🧠 AI Boosts 129Xe MRI: Sharper Images, Cost Savings

A recent study evaluated whether deep learning (DL)-accelerated acquisition and reconstruction preserves key lung imaging metrics in 129Xe diffusion-weighted MRI, finding improved image quality and comparable quantitative values to conventional methods.

Why It Matters To Your Practice

▪ DL methods can improve sharpness and signal-to-noise ratio (SNR) in lung MRI imaging, critical for diagnosing respiratory diseases.

▪ Enables use of natural-abundance xenon, potentially reducing costs and logistical barriers for advanced lung morphometry.

▪ Higher acceleration factors may speed up scan times, benefiting busy clinical workflows.

▪ Supports broader adoption of 129Xe MRI in real-world settings, including for COPD, asthma, and IPF evaluation.

Clinical Implications

▪ DL reconstructions introduce only a slight bias in diffusion metrics (ADC and LmD), supporting their clinical reliability.

▪ Improved SNR with natural-abundance xenon could make advanced imaging more accessible without specialized gas supplies.

▪ Potential for more accurate assessment of lung microstructure in diverse pulmonary diseases.

▪ May enable new protocols for faster, higher-quality lung imaging in routine practice.

Insights

▪ DL-based compressed sensing, denoising, and de-ringing outperformed conventional CS in both retrospective and prospective cohorts.

▪ Metrics for healthy volunteers using natural-abundance xenon aligned closely with those from enriched xenon imaging.

▪ Minor bias in ADC and LmD should be considered when interpreting results across reconstruction methods.

▪ AI-driven workflows may spur innovation in hyperpolarized gas imaging and other advanced MRI applications.

The Bottom Line

▪ DL-accelerated 129Xe MRI enhances image quality and feasibility, with only minimal quantitative bias, paving the way for broader, cost-effective clinical use in lung disease assessment.