Automatic Detection of Airway Invasion from Videofluoroscopy via Deep Learning Technology

Seong Jae Lee,Sang-Il Choi,Hyun Il Kim,Joo Young Ko

doi:10.3390/app10186179

Abstract

In dysphagia, food materials frequently invade the laryngeal airway, potentially resulting in serious consequences, such as asphyxia or pneumonia. The VFSS (videofluoroscopic swallowing study) procedure can be used to visualize the occurrence of airway invasion, but its reliability is limited by human errors and fatigue. Deep learning technology may improve the efficiency and reliability of VFSS analysis by reducing the human effort required. A deep learning model has been developed that can detect airway invasion from VFSS images in a fully automated manner. The model consists of three phases: (1) image normalization, (2) dynamic ROI (region of interest) determination, and (3) airway invasion detection. Noise induced by movement and learning from unintended areas is minimized by defining a “dynamic” ROI with respect to the center of the cervical spinal column as segmented using U-Net. An Xception module, trained on a dataset consisting of 267,748 image frames obtained from 319 VFSS video files, is used for the detection of airway invasion. The present model shows an overall accuracy of 97.2% in classifying image frames and 93.2% in classifying video files. It is anticipated that the present model will enable more accurate analysis of VFSS data.

Highlights

Dysphagia can occur as a result of a wide variety of neurological, structural, or psychiatric disorders [1]
The present study focuses on the detection of airway invasion that has the utmost clinical importance in the analysis of VFSS images
We evaluated the performance of the U-Net model for the global localization of the cervical spinal column and the Xception-based CNN for airway invasion detection on the basis of the receiver operating characteristic (ROC) curve, which is widely used in detection problems

Summary

Introduction

Dysphagia can occur as a result of a wide variety of neurological, structural, or psychiatric disorders [1]. Comprehensive and objective evaluation of swallowing function is of vital importance for choosing appropriate measures to prevent or decrease the mortality and morbidity of patients with dysphagia [3]. The VFSS (videofluoroscopic swallowing study) procedure, known as an MBS (modified barium swallowing) examination, is regarded as the gold standard for the evaluation of dysphagia and is the examination method that is most commonly used by clinicians for this purpose [3]. The VFSS is a radiographic procedure in which fluoroscopic images are acquired while the patient swallows boluses mixed with contrast media impenetrable to radiation beams [1]. The VFSS permits the direct visualization of bolus flow and the detection of the occurrence and timing of airway invasion, such as laryngeal penetration or aspiration, and it assists in identifying the physiological and often treatable causes of dysphagia [3]. Intra-rater and inter-rater reliability have been reported to be only poor to fair

Objectives

Methods

Results

Discussion

Conclusion