Skip to main content
SCMR 29th Annual Scientific Sessions Main banner
Final Program


Program-At-A-Glance

Rapid Fire Session

Session: Rapid Fire Session 4: New Contrast Techniques (Monitor 14)

Dual-adiabatic spin-lock measurements for single-scan dispersion mapping in the human heart at 3T

Friday, February 6, 2026
3:52 PM - 3:59 PM  
Location: Galapagos / Aruba / Ilhabela (2nd Lower Level)
Background: T1ρ mapping has emerged as a promising non-contrast biomarker for various cardiomyopathies[1-3]. Measurement during the application of a spin-lock pulse renders T1ρ sensitive to slow molecular interactions (~100Hz-1kHz). However, T1ρ measurements are not specific to a single relaxation mechanism as multiple effects, including chemical exchange and molecular diffusion, fall in this frequency range. R1ρ (=1/T1ρ) dispersion mapping aims to isolate relaxation mechanisms. By probing R1ρ at different spin-lock frequencies[4], dispersion mapping suppresses contributions outside of a narrow frequency window. Both T1ρ and R1ρ dispersion mapping typically rely on continuous wave (CW) spin-locking. However, in cardiac imaging, CW is often limited by its susceptibility to field inhomogeneities[5]. Adiabatic spin-lock preparations have been shown to alleviate this[6]. Thus, in this study, we investigate the use of a dual adiabatic spin-lock sequence to derive robust R1ρ dispersion maps in the heart at 3T.

Methods: Dispersion maps were acquired in a single scan with ECG-triggered, single-shot bSSFP (TE/TR=1.5/3.1ms,res=2x2x8mm3). Pairs of phase-cycled adiabatic hyperbolic secant (HS)-pulses were used for spin-locking[6,7] (Fig.1B). Four images were acquired in a 14s breath-hold: one without preparation, two HS-prepared images with different pulse parameters, and a saturation-prepared image (Fig 1A). Pairs of R1ρ,adiab maps were generated for each scan using three-parameter fitting[9]. Dispersion maps were computed by pixel-wise subtraction (Fig.1C).

Imaging was performed at 3T (Ingenia, Philips). The effect of the HS-pulse parameters on R1ρ was studied in the T1MES phantom[8], for fmax=100,...,500 Hz, and β=2.5,...,8.5, with 10 repetitions per scan.

R1ρ,adiab dispersion maps were acquired in 3 SAX slices of 3 healthy volunteers (26±2y, 2m, 1f). The mid-slice was acquired in 3 repetitions to assess repeatability in manually drawn septal ROIs.

Results: Phantom data shows good R1ρ,adiab map quality, and a consistent decrease in R1ρ values for increasing fmax and β (Fig.2). R1ρ,adiab repeatability significantly deteriorates for fmax ≤100 Hz, indicating insufficient pulse efficiency.

In vivo maps were acquired with fmax/β=3.5/200Hz and 8.5/450Hz for the two pulses per scan, respectively. Adiabatic R1ρ maps show high map quality with a homogenous myocardial signal for both pulse configurations (Fig.3A). R1ρ,adiab dispersion maps show good contrast, with visually largely homogenous myocardial signal, albeit at increased noise levels compared to individual R1ρ,adiab maps. Good repeatability was observed in the septal measurements with 10.8±7.3% coefficient of variation and 12.5% inter-subject reproducibility.

Conclusion: Adiabatic spin-lock preparations can provide robust R1ρ dispersion mapping of the myocardium in a single breath-hold scan. The resulting maps hold promise for increased contrast specificity in quantitative tissue characterization compared with conventional T1ρ mapping techniques.

a) Dual-adiabatic spin-lock sequence, for acquisition of adiabatic dispersion maps in a single scan. Four bSSFP images are acquired in a single breath-hold, separated by 4s delay for magnetization recovery. Two different adiabatic preparations were obtained with different HS-values. b) Dual HS pulses waveforms for amplitude (top) and frequency modulation (bottom). Each pulse is composed of two phase-cycled adiabatic full passages of 30ms length, for a total of tHS = 60 ms. β and fmax parameters permit modulation of the pulses. c) Processing pipeline. Two R1ρ,adiab maps were calculated analytically from the two HS-prepared images, using a three-parameter model and the saturation and the unprepared image. R1ρ,adiab dispersion maps are computed from the generated maps by pixel-wise subtraction.
Effect of HS pulse parameters on R1ρ,adiab in T1MES phantom. a) An example of an R1ρ,adiab map, acquired with (β= 5.5, fmax = 200Hz) shows robust map quality. b) Heatmap of the parameter space indicates lower R1ρ,adiab for increasing β and fmax. c) and d) show the effect of changing β and fmax respectively, for three representative vials (see a). Higher R1ρ,adiab, and higher variability is observed for lower β and fmax.
n vivo R1ρ,adiab and ΔR1ρ,adiab dispersion maps acquired in three short axis slices from a representative subject. R1ρ,adiab maps show high map quality, with homogenous myocardial signal in visual assessment. ΔR1ρ,adiab dispersion maps provide good blood-myocardium contrast, and visually largely uniform myocardial values, albeit at increased noise levels compared to R1ρ,adiab. b) Septal ΔR1ρ,adiab values show low coefficients of variation (10.8±7.5%) and good inter-subject reproducibility (12.5%) across three healthy volunteers.