessment, as well as the predictive value of ultrasonography. We also address the narrative that AI-based algorithms show high sensitivity and specificity, with which they significantly outperform classical tests based on complex scales and indices.

The Future of Artificial Intelligence Using Images and Clinical Assessment for Difficult Airway Management

Artificial intelligence (AI) algorithms, particularly deep learning, are automatic and sophisticated methods that recognize complex patterns in imaging data providing high qualitative assessments. Several machine-learning and deep-learning models using imaging techniques have been recently developed and validated to predict difficult airways. Despite advances in AI modeling. In this review article, we describe the advantages of using AI models. We explore how these methods could impact clinical practice. Finally, we discuss predictive modeling for difficult laryngoscopy using machine-learning and the future approach with intelligent intubation devices.

Artificial intelligence in anesthesiology: a bibliometric analysis

The application of artificial intelligence (AI) in anesthesiology has become increasingly widespread. However, no previous study has analyzed this field from the bibliometric analysis dimension. The objective of this paper was to assess the global research trends in AI in anesthesiology using bibliometric software. Literatures relevant to AI and anesthesiology were retrieved from the Web of Science until 10 April 2024 and were visualized and analyzed using Excel, CiteSpace, and VOSviewer. After screening, 491 studies were included in the final bibliometric analysis. The growth rate of publications, countries, institutions, authors, journals, literature co-citations, and keyword co-occurrences was computed. The number of publications increased annually since 2018, with the most significant contributions from the USA, China, and England. The top 3 institutions were Yuan Ze University, National Taiwan University, and Brunel University London. The top three journals were Anesthesia & Analgesia, BMC Anesthesiology, and the British Journal of Anaesthesia. The researches on the application of AI in predicting hypotension have been extensive and represented a hotspot and frontier. In terms of keyword co-occurrence cluster analysis, keywords were categorized into four clusters: ultrasound-guided regional anesthesia, postoperative pain and airway management, prediction, depth of anesthesia (DoA), and intraoperative drug infusion. This analysis provides a systematic analysis on the literature regarding the AI-related research in the field of anesthesiology, which may help researchers and anesthesiologists better understand the research trend of anesthesia-related AI.

Unravelling intubation challenges: a machine learning approach incorporating multiple predictive parameters

BACKGROUND

To protect patients during anesthesia, difficult airway management is a serious issue that needs to be carefully planned for and carried out. Machine learning prediction tools have recently become increasingly common in medicine, frequently surpassing more established techniques. This study aims to utilize machine learning techniques on predictive parameters for challenging airway management.

METHODS

This study was cross-sectional. The Shahid Beheshti University of Medical Sciences in Iran's Loghman Hakim and Shahid Labbafinezhad hospitals provided 622 records in total for analysis. Using the forest of trees approach and feature importance, important features were chosen. The Synthetic Minority Oversampling Technique (SMOTE) and repeated edited nearest neighbor under-sampling were used to balance the data. Using Python and 10-fold cross-validation, seven machine learning algorithms were assessed: Logistic Regression, Support Vector Machines (SVM), Random Forest (INFORMATION-GAIN and GINI-INDEX), Decision Tree, and K-Nearest Neighbors (KNN). Metrics like F-measure, AUC, Recall, Accuracy, Specificity, and Precision were used to evaluate the performance of the model.

RESULTS

Twenty-four important features were chosen from the original 32 features. The under-sampling strategy produced better results than SMOTE. Among the algorithms, KNN (Euclidean, Minkowski) had better performance than other algorithms. The highest values ​​for accuracy, precision, recall, F-measure, and AUC were obtained at 0.87, 0.88, 0.82, 0.85, and 0.87, respectively.

CONCLUSION

Algorithms for machine learning provide insightful information for anticipating challenging airway management. By making it possible to forecast airway difficulties more accurately, these techniques can potentially improve clinical practice and patient outcomes.

G2LCPS: End-to-end semi-supervised landmark prediction with global-to-local cross pseudo supervision for airway difficulty assessment

Difficult tracheal intubation is a major cause of anesthesia-related injuries, including brain damage and death. While deep neural networks have improved difficult airways (DA) predictions over traditional assessment methods, existing models are often black boxes, making them difficult to trust in critical medical settings. Traditional DA assessment relies on facial and neck features, but detecting neck landmarks is particularly challenging. This paper introduces a novel semi-supervised method for landmark prediction, namely G2LCPS, which leverages hierarchical filters and cross-supervised signals. The novelty lies in ensuring that the networks select good unlabeled samples at the image level and generate high-quality pseudo heatmaps at the pixel level for cross-pseudo supervision. The extended versions of the public AFLW, CFP, CPLFW and CASIA-3D FaceV1 face datasets and show that G2LCPS achieves superior performance compared to other state-of-the-art semi-supervised methods, achieving the lowest normalized mean error (NME) of 3.588 when only 1/8 of data is labeled. Notably, the inclusion of the local filter improved the prediction by at least 0.199 NME, whereas the global filter contributed an additional improvement of at least 0.216 NME. These findings underscore the effectiveness of our approach, particularly in scenarios with limited labeled data, and suggest that G2LCPS can significantly enhance the reliability and accuracy of DA predictions in clinical practice. The results highlight the potential of our method to improve patient safety by providing more trustworthy and precise predictions for difficult airway management.

Voice Analysis as a Method for Preoperatively Predicting a Difficult Airway Based on Machine Learning Algorithms: Original Research Report

Background and Aims

An unanticipated difficult airway is one of the greatest challenges for anesthesiologists. Proper preoperative airway assessment is crucial to reducing complications. However, current screening tests based on anthropometric features are of uncertain benefit. Therefore, our study explores using voice analysis with machine learning algorithms to predict a difficult airway.

Methods

Observational, multicenter study with N = 438 patients initially enrolled at Centro Medico Teknon and Institut Universitari Dexeus (2019-2022) for the research study. After excluding 125 patients, N = 313 were included. Ethics committee approval was obtained. Adults ASA I-III scheduled for elective procedures under general anesthesia with endotracheal intubation were selected. Patient clinical features and traditional predictive tests were collected. Vowels "A, E, I, O, U" were recorded in normal, flexion, and extension positions. Cormack grade was assessed, and data were analyzed using KNIME, resulting in multiple models based on demographics and voice data. ROC curves and other metrics were evaluated for each model.

Results

Among multiple models evaluated, two yielded the best performance to predict a difficult airway both exclusively analyzing Cormack I and IV cases which showed the most distinct differences. The variables included in each model were the following: Model 1; included demographic data, vowel "A" in all positions and harmonics of the voice achieving an AUC of 0.91. Model 2; Included demographic data, vowel "O" in normal positions and voice parameters (Shimmer, Jitter, HNR); achieving in an AUC of 0.90. In contrast, models which focused on analyzing all Cormack grades (I, II, III, IV) cases performed less effectively.

Conclusions

Acoustic parameters of the voice together with the demographic data of the patients, when introduced into classification algorithms based on machine learning showed promising signs of predicting a difficult airway.

Beyond the stereotypes: Artificial Intelligence image generation and diversity in anesthesiology

Introduction

Artificial Intelligence (AI) is increasingly being integrated into anesthesiology to enhance patient safety, improve efficiency, and streamline various aspects of practice.

Objective

This study aims to evaluate whether AI-generated images accurately depict the demographic racial and ethnic diversity observed in the Anesthesia workforce and to identify inherent social biases in these images.

Methods

This cross-sectional analysis was conducted from January to February 2024. Demographic data were collected from the American Society of Anesthesiologists (ASA) and the European Society of Anesthesiology and Intensive Care (ESAIC). Two AI text-to-image models, ChatGPT DALL-E 2 and Midjourney, generated images of anesthesiologists across various subspecialties. Three independent reviewers assessed and categorized each image based on sex, race/ethnicity, age, and emotional traits.

Results

A total of 1,200 images were analyzed. We found significant discrepancies between AI-generated images and actual demographic data. The models predominantly portrayed anesthesiologists as White, with ChatGPT DALL-E2 at 64.2% and Midjourney at 83.0%. Moreover, male gender was highly associated with White ethnicity by ChatGPT DALL-E2 (79.1%) and with non-White ethnicity by Midjourney (87%). Age distribution also varied significantly, with younger anesthesiologists underrepresented. The analysis also revealed predominant traits such as "masculine, ""attractive, "and "trustworthy" across various subspecialties.

Conclusion

AI models exhibited notable biases in gender, race/ethnicity, and age representation, failing to reflect the actual diversity within the anesthesiologist workforce. These biases highlight the need for more diverse training datasets and strategies to mitigate bias in AI-generated images to ensure accurate and inclusive representations in the medical field.

Machine Learning Predictions and Identifying Key Predictors for Safer Intubation: A Study on Video Laryngoscopy Views

This study develops a predictive model for video laryngoscopic views using advanced machine learning techniques, aiming to enhance airway management's efficiency and safety. A total of 212 participants were involved, with 169 in the training set and 43 in the test set. We assessed outcomes using the percentage of glottic opening (POGO) score and considered factors like the modified Mallampati classification, thyromental height and distance, sternomental distance, mouth opening distance, and neck circumference. A range of machine learning algorithms was employed for data analysis, including Random Forest, Light Gradient Boosting Machine, K-Nearest Neighbors, Support Vector Regression, Ridge Regression, and Lasso Regression. The models' performance was evaluated on the test set, with Root Mean Squared Error values ranging from 20.4 to 21.9. SHapley Additive exPlanations value analysis revealed that age is a consistent and significant predictor of POGO score across all models, highlighting its critical role in the predictive accuracy of these techniques.

Evaluation of preoperative difficult airway prediction methods for adult patients without obvious airway abnormalities: a systematic review and meta-analysis

BACKGROUND

This systematic review aims to assist clinical decision-making in selecting appropriate preoperative prediction methods for difficult tracheal intubation by identifying and synthesizing literature on these methods in adult patients undergoing all types of surgery.

METHODS

A systematic review and meta-analysis were conducted following PRISMA guidelines. Comprehensive electronic searches across multiple databases were completed on March 28, 2023. Two researchers independently screened, selected studies, and extracted data. A total of 227 articles representing 526 studies were included and evaluated for bias using the QUADAS-2 tool. Meta-Disc software computed pooled sensitivity (SEN), specificity (SPC), positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic odds ratio (DOR). Heterogeneity was assessed using the Spearman correlation coefficient, Cochran's-Q, and I2 index, with meta-regression exploring sources of heterogeneity. Publication bias was evaluated using Deeks' funnel plot.

RESULTS

Out of 2906 articles retrieved, 227 met the inclusion criteria, encompassing a total of 686,089 patients. The review examined 11 methods for predicting difficult tracheal intubation, categorized into physical examination, multivariate scoring system, and imaging test. The modified Mallampati test (MMT) showed a SEN of 0.39 and SPC of 0.86, while the thyromental distance (TMD) had a SEN of 0.38 and SPC of 0.83. The upper lip bite test (ULBT) presented a SEN of 0.52 and SPC of 0.84. Multivariate scoring systems like LEMON and Wilson's risk score demonstrated moderate sensitivity and specificity. Imaging tests, particularly ultrasound-based methods such as the distance from the skin to the epiglottis (US-DSE), exhibited higher sensitivity (0.80) and specificity (0.77). Significant heterogeneity was identified across studies, influenced by factors such as sample size and study design.

CONCLUSION

No single preoperative prediction method shows clear superiority for predicting difficult tracheal intubation. The evidence supports a combined approach using multiple methods tailored to specific patient demographics and clinical contexts. Future research should focus on integrating advanced technologies like artificial intelligence and deep learning to improve predictive models. Standardizing testing procedures and establishing clear cut-off values are essential for enhancing prediction reliability and accuracy. Implementing a multi-modal predictive approach may reduce unanticipated difficult intubations, improving patient safety and outcomes.

Difficult Airway Assessment Based on Multi-View Metric Learning

The preoperative assessment of difficult airways is of great significance in the practice of anesthesia intubation. In recent years, although a large number of difficult airway recognition algorithms have been investigated, defects such as low recognition accuracy and poor recognition reliability still exist. In this paper, we propose a Dual-Path Multi-View Fusion Network (DMF-Net) based on multi-view metric learning, which aims to predict difficult airways through multi-view facial images of patients. DMF-Net adopts a dual-path structure to extract features by grouping the frontal and lateral images of the patients. Meanwhile, a Multi-Scale Feature Fusion Module and a Hybrid Co-Attention Module are designed to improve the feature representation ability of the model. Consistency loss and complementarity loss are utilized fully for the complementarity and consistency of information between multi-view data. Combined with Focal Loss, information bias is effectively avoided. Experimental validation illustrates the effectiveness of the proposed method, with the accuracy, specificity, sensitivity, and F1 score reaching 77.92%, 75.62%, 82.50%, and 71.35%, respectively. Compared with methods such as clinical bedside screening tests and existing artificial intelligence-based methods, our method is more accurate and reliable and can provide a reliable auxiliary tool for clinical healthcare personnel to effectively improve the accuracy and reliability of preoperative difficult airway assessments. The proposed network can help to identify and assess the risk of difficult airways in patients before surgery and reduce the incidence of postoperative complications.

Explainable artificial intelligence (XAI) for predicting the need for intubation in methanol-poisoned patients: a study comparing deep and machine learning models

The need for intubation in methanol-poisoned patients, if not predicted in time, can lead to irreparable complications and even death. Artificial intelligence (AI) techniques like machine learning (ML) and deep learning (DL) greatly aid in accurately predicting intubation needs for methanol-poisoned patients. So, our study aims to assess Explainable Artificial Intelligence (XAI) for predicting intubation necessity in methanol-poisoned patients, comparing deep learning and machine learning models. This study analyzed a dataset of 897 patient records from Loghman Hakim Hospital in Tehran, Iran, encompassing cases of methanol poisoning, including those requiring intubation (202 cases) and those not requiring it (695 cases). Eight established ML (SVM, XGB, DT, RF) and DL (DNN, FNN, LSTM, CNN) models were used. Techniques such as tenfold cross-validation and hyperparameter tuning were applied to prevent overfitting. The study also focused on interpretability through SHAP and LIME methods. Model performance was evaluated based on accuracy, specificity, sensitivity, F1-score, and ROC curve metrics. Among DL models, LSTM showed superior performance in accuracy (94.0%), sensitivity (99.0%), specificity (94.0%), and F1-score (97.0%). CNN led in ROC with 78.0%. For ML models, RF excelled in accuracy (97.0%) and specificity (100%), followed by XGB with sensitivity (99.37%), F1-score (98.27%), and ROC (96.08%). Overall, RF and XGB outperformed other models, with accuracy (97.0%) and specificity (100%) for RF, and sensitivity (99.37%), F1-score (98.27%), and ROC (96.08%) for XGB. ML models surpassed DL models across all metrics, with accuracies from 93.0% to 97.0% for DL and 93.0% to 99.0% for ML. Sensitivities ranged from 98.0% to 99.37% for DL and 93.0% to 99.0% for ML. DL models achieved specificities from 78.0% to 94.0%, while ML models ranged from 93.0% to 100%. F1-scores for DL were between 93.0% and 97.0%, and for ML between 96.0% and 98.27%. DL models scored ROC between 68.0% and 78.0%, while ML models ranged from 84.0% to 96.08%. Key features for predicting intubation necessity include GCS at admission, ICU admission, age, longer folic acid therapy duration, elevated BUN and AST levels, VBG_HCO3 at initial record, and hemodialysis presence. This study as the showcases XAI's effectiveness in predicting intubation necessity in methanol-poisoned patients. ML models, particularly RF and XGB, outperform DL counterparts, underscoring their potential for clinical decision-making.

Artificial intelligence in pediatric airway - A scoping review

Artificial intelligence is an ever-growing modality revolutionizing the field of medical science. It utilizes various computational models and algorithms and helps out in different sectors of healthcare. Here, in this scoping review, we are trying to evaluate the use of Artificial intelligence (AI) in the field of pediatric anesthesia, specifically in the more challenging domain, the pediatric airway. Different components within the domain of AI include machine learning, neural networks, deep learning, robotics, and computer vision. Electronic databases like Google Scholar, Cochrane databases, and Pubmed were searched. Different studies had heterogeneity of age groups, so all studies with children under 18 years of age were included and assessed. The use of AI was reviewed in the preoperative, intraoperative, and postoperative domains of pediatric anesthesia. The applicability of AI needs to be supplemented by clinical judgment for the final anticipation in various fields of medicine.

Improving difficult direct laryngoscopy prediction using deep learning and minimal image analysis: a single-center prospective study

Accurate prediction of difficult direct laryngoscopy (DDL) is essential to ensure optimal airway management and patient safety. The present study proposed an AI model that would accurately predict DDL using a small number of bedside pictures of the patient's face and neck taken simply with a smartphone. In this prospective single-center study, adult patients scheduled for endotracheal intubation under general anesthesia were included. Patient pictures were obtained in frontal, lateral, frontal-neck extension, and open mouth views. DDL prediction was performed using a deep learning model based on the EfficientNet-B5 architecture, incorporating picture view information through multitask learning. We collected 18,163 pictures from 3053 patients. After under-sampling to achieve a 1:1 image ratio of DDL to non-DDL, the model was trained and validated with a dataset of 6616 pictures from 1283 patients. The deep learning model achieved a receiver operating characteristic area under the curve of 0.81-0.88 and an F1-score of 0.72-0.81 for DDL prediction. Including picture view information improved the model's performance. Gradient-weighted class activation mapping revealed that neck and chin characteristics in frontal and lateral views are important factors in DDL prediction. The deep learning model we developed effectively predicts DDL and requires only a small set of patient pictures taken with a smartphone. The method is practical and easy to implement.

Personalized anesthesia and precision medicine: a comprehensive review of genetic factors, artificial intelligence, and patient-specific factors

Precision medicine, characterized by the personalized integration of a patient's genetic blueprint and clinical history, represents a dynamic paradigm in healthcare evolution. The emerging field of personalized anesthesia is at the intersection of genetics and anesthesiology, where anesthetic care will be tailored to an individual's genetic make-up, comorbidities and patient-specific factors. Genomics and biomarkers can provide more accurate anesthetic protocols, while artificial intelligence can simplify anesthetic procedures and reduce anesthetic risks, and real-time monitoring tools can improve perioperative safety and efficacy. The aim of this paper is to present and summarize the applications of these related fields in anesthesiology by reviewing them, exploring the potential of advanced technologies in the implementation and development of personalized anesthesia, realizing the future integration of new technologies into clinical practice, and promoting multidisciplinary collaboration between anesthesiology and disciplines such as genomics and artificial intelligence.

Reliable prediction of difficult airway for tracheal intubation from patient preoperative photographs by machine learning methods

BACKGROUND

Estimating the risk of a difficult tracheal intubation should help clinicians in better anaesthesia planning, to maximize patient safety. Routine bedside screenings suffer from low sensitivity.

OBJECTIVE

To develop and evaluate machine learning (ML) and deep learning (DL) algorithms for the reliable prediction of intubation risk, using information about airway morphology.

METHODS

Observational, prospective cohort study enrolling n=623 patients who underwent tracheal intubation: 53/623 difficult cases (prevalence 8.51%). First, we used our previously validated deep convolutional neural network (DCNN) to extract 2D image coordinates for 27 + 13 relevant anatomical landmarks in two preoperative photos (frontal and lateral views). Here we propose a method to determine the 3D pose of the camera with respect to the patient and to obtain the 3D world coordinates of these landmarks. Then we compute a novel set of dM=59 morphological features (distances, areas, angles and ratios), engineered with our anaesthesiologists to characterize each individual's airway anatomy towards prediction. Subsequently, here we propose four ad hoc ML pipelines for difficult intubation prognosis, each with four stages: feature scaling, imputation, resampling for imbalanced learning, and binary classification (Logistic Regression, Support Vector Machines, Random Forests and eXtreme Gradient Boosting). These compound ML pipelines were fed with the dM=59 morphological features, alongside dD=7 demographic variables. Here we trained them with automatic hyperparameter tuning (Bayesian search) and probability calibration (Platt scaling). In addition, we developed an ad hoc multi-input DCNN to estimate the intubation risk directly from each pair of photographs, i.e. without any intermediate morphological description. Performance was evaluated using optimal Bayesian decision theory. It was compared against experts' judgement and against state-of-the-art methods (three clinical formulae, four ML, four DL models).

RESULTS

Our four ad hoc ML pipelines with engineered morphological features achieved similar discrimination capabilities: median AUCs between 0.746 and 0.766. They significantly outperformed both expert judgement and all state-of-the-art methods (highest AUC at 0.716). Conversely, our multi-input DCNN yielded low performance due to overfitting. This same behaviour occurred for the state-of-the-art DL algorithms. Overall, the best method was our XGB pipeline, with the fewest false negatives at the optimal Bayesian decision threshold.

CONCLUSIONS

We proposed and validated ML models to assist clinicians in anaesthesia planning, providing a reliable calibrated estimate of airway intubation risk, which outperformed expert assessments and state-of-the-art methods. Our novel set of engineered features succeeded in providing informative descriptions for prognosis.

Deep learning-based facial analysis for predicting difficult videolaryngoscopy: a feasibility study

While videolaryngoscopy has resulted in better overall success rates of tracheal intubation, airway assessment is still an important prerequisite for safe airway management. This study aimed to create an artificial intelligence model to identify difficult videolaryngoscopy using a neural network. Baseline characteristics, medical history, bedside examination and seven facial images were included as predictor variables. ResNet-18 was introduced to recognise images and extract features. Different machine learning algorithms were utilised to develop predictive models. A videolaryngoscopy view of Cormack-Lehane grade of 1 or 2 was classified as 'non-difficult', while grade 3 or 4 was classified as 'difficult'. A total of 5849 patients were included, of whom 5335 had non-difficult and 514 had difficult videolaryngoscopy. The facial model (only including facial images) using the Light Gradient Boosting Machine algorithm showed the highest area under the curve (95%CI) of 0.779 (0.733-0.825) with a sensitivity (95%CI) of 0.757 (0.650-0.845) and specificity (95%CI) of 0.721 (0.626-0.794) in the test set. Compared with bedside examination and multivariate scores (El-Ganzouri and Wilson), the facial model had significantly higher predictive performance (p < 0.001). Artificial intelligence-based facial analysis is a feasible technique for predicting difficulty during videolaryngoscopy, and the model developed using neural networks has higher predictive performance than traditional methods.

A review of the current status and progress in difficult airway assessment research

A difficult airway is a situation in which an anesthesiologist with more than 5 years of experience encounters difficulty with intubation or mask ventilation. According to the 2022 American Society of Anesthesiologists Practice Guidelines for the Management of Difficult Airway, difficult airways are subdivided into seven detailed categories. This condition can lead to serious adverse events and therefore must be diagnosed accurately and quickly. In this review, we comprehensively summarize and discuss the different methods used in clinical practice and research to assess difficult airways, including medical history, simple bedside assessment, comprehensive assessment of indicators, preoperative endoscopic airway examination, imaging, computer-assisted airway reconstruction, and 3D-printing techniques. We also discuss in detail the latest trends in difficult airway assessment through mathematical methods and artificial intelligence. With the continuous development of artificial intelligence and other technologies, in the near future, we will be able to predict whether a patient has a difficult airway simply by taking an image of the patient's face through a cell phone program. Artificial intelligence and other technologies will bring great changes to the development of airway assessment, and at the same time raise some new questions that we should think about.

Scope of artificial intelligence in airway management

The evolution of artificial intelligence (AI) systems in the field of anaesthesiology owes to notable advancements in data processing, databases, algorithmic programs, and computation power. Over the past decades, its accelerated progression has enhanced safety in anaesthesia by improving the efficiency of equipment, perioperative risk assessments, monitoring, and drug administration systems. AI in the field of anaesthesia aims to improve patient safety, optimise resources, and improve the quality of anaesthesia management in all phases of perioperative care. The use of AI is likely to impact difficult airway management and patient safety considerably. AI has been explored to predict difficult intubation to outperform conventional airway examinations by integrating subjective factors, such as facial appearance, speech features, habitus, and other poorly known features. This narrative review delves into the status of AI in airway management, the most recent developments in this field, and its future clinical applications.

Remote, automatic, digital preanesthetic evaluation - are we there yet?

Recent years have witnessed multiple advancements in the field of information technology in medicine. The need to ensure patient and doctor safety during COVID-19 resulted in improved telemedicine adaptation across various fields, including anaesthesiology. In this review, the authors examine the current state of the elements of preanesthetic evaluation and their remote execution using current and future telemedical facilities and technologies, as well as the potential of future advancements in this field.

New Devices, Innovative Technologies, and Non-Standard Techniques for Airway Management: A Narrative Review

A wide range of airway devices and techniques have been created to enhance the safety of airway management. However, airway management remains a challenge. All techniques are susceptible to failure. Therefore, it is necessary to have and know the greatest number of alternatives to treat even the most challenging airway successfully. The aim of this narrative review is to describe some new devices, such as video laryngeal masks, articulated stylets, and non-standard techniques, for laryngeal mask insertion and endotracheal intubation that are not applied in daily practice, but that could be highly effective in overcoming a difficulty related to airway management. Artificial intelligence and 3D technology for airway management are also discussed.

Artificial intelligence in anesthesiology

Artificial intelligence (AI) is rapidly evolving and gaining attention in the medical world. Our aim is to provide readers with insights into this quickly changing medical landscape and the role of clinicians in the middle of this popular technology. In this review, our aim is to explain some of the increasingly frequently used AI terminology explicitly for physicians. Next, we give a summation, an overview of currently existing applications, future possibilities for AI in the medical field of anesthesiology and thoroughly highlight possible problems that could arise from implementing this technology in daily practice. Therefore, we conducted a literature search, including all types of articles published between the first of January 2010 and the 1st of May 2023, written in English, and having a free full text available. We searched Pubmed, Medline, and Embase using “artificial intelligence”, “machine learning”, “deep learning”, “neural networks” and “anesthesiology” as MESH terms. To structure these findings, we divided the results into five categories: preoperatively, perioperatively, postoperatively, AI in the intensive care unit and finally, AI used for teaching purposes. In the first category, we found AI applications for airway assessment, risk prediction, and logistic support. Secondly, we made a summation of AI applications used during the operation. AI can predict hypotensive events, delivering automated anesthesia, reducing false alarms, and aiding in the analysis of ultrasound anatomy in locoregional anesthesia and echocardiography. Thirdly, namely postoperatively, AI can be applied in predicting acute kidney injury, pulmonary complications, postoperative cognitive dysfunction and can help to diagnose postoperative pain in children. At the intensive care unit, AI tools discriminate acute respiratory distress syndrome (ARDS) from pulmonary oedema in pleural ultrasound, predict mortality and sepsis more accurately, and predict survival rates in severe Coronavirus-19 (COVID-19). Finally, AI has been described in training residents in spinal ultrasound, simulation, and plexus block anatomy. Several concerns must be addressed regarding the use of AI. Firstly, this software does not explain its decision process (i.e., the ‘black box problem’). Secondly, to develop AI models and decision support systems, we need big and accurate datasets, unfortunately with potential unknown bias. Thirdly, we need an ethical and legal framework before implementing this technology. At the end of this paper, we discuss whether this technology will be able to replace the clinician one day. This paper adds value to already existing literature because it not only offers a summation of existing literature on AI applications in anesthesiology but also gives clear definitions of AI itself and critically assesses implementation of this technology.

Geometric morphometrics and machine learning from three-dimensional facial scans for difficult mask ventilation prediction

Background

Unanticipated difficult mask ventilation (DMV) is a potentially life-threatening event in anesthesia. Nevertheless, predicting DMV currently remains a challenge. This study aimed to verify whether three dimensional (3D) facial scans could predict DMV in patients scheduled for general anesthesia.

Methods

The 3D facial scans were taken on 669 adult patients scheduled for elective surgery under general anesthesia. Clinical variables currently used as predictors of DMV were also collected. The DMV was defined as the inability to provide adequate and stable ventilation. Spatially dense landmarks were digitized on 3D scans to describe sufficient details for facial features and then processed by 3D geometric morphometrics. Ten different machine learning (ML) algorithms, varying from simple to more advanced, were introduced. The performance of ML models for DMV prediction was compared with that of the DIFFMASK score. The area under the receiver operating characteristic curves (AUC) with its 95% confidence interval (95% CI) as well as the specificity and sensitivity were used to evaluate the predictive value of the model.

Results

The incidence of DMV was 35/669 (5.23%). The logistic regression (LR) model performed best among the 10 ML models. The AUC of the LR model was 0.825 (95% CI, 0.765-0.885). The sensitivity and specificity of the model were 0.829 (95% CI, 0.629-0.914) and 0.733 (95% CI, 0.532-0.819), respectively. The LR model demonstrated better predictive performance than the DIFFMASK score, which obtained an AUC of 0.785 (95% CI, 0.710-0.860) and a sensitivity of 0.686 (95% CI, 0.578-0.847). Notably, we identified a significant morphological difference in the mandibular region between the DMV group and the easy mask ventilation group.

Conclusion

Our study indicated a distinct morphological difference in the mandibular region between the DMV group and the easy mask ventilation group. 3D geometric morphometrics with ML could be a rapid, efficient, and non-invasive tool for DMV prediction to improve anesthesia safety.

Expert consensus on difficult airway assessment

Background

Identifying a potentially difficult airway is crucial both in anaesthesia in the operating room (OR) and non-operation room sites. There are no guidelines or expert consensus focused on the assessment of the difficult airway before, so this expert consensus is developed to provide guidance for airway assessment, making this process more standardized and accurate to reduce airway-related complications and improve safety.

Methods

Seven members from the Airway Management Group of the Chinese Society of Anaesthesiology (CSA) met to discuss the first draft and then this was sent to 15 international experts for review, comment, and approval. The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) is used to determine the level of evidence and grade the strength of recommendations. The recommendations were revised through a three-round Delphi survey from experts.

Results

This expert consensus provides a comprehensive approach to airway assessment based on the medical history, physical examination, comprehensive scores, imaging, and new developments including transnasal endoscopy, virtual laryngoscopy, and 3D printing. In addition, this consensus also reviews some new technologies currently under development such as prediction from facial images and voice information with the aim of proposing new research directions for the assessment of difficult airway.

Conclusions

This consensus applies to anesthesiologists, critical care, and emergency physicians refining the preoperative airway assessment and preparing an appropriate intubation strategy for patients with a potentially difficult airway.

Endotracheal Intubation of Difficult Airways in Emergency Settings: A Guide for Innovators

Over 400,000 Americans are intubated in emergency settings annually, with indications ranging from respiratory failure to airway obstructions to anaphylaxis. About 12.7% of emergency intubations are unsuccessful on the first attempt. Failure to intubate on the first attempt is associated with a higher likelihood of adverse events, including oxygen desaturation, aspiration, trauma to soft tissue, dysrhythmia, hypotension, and cardiac arrest. Difficult airways, as classified on an established clinical scale, are found in up to 30% of emergency department (ED) patients and are a significant contributor to failure to intubate. Difficult intubations have been associated with longer lengths of stay and significantly greater costs than standard intubations. There exists a wide range of airway management devices, both invasive and noninvasive, which are available in the emergency setting to accommodate difficult airways. Yet, first-pass success rates remain variable and leave room for improvement. In this article, we review the disease states most correlated with intubation, the current landscape of emergency airway management technologies, and the market potential for innovation. The aim of this review is to inspire new technologies to assist difficult airway management, given the substantial opportunity for translation due to two key-value signposts of medical innovation: the potential to decrease cost and the potential to improve clinical outcomes.

Emerging Robotic Innovations and Artificial Intelligence in Endotracheal Intubation and Airway Management: Current State of the Art

Robotic sciences have rapidly advanced and revolutionized various aspects of medicine, including the field of airway management. Robotic endotracheal intubation is an innovative method that utilizes robotic systems to aid in the accurate placement of an endotracheal tube within the trachea. This cutting-edge technique shows great promise in improving procedural precision and ensuring patient safety. In this comprehensive overview, we delve into the present status of robotic-assisted endotracheal intubation, examining its advantages, obstacles, and the potential implications it holds for the future. In addition, this review encompasses a comprehensive analysis of the existing literature and references on recent advances in robotic technology and artificial intelligence related to airway management.

Predicting case difficulty in endodontic microsurgery using machine learning algorithms

OBJECTIVES

The study aimed to develop and validate machine learning models for case difficulty prediction in endodontic microsurgery, assisting clinicians in preoperative analysis.

METHODS

The cone-beam computed tomographic images were collected from 261 patients with 341 teeth and used for radiographic examination and measurement. Through linear regression (LR), support vector regression (SVR), and extreme gradient boosting (XGBoost) algorithms, four models were established according to different loss functions, including the L1-loss LR model, L2-loss LR model, SVR model and XGBoost model. Five-fold cross-validation was applied in model training and validation. Explained variance score (EVS), coefficient of determination (R²), mean absolute error (MAE), mean squared error (MSE) and median absolute error (MedAE) were calculated to evaluate the prediction performance.

RESULTS

The MAE, MSE and MedAE values ​​of the XGBoost model were the lowest, which were 0.1010, 0.0391 and 0.0235, respectively. The EVS and R² values ​​of the XGBoost model were the highest, which were 0.7885 and 0.7967, respectively. The factors used to predict the case difficulty in endodontic microsurgery were ordered according to their relative importance, including lesion size, the distance between apex and adjacent important anatomical structures, root filling density, root apex diameter, root resorption, tooth type, tooth length, root filling length, root canal curvature and the number of root canals.

CONCLUSIONS

The XGBoost model outperformed the LR and SVR models on all evaluation metrics, which can assist clinicians in preoperative analysis. The relative feature importance provides a reference to develop the scoring system for case difficulty assessment in endodontic microsurgery.

CLINICAL SIGNIFICANCE

Preoperative case assessment is a crucial step to identify potential risks and make referral decisions. Machine learning models for case difficulty prediction in endodontic microsurgery can assist clinicians in preoperative analysis efficiently and accurately.

Automated location of orofacial landmarks to characterize airway morphology in anaesthesia via deep convolutional neural networks

BACKGROUND

A reliable anticipation of a difficult airway may notably enhance safety during anaesthesia. In current practice, clinicians use bedside screenings by manual measurements of patients' morphology.

OBJECTIVE

To develop and evaluate algorithms for the automated extraction of orofacial landmarks, which characterize airway morphology.

METHODS

We defined 27 frontal + 13 lateral landmarks. We collected n=317 pairs of pre-surgery photos from patients undergoing general anaesthesia (140 females, 177 males). As ground truth reference for supervised learning, landmarks were independently annotated by two anaesthesiologists. We trained two ad-hoc deep convolutional neural network architectures based on InceptionResNetV2 (IRNet) and MobileNetV2 (MNet), to predict simultaneously: (a) whether each landmark is visible or not (occluded, out of frame), (b) its 2D-coordinates (x,y). We implemented successive stages of transfer learning, combined with data augmentation. We added custom top layers on top of these networks, whose weights were fully tuned for our application. Performance in landmark extraction was evaluated by 10-fold cross-validation (CV) and compared against 5 state-of-the-art deformable models.

RESULTS

With annotators' consensus as the 'gold standard', our IRNet-based network performed comparably to humans in the frontal view: median CV loss L=1.277·10^-3, inter-quartile range (IQR) [1.001, 1.660]; versus median 1.360, IQR [1.172, 1.651], and median 1.352, IQR [1.172, 1.619], for each annotator against consensus, respectively. MNet yielded slightly worse results: median 1.471, IQR [1.139, 1.982]. In the lateral view, both networks attained performances statistically poorer than humans: median CV loss L=2.141·10^-3, IQR [1.676, 2.915], and median 2.611, IQR [1.898, 3.535], respectively; versus median 1.507, IQR [1.188, 1.988], and median 1.442, IQR [1.147, 2.010] for both annotators. However, standardized effect sizes in CV loss were small: 0.0322 and 0.0235 (non-significant) for IRNet, 0.1431 and 0.1518 (p<0.05) for MNet; therefore quantitatively similar to humans. The best performing state-of-the-art model (a deformable regularized Supervised Descent Method, SDM) behaved comparably to our DCNNs in the frontal scenario, but notoriously worse in the lateral view.

CONCLUSIONS

We successfully trained two DCNN models for the recognition of 27 + 13 orofacial landmarks pertaining to the airway. Using transfer learning and data augmentation, they were able to generalize without overfitting, reaching expert-like performances in CV. Our IRNet-based methodology achieved a satisfactory identification and location of landmarks: particularly in the frontal view, at the level of anaesthesiologists. In the lateral view, its performance decayed, although with a non-significant effect size. Independent authors had also reported lower lateral performances; as certain landmarks may not be clear salient points, even for a trained human eye.

Attitudes of Anesthesiologists toward Artificial Intelligence in Anesthesia: A Multicenter, Mixed Qualitative-Quantitative Study

Artificial intelligence (AI) is predicted to play an increasingly important role in perioperative medicine in the very near future. However, little is known about what anesthesiologists know and think about AI in this context. This is important because the successful introduction of new technologies depends on the understanding and cooperation of end users. We sought to investigate how much anesthesiologists know about AI and what they think about the introduction of AI-based technologies into the clinical setting. In order to better understand what anesthesiologists think of AI, we recruited 21 anesthesiologists from 2 university hospitals for face-to-face structured interviews. The interview transcripts were subdivided sentence-by-sentence into discrete statements, and statements were then grouped into key themes. Subsequently, a survey of closed questions based on these themes was sent to 70 anesthesiologists from 3 university hospitals for rating. In the interviews, the base level of knowledge of AI was good at 86 of 90 statements (96%), although awareness of the potential applications of AI in anesthesia was poor at only 7 of 42 statements (17%). Regarding the implementation of AI in anesthesia, statements were split roughly evenly between pros (46 of 105, 44%) and cons (59 of 105, 56%). Interviewees considered that AI could usefully be used in diverse tasks such as risk stratification, the prediction of vital sign changes, or as a treatment guide. The validity of these themes was probed in a follow-up survey of 70 anesthesiologists with a response rate of 70%, which confirmed an overall positive view of AI in this group. Anesthesiologists hold a range of opinions, both positive and negative, regarding the application of AI in their field of work. Survey-based studies do not always uncover the full breadth of nuance of opinion amongst clinicians. Engagement with specific concerns, both technical and ethical, will prove important as this technology moves from research to the clinic.

New technologies in airway management: A review

The evolution of medical knowledge and technological growth have contributed to the development of different techniques and devices for airway management. These appear to play a role in optimizing the number of attempts and overall success, ultimately reducing the negative consequences of airway manipulation. In this literature review, we highlight the recent evidence regarding new technologies applied to airway management. Before intubation, every patient should have an individualized structured airway management plan. Technology can help with both airway evaluation and tracheal intubation. Point-of-care cervical ultrasound and artificial intelligence models with automated facial analysis have been used to predict difficult airways. Various devices can be used in airway management. This includes a robotic video endoscope that guides intubation based on real image recognition, a laryngeal mask with a non-inflatable cuff that tries to reduce local complications, video laryngeal masks that are able to confirm the correct position and facilitate intubation, Viescope™, a videolaryngoscope developed for combat medicine with a unique circular blade, a system that uses cervical transillumination for glottis identification in difficult airways and Vivasight SL™ tracheal tube, which has a high-resolution camera at its tip guaranteeing visual assurance of tube position as well as guiding bronchial blocker position. To conclude, we detailed the challenges in airway management outside the operating room as well as described suction-assisted laryngoscopy and airway decontamination technique for contaminated airways. Further research in the clinical setting is recommended to better support the use of these technologies.

Performance of novice intubators in using direct laryngoscope with 3 stylets on a manikin model

BACKGROUND

Tracheal intubation is an important clinical skill for medical students and junior residents (novice intubators). They are usually trained to use a direct laryngoscope (DL) with straight-to-cuff styletted tracheal tubes first. Only later are they exposed to the bougie as an airway adjunct and videolaryngoscope (VL) with either a standard blade or a hyperangulated blade. The purpose of this study was to investigate the performance of novice intubators in using DL with 3 common stylets.

METHODS

We conducted a prospective study to compare the performance of DL with 3 common stylets, namely the straight-to-cuff stylet (S), hyperangulated VL stylet (G), and bougie (B), on a manikin model.

RESULTS

Among 72 participants, no significant difference was observed between the success rates of S, G, and B at the first attempt (84.72%, 81.94%, and 86.11%, respectively [P = .78]) or within 2 minutes (91.67%, 93.06%, and 91.67%, respectively [P = .94]). For participants with successful intubation within 2 minutes, the average total intubation times for S, G, and B were 25.05, 24.39, and 37.45 seconds, respectively. Among the 3 stylets, B had the longest intubation time, which differed significantly from S and G (P < .01).

CONCLUSIONS

The performances of novice intubators with 3 different stylets were similar. The success rates for DL with either hyperangulated VL stylet or bougie were not inferior compared with the straight-to-cuff stylet on manikin airway training model. If we properly trained novice intubators to use corresponding maneuvers, they can learn to use the 3 stylets early in their airway learning course.

Predicting difficult airway intubation in thyroid surgery using multiple machine learning and deep learning algorithms

BACKGROUND

In this paper, we examine whether machine learning and deep learning can be used to predict difficult airway intubation in patients undergoing thyroid surgery.

METHODS

We used 10 machine learning and deep learning algorithms to establish a corresponding model through a training group, and then verify the results in a test group. We used R for the statistical analysis and constructed the machine learning prediction model in Python.

RESULTS

The top 5 weighting factors for difficult airways identified by the average algorithm in machine learning were age, sex, weight, height, and BMI. In the training group, the AUC values and accuracy and the Gradient Boosting precision were 0.932, 0.929, and 100%, respectively. As for the modeled effects of predicting difficult airways in test groups, among the models constructed by the 10 algorithms, the three algorithms with the highest AUC values were Gradient Boosting, CNN, and LGBM, with values of 0.848, 0.836, and 0.812, respectively; In addition, among the algorithms, Gradient Boosting had the highest accuracy with a value of 0.913; Additionally, among the algorithms, the Gradient Boosting algorithm had the highest precision with a value of 100%.

CONCLUSION

According to our results, Gradient Boosting performed best overall, with an AUC >0.8, an accuracy >90%, and a precision of 100%. Besides, the top 5 weighting factors identified by the average algorithm in machine learning for difficult airways were age, sex, weight, height, and BMI.