Assessment of Critical Feeding Tube Malpositions on Radiographs Using Deep Learning

Abstract

Assess the efficacy of deep convolutional neural networks (DCNNs) in detection of critical enteric feeding tube malpositions on radiographs. 5475 de-identified HIPAA compliant frontal view chest and abdominal radiographs were obtained, consisting of 174 x-rays of bronchial insertions and 5301 non-critical radiographs, including normal course, normal chest, and normal abdominal x-rays. The ground-truth classification for enteric feeding tube placement was performed by two board-certified radiologists. Untrained and pretrained deep convolutional neural network models for Inception V3, ResNet50, and DenseNet 121 were each employed. The radiographs were fed into each deep convolutional neural network, which included untrained and pretrained models. The Tensorflow framework was used for Inception V3, ResNet50, and DenseNet. Images were split into training (4745), validation (630), and test (100). Both real-time and preprocessing image augmentation strategies were performed. Receiver operating characteristic (ROC) and area under the curve (AUC) on the test data were used to assess the models. Statistical differences among the AUCs were obtained. p < 0.05 was considered statistically significant. The pretrained Inception V3, which had an AUC of 0.87 (95 CI; 0.80–0.94), performed statistically significantly better ( p < .001) than the untrained Inception V3, with an AUC of 0.60 (95 CI; 0.52–0.68). The pretrained Inception V3 also had the highest AUC overall, as compared with ResNet50 and DenseNet121, with AUC values ranging from 0.82 to 0.85. Each pretrained network outperformed its untrained counterpart. ( p < 0.05). Deep learning demonstrates promise in differentiating critical vs. non-critical placement with an AUC of 0.87. Pretrained networks outperformed untrained ones in all cases. DCNNs may allow for more rapid identification and communication of critical feeding tube malpositions.

Effect of a comprehensive deep-learning model on the accuracy of chest x-ray interpretation by radiologists: a retrospective, multireader multicase study

“…Furthermore, lateral radiographs are often not assessed, despite clear evidence that they contain clinically important information. 11 Deep-learning chest x-ray analysis systems have been developed to automate lung segmentation and bone exclusion; 12 diagnose tuberculosis; 13 detect pneumonia, 14,15 COVID-19, 16 pneumothorax, 17 pneumoconiosis, 18 and lung cancer; 19 identify the position of feeding tubes; 20 and to predict temporal changes in imaging findings. 21 Deeplearning diagnostic tools have also been shown to improve the classification accuracy of radiologists in the detection of pulmonary nodules, 22 pneumoconiosis, 18 pneumonia, 14,15 emphysema, 7 and pleural effusion.…”

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

“…The development of AI systems for object detection and recognition in X-rays is an emerging field, and many such studies have used transfer learning. Unlike disease diagnosis, the use of AI for object detection focuses mainly on the therapeutic tube and catheter [20][21][22]. In a similar study on the placement of feeding tubes, pre-trained deep CNN models (Inception V3, ResNet50, and DenseNet121) achieved AUC values of 0.82-0.87, which is far lower than that of the models developed for disease diagnosis [22].…”

A review of transfer learning for medical image classification

“…However, there is no consensus among studies concerning the global optimum configuration for fine-tuning. [24] concluded that fine-tuning the last fully connected layers of Inception3, ResNet50, and DenseNet121 outperformed fine-tuning from scratch in all cases with AUC values ranging from 0.82 to 0.85. On the other hand, Yu et al [25] found that retraining from scratch of DenseNet201 achieved the highest diagnostic accuracy of 92.73%.…”