Click here to

Session: MRI [Return to Session]

Novel Sampling Strategy for Improved Reproducibility in Quantitative T1rho Mapping

C Wu1*, Q Peng2, (1) Department of Medical Physics, Memorial Sloan Kettering Cancer Center, New York, NY, USA (2) Department of Radiology, Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, NY, USA


TH-C-TRACK 3-5 (Thursday, 7/29/2021) 1:00 PM - 2:00 PM [Eastern Time (GMT-4)]

Purpose: T1rho (T1ρ) mapping emerges as a promising MRI tool for quantitative evaluation of low-frequency biochemical changes. However, the accuracy and reproducibility of T1ρ quantification can be compromised by B0 and B1 field inhomogeneities. The purpose of this study was to develop a novel sampling strategy of spin-locking time (TSL) for improved reproducibility in quantitative T1ρ mapping.

Methods: Two TSL sampling strategies were compared in simulation and phantom experiments. The first one was the standard sampling strategy where TSLs were sampled in a linearly spaced fashion (TSL-LS). The second one was a novel TSL sampling strategy called precision- and reproducibility-guided TSL sampling (TSL-PRG) where all TSLs, except TSL=0 ms, were sampled near the expected T1ρ value of interest at a constant interval defined by the Mz oscillation frequency. Simulation was repeated 100 times with field offsets of B0 (±600 Hz) an B1 (±50%) in a 50×50 grid where the percentage of T1ρ errors < 5% (T1ρ_Error_5pt) and the coefficients of variation < 5% (T1ρ_CV_5pt) were compared. For phantom experiment, T1ρ standard deviations in the three tubes at 150 Hz B0 off-resonance were compared.

Results: Simulation showed that TSL-PRG had slightly better T1ρ quantification accuracy within the range of B0 and B1 offsets (T1ρ_Error_5pt: 30.4% versus 32.8% for TSL-LS and TSL-PRG, respectively). Moreover, T1ρ variation was substantially reduced for TSL-PRG (T1ρ_CV_5pt: 44.3% versus 67.0% for TSL-LS and TSL-PRG, respectively). In the phantom experiment, standard deviation of T1ρ values at about 150 Hz B0 off-resonance was significantly reduced for TSL-PRG compared to TSL-LS in all three tubes (average of 5.40±0.75 ms versus 8.70±2.25 ms, P<0.001).

Conclusion: This study demonstrated that the proposed TSL-PRG sampling strategy produced more reproducible T1ρ quantification compared to standard linearly spaced TSL sampling (TSL-LS), suggesting the importance of TSL selection in quantitative T1ρ mapping.



    MRI, Quantitative Imaging, Magnetic Fields


    IM- MRI : Quantitative Imaging

    Contact Email