Spin-locked MOLED with deep learning enables fast, robust T1ρ, T2, and T2* mapping

A spin-locked MOLED technique with deep learning reconstruction maps T1ρ, T2, and T2* in ~11 sec/slice, with phantom correlation r>0.997 and in vivo repeatability CV<3.3%. In an MS patient, T1ρ was elevated in lesions relative to normal white matter.

• Numerical experiments: artifact-aware training boosted SSIM by 0.1–0.2 and reduced MAE by 2–5 ms for T1ρ, T2, T2*.
• In a patient with multiple sclerosis, T1ρ elevation in lesions was more prominent than T2/T2* changes, suggesting different pathological sensitivity.

Automated summary

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