Please use this identifier to cite or link to this item: https://ahro.austin.org.au/austinjspui/handle/1/34184
Title: Assessment of inspiration and technical quality in anteroposterior thoracic radiographs using machine learning.
Austin Authors: Sorace, L;Raju, N;O'Shaughnessy, J;Kachel, S;Jansz, K;Yang, N;Lim, R P
Affiliation: Radiology
The University of Melbourne, Parkville, Australia; Columbia University, New York, NY, USA.
The University of Melbourne, Parkville, Australia.
Issue Date: Jan-2024
Date: 2023
Publication information: Radiography (London, England : 1995) 2024-01; 30(1)
Abstract: Chest radiographs are the most performed radiographic procedure, but suboptimal technical factors can impact clinical interpretation. A deep learning model was developed to assess technical and inspiratory adequacy of anteroposterior chest radiographs. Adult anteroposterior chest radiographs (n = 2375) were assessed for technical adequacy, and if otherwise technically adequate, for adequacy of inspiration. Images were labelled by an experienced radiologist with one of three ground truth labels: inadequate technique (n = 605, 25.5 %), adequate inspiration (n = 900, 37.9 %), and inadequate inspiration (n = 870, 36.6 %). A convolutional neural network was then iteratively trained to predict these labels and evaluated using recall, precision, F1 and micro-F1, and Gradient-weighted Class Activation Mapping analysis on a hold-out test set. Impact of kyphosis on model accuracy was assessed. The model performed best for radiographs with adequate technique, and worst for images with inadequate technique. Recall was highest (89 %) for radiographs with both adequate technique and inspiration, with recall of 81 % for images with adequate technique and inadequate inspiration, and 60 % for images with inadequate technique, although precision was highest (85 %) for this category. Per-class F1 was 80 %, 81 % and 70 % for adequate inspiration, inadequate inspiration, and inadequate technique respectively. Weighted F1 and Micro F1 scores were 78 %. Presence or absence of kyphosis had no significant impact on model accuracy in images with adequate technique. This study explores the promising performance of a machine learning algorithm for assessment of inspiratory adequacy and overall technical adequacy for anteroposterior chest radiograph acquisition. With further refinement, machine learning can contribute to education and quality improvement in radiology departments.
URI: https://ahro.austin.org.au/austinjspui/handle/1/34184
DOI: 10.1016/j.radi.2023.10.014
ORCID: 
Journal: Radiography (London, England : 1995)
Start page: 107
End page: 115
PubMed URL: 37918335
ISSN: 1532-2831
Type: Journal Article
Subjects: Artificial intelligence
Convolutional neural network
Machine learning
Quality
Thoracic radiology
Appears in Collections:Journal articles

Show full item record

Page view(s)

30
checked on Oct 14, 2024

Google ScholarTM

Check


Items in AHRO are protected by copyright, with all rights reserved, unless otherwise indicated.