Predicting Cormack classification based on neural network with multiple anthropometric features

Cormack classification is believed as a golden indicator for predicting tracheal intubation is difficult or not in clinic. Some anaesthetists usually estimate the airway state by examining single airway features. However, specialists agree that prediction accuracy of a difficult airway may be improved if multiple static and dynamic metrical airway features were considered. In this paper, we developed a medical decision support system based on multilayer perceptron network for Cormark classification predication with 13 input features. A tracheal intubation database consisting of 824 cases was used to train and test the system. The results showed that the multilayer perceptron based decision support system we proposed could achieve 91.9% average classification accuracy, manifesting its great application prospect of supporting clinic aided diagnosis with full consideration of multiple features of airway physical examination.