VQ ‐Wave: A Physics‐Driven Spatiotemporal Deep Learning Approach for Noncontrast‐Enhanced Lung Ventilation and Perfusion MRI
Magnetic Resonance in Medicine · 2026-09 · 96 (3) · 1427-1442
DOI: 10.1002/mrm.70423 ↗Kurz gefasst
Fachartikel in der Zeitschrift Magnetic Resonance in Medicine, online erschienen am 10.05.2026. Verfasst von Grzegorz Bauman und 3 weiteren Autor:innen. Beteiligte Einrichtung: Division of Radiological Physics, Department of Radiology, University of Basel Hospital, Basel, Switzerland. Hauptthemen laut MEDLINE: Deep Learning, Lung, Image Processing, Computer-Assisted, Magnetic Resonance Imaging. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Purpose
To develop a robust deep learning framework for noncontrast‐enhanced functional lung MRI, overcoming the limitations of spectral decomposition in the presence of physiological nonstationarity.
Methods
We introduce VQ‐Wave (Ventilation/Q‐perfusion Waveform‐based Assessment of Variable Evolutions), a physics‐driven spatiotemporal inception neural network trained on synthetic signal models to estimate ventilation and perfusion parameters. By processing local spatial context alongside temporal evolution, the network learns to robustly decouple physiological signals from noise. The training generator simulated realistic nonstationary dynamics, including amplitude modulations, frequency drifts, and noise. Performance was validated against matrix pencil (MP) decomposition using numerical phantoms and in vivo functional lung MRI acquired in four healthy volunteers and two children with cystic fibrosis (CF) at 1.5 T. Robustness was assessed across varying noise levels, physiological instabilities, and scan durations (truncating the acquisition from N = 140 to N = 40 images).
Results
In numerical benchmarks, VQ‐Wave demonstrated superior robustness to nonstationarity, maintaining low global and regional error rates where MP exhibited stochastic instability due to spectral leakage. In vivo, VQ‐Wave accurately captured functional defects in patients with CF yielding diagnostically consistent ventilation and perfusion maps with high quantitative stability (global mean variation < 12%) even when scan time was reduced from 45 s to 15 s ( N = 40). Conversely, under irregular physiology and short scan lengths, MP decomposition severely degraded, exhibiting systematic amplitude instability, overestimation bias, and regional signal dropouts.
Conclusion
VQ‐Wave offers a robust, physics‐driven neural network‐based alternative to spectral decomposition. By effectively handling physiological irregularity and noise, it enables reliable functional lung imaging with substantially shortened acquisition protocols.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Division of Radiological Physics, Department of Radiology, University of Basel Hospital, Basel, Switzerland.
- Department of Biomedical Engineering, University of Basel, Allschwil, Switzerland.
- Pediatric Respiratory Medicine, Department of Pediatrics, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland.
ORCID-Profile: Pavlos Panos, Oliver Bieri
Zitieren
APA
Bauman, G., Panos, P., Latzin, P., & Bieri, O. (2026). VQ ‐Wave: A Physics‐Driven Spatiotemporal Deep Learning Approach for Noncontrast‐Enhanced Lung Ventilation and Perfusion MRI. Magnetic Resonance in Medicine, 96(3), 1427-1442. https://doi.org/10.1002/mrm.70423
Vancouver
Bauman G, Panos P, Latzin P, Bieri O. VQ ‐Wave: A Physics‐Driven Spatiotemporal Deep Learning Approach for Noncontrast‐Enhanced Lung Ventilation and Perfusion MRI. Magnetic Resonance in Medicine. 2026;96(3):1427-1442. doi:10.1002/mrm.70423
BibTeX
@article{Bauman2026,
author = {Bauman, Grzegorz and Panos, Pavlos and Latzin, Philipp and Bieri, Oliver},
title = {VQ ‐Wave: A Physics‐Driven Spatiotemporal Deep Learning Approach for Noncontrast‐Enhanced Lung Ventilation and Perfusion MRI},
journal = {Magnetic Resonance in Medicine},
year = {2026},
volume = {96},
number = {3},
pages = {1427--1442},
doi = {10.1002/mrm.70423}
}
EUVIMED