Using Transfer Learning Method to Develop an Artificial Intelligence Assisted Triaging for Endotracheal Tube Position on Chest X-ray.

Kuo-Ching Yuan,Lung-Wen Tsai,Kevin Lai,Yu-Sheng Lo,Syu-Jyun Peng,Sing-Teck Teng

doi:10.3390/diagnostics11101844

Abstract

Endotracheal tubes (ETTs) provide a vital connection between the ventilator and patient; however, improper placement can hinder ventilation efficiency or injure the patient. Chest X-ray (CXR) is the most common approach to confirming ETT placement; however, technicians require considerable expertise in the interpretation of CXRs, and formal reports are often delayed. In this study, we developed an artificial intelligence-based triage system to enable the automated assessment of ETT placement in CXRs. Three intensivists performed a review of 4293 CXRs obtained from 2568 ICU patients. The CXRs were labeled “CORRECT” or “INCORRECT” in accordance with ETT placement. A region of interest (ROI) was also cropped out, including the bilateral head of the clavicle, the carina, and the tip of the ETT. Transfer learning was used to train four pre-trained models (VGG16, INCEPTION_V3, RESNET, and DENSENET169) and two models developed in the current study (VGG16_Tensor Projection Layer and CNN_Tensor Projection Layer) with the aim of differentiating the placement of ETTs. Only VGG16 based on ROI images presented acceptable performance (AUROC = 92%, F1 score = 0.87). The results obtained in this study demonstrate the feasibility of using the transfer learning method in the development of AI models by which to assess the placement of ETTs in CXRs.

Highlights

Mechanical ventilation is a life support modality commonly used in intensive care units (ICUs) for a wide range of situations, from scheduled surgical procedures to acute organ failure [1]
We developed a novel convolutional neural network (CNN) with a tensor projection layer (CNN_TPL) and VGG16 with a tensor projection layer (VGG16_TPL)
We developed an artificial intelligence (AI)-based triage system to assess the placement of Endotracheal tubes (ETTs) in Chest X-ray (CXR)

Summary

Introduction

Mechanical ventilation is a life support modality commonly used in intensive care units (ICUs) for a wide range of situations, from scheduled surgical procedures to acute organ failure [1]. Mechanical ventilation requires an artificial connection between the ventilator and the patient’s airway, involving tracheostomy, a jet needle, or most commonly an endotracheal tube (ETT) [3]. A number of methods have been developed to confirm ETT placement using a stethoscope, end-tidal CO2 levels, or portable chest X-rays (CXRs). Portable CXRs are currently the gold standard to confirm ETT placement, due to the fact that they are highly informative, inexpensive, and immediately available at the patient’s bedside in any location of the hospital [4]. The overwhelming workload of experienced radiologists often delays the preparation of formal reports. This situation has led to the development of various point-of-care methods to facilitate the timely assessment of tube placement [5]

Methods

Results

Discussion

Conclusion