Kidney Week

Abstract: TH-PO0027

MRI-Based Regional Kidney Blood Flow Mapping Is More Accurate Using Kidney-Specific Arterial Spin Labeling Model

Session Information

• Bioengineering: MPS, Flow, and Delivery
  November 06, 2025 | Location: Exhibit Hall, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Bioengineering

• 400 Bioengineering

Authors

• Wang, Jiachen, The University of Texas at Austin, Austin, Texas, United States

Jiachen Wang, PhD

• Bush, Adam, The University of Texas at Austin, Austin, Texas, United States

Adam Bush, PhD

• Margain, Corina, The University of Texas at Austin, Austin, Texas, United States

Corina Margain

Background

Arterial Spin Labeling (ASL) is a noncontrast MRI technique using magnetically labeled water to map renal blood flow (RBF). ASL is widely used in brain but to a lesser degree in the kidneys due to motion and inaccuracy. RBF quantification relies on kinetic models converting measured signals into perfusion (ml/min/100g). The single-compartment Buxton model is commonly used but lacks thorough kidney-specific validation. This study proposes an improved kinetic model incorporating an arterial outflow term for enhanced accuracy and physiological relevance.

Methods

Healthy subjects underwent MRI (Siemens 3T Vida) under IRB approval: labeling duration=1.5s; delays=0.5,0.75,1,1.25,1.5,2,2.5s; TR/TE=6000/16ms; 10 dynamics; single-shot 3D TGSE readout. Two studies were performed:

A: Baseline-11 subjects had renal ASL and phase-contrast (PC) MRI.
B: Stress test-5 subjects underwent renal ASL and PC MRI before and after an oral protein-load inducing RBF changes.

RBF was estimated using Buxton and our model:
dΔM/dt = f/λ[I(t)-OFFe^-OFT/T1I(t-OFT)]-1/T₁ΔM
where ΔM=ASL difference signal, f=RBF, λ=partition coefficient, I(t)=arterial input function, OFT/OFF=outflow delay/fraction, T1=longitudinal relaxation time.

Results

A: Fig1 illustrates arterial outflow contrasts (OFF, OFT) introduced by our model. Fig2A shows compared to Buxton, our model improves RBF estimation accuracy (397.87±68.42 vs. 167.59±63.07), better matching PC measurements (612.06±64.03). AIC favors our model.
B: Fig2B shows our model accurately detects RBF changes post-protein-load, correlating strongly with PC (r=0.62, p=0.27), whereas Buxton fails (r=-0.31, p=0.62).

Conclusion

Including arterial outflow in the ASL model enhances physiological relevance and improves RBF estimation.

Illustrative images from our model.

Statistical analyses.

Funding

• NIDDK Support

Digital Object Identifier (DOI)

• doi: 10.1681/ASN.20256ae485nw