Back to the list
Congress: ECR24
Poster Number: C-19297
Type: EPOS Radiologist (scientific)
Authorblock: K. M. Moon1, D. Hwang2, H-S. Choi3, J. Y. Hong4, H. Ko4, Y. Kim2, W. J. Kim4, D. Lee2; 1Gangneung/KR, 2Chuncheon 24341/KR, 3Seoul/KR, 4Chuncheon/KR
Disclosures:
Kyoung Min Moon: Nothing to disclose
Donghwan Hwang: Nothing to disclose
Hyun-Soo Choi: Nothing to disclose
Ji Young Hong: Nothing to disclose
Hongseok Ko: Nothing to disclose
Yoon Kim: Nothing to disclose
Woo Jin Kim: Nothing to disclose
Doohee Lee: Nothing to disclose
Keywords: Artificial Intelligence, Emergency, Gastrointestinal tract, Digital radiography, Neural networks, PACS, Catheters, Complications, Dynamic swallowing studies, Education and training, Image verification, Swallowing disorders
Purpose The accurate placement of a nasogastric tube is a critical procedure in medical practice, where precision is paramount. Incorrect positioning of the tube can lead to severe complications, including pulmonary aspiration, tissue perforation, and inadequate nutritional delivery. Recognizing this, our study introduces a groundbreaking approach to automate the detection of nasogastric tube positions in chest X-ray images. This method leverages advanced deep learning techniques, with a particular focus on the application of active learning methodologies. The primary objective of this...
Read more Methods and materials In pursuit of our objective, we conducted a comprehensive retrospective analysis. Our dataset comprised 2,262 chest radiographs, which included images featuring nasogastric tubes. These radiographs were sourced from two prominent medical institutions: Gangneung Asan Hospital and Chuncheon Sacred Heart Hospital. To ensure the accuracy and reliability of our dataset, each of the 2,262 chest radiographs was meticulously annotated by two experienced pulmonologists.The cornerstone of our methodology was the division of 2,062 of these chest X-rays into eight distinct queries, as...
Read more Results Our findings were noteworthy, particularly in the realm of model performance and efficiency. The active learning methodology, initiated by query 2 in our training set, significantly enhanced the performance of our segmentation model. Notably, this improvement was not just in accuracy but also in the reduction of preprocessing time required for model creation, signifying a leap in efficiency.In the testing set, our model demonstrated exceptional performance with an Area Under the Receiver Operating Characteristic (AUROC) of 0.99 for classification using...
Read more Conclusion The outcomes of our study firmly establish the efficacy of employing a deep learning model, augmented with active learning techniques, for the detection of nasogastric tubes in chest X-rays. This advancement not only underscores the potential of artificial intelligence in medical diagnostics but also marks a significant step forward in enhancing patient safety. The development and implementation of a classifier for nasogastric tube position detection promise to be a valuable tool in the medical field, offering a more reliable, efficient,...
Read more References 1An Artificial Neural Network for Nasogastric Tube Position Decision Support.Drozdov I, Dixon R, Szubert B, Dunn J, Green D, Hall N, Shirandami A, Rosas S, Grech R, Puttagunta S, Hall M, Lowe DJ.Radiol Artif Intell. 2023 Feb 1;5(2):e220165. doi: 10.1148/ryai.220165. eCollection 2023 Mar.PMID: 37035435 Free PMC article. 2Automatic detection of supporting device positioning in intensive care unit radiography.Sheng C, Li L, Pei W.Int J Med Robot. 2009 Sep;5(3):332-40. doi: 10.1002/rcs.265.PMID: 19449314 Clinical Trial. 3Review of robot-assisted laparoscopic surgery in management of infant...
Read more
GALLERY