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

Abstract

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

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

“…For example, in a study on predicting difficult airways for thyroid surgery, the authors used 10 AI algorithms trained on labeled input features, ultimately concluding that age, gender, weight, height, and body mass index were the five most important factors in identifying difficult airways. However, this method of predicting difficult airways is still semi-automated and requires us to collect, extract and input data, not quite the same as the fully automated analysis we originally envisioned [ 50 ]. Another direction to predict difficult airways through AI is based on the digitization of artificial intelligence and the number and availability of medical images as a source of data [ 51 ].…”

Construction of Prediction Model of Early Glottic Cancer Based on machine learning

“…Accuracy, precision (representing positive and negative predictive values in digital diagnostic tests), recall (indicating the performance of binary sensitivity) and F1 scores (harmonic factor) are used to test the model, and receiver operating characteristics (ROC) are used to control the performance of the above prediction model [21] . The ROC curve is a graphical description, which shows the diagnostic ability when the discrimination threshold of the binary classi er system changes.…”

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

“…TMD, SMD, MO-IIG or neck circumference) -these latter magnitudes having been measured manually, as in the bedside screenings from current clinical practise. Works in this category include Yan et al [11] (who trained a Support Vector Machine), Langeron et al [12] (who employed Tree Boosting), Kim et al [13] (5 ML algorithms, among which Random Forests performed best), Yamanaka et al [14] (with an ensemble of 7 ML algorithms), or Zhou et al [15] (10 algorithms, with Gradient Boosting as their best).…”