Deep Learning Nearly Halves Scan Time While Improving Image Quality in 3T EDGE-MP2RAGE MRI

Deep learning reconstruction for 3T EDGE-MP2RAGE reduced scan time by 47.7%, increased contrast-to-noise ratio by 85.9% (P<.001) and decreased noise by 22.6% (P=.005) vs standard 3T, with image quality comparable to 7T but better contrast uniformity.

• In 117 patients (39 per group), DL reconstruction reduced 3T EDGE-MP2RAGE scan time by 47.7%, increased CNR by 85.9% (P<.001), and reduced noise by 22.6% (P=.005).
• Qualitative scores for overall image quality, sharpness, structural conspicuity, and contrast uniformity were significantly higher with DL (all P<.001) without added artifacts.
• 3T DL images had comparable CNR and noise to 7T (P=.17 and P=.77) and superior contrast uniformity (P<.001).

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.