Per-Pixel Lung Thickness and Lung Capacity Estimation on Chest X-Rays using Convolutional Neural Networks
Authors:
Manuel Schultheiss,
Philipp Schmette,
Thorsten Sellerer,
Rafael Schick,
Kirsten Taphorn,
Korbinian Mechlem,
Lorenz Birnbacher,
Bernhard Renger,
Marcus R. Makowski,
Franz Pfeiffer,
Daniela Pfeiffer
Abstract:
Estimating the lung depth on x-ray images could provide both an accurate opportunistic lung volume estimation during clinical routine and improve image contrast in modern structural chest imaging techniques like x-ray dark-field imaging. We present a method based on a convolutional neural network that allows a per-pixel lung thickness estimation and subsequent total lung capacity estimation. The n…
▽ More
Estimating the lung depth on x-ray images could provide both an accurate opportunistic lung volume estimation during clinical routine and improve image contrast in modern structural chest imaging techniques like x-ray dark-field imaging. We present a method based on a convolutional neural network that allows a per-pixel lung thickness estimation and subsequent total lung capacity estimation. The network was trained and validated using 5250 simulated radiographs generated from 525 real CT scans. The network was evaluated on a test set of 131 synthetic radiographs and a retrospective evaluation was performed on another test set of 45 standard clinical radiographs. The standard clinical radiographs were obtained from 45 patients, who got a CT examination between July 1, 2021 and September 1, 2021 and a chest x-ray 6 month before or after the CT. For 45 standard clinical radiographs, the mean-absolute error between the estimated lung volume and groundtruth volume was 0.75 liter with a positive correlation (r = 0.78). When accounting for the patient diameter, the error decreases to 0.69 liter with a positive correlation (r = 0.83). Additionally, we predicted the lung thicknesses on the synthetic test set, where the mean-absolute error between the total volumes was 0.19 liter with a positive correlation (r = 0.99). The results show, that creation of lung thickness maps and estimation of approximate total lung volume is possible from standard clinical radiographs.
△ Less
Submitted 29 March, 2022; v1 submitted 24 October, 2021;
originally announced October 2021.
