Recommendations for quantitative cerebral perfusion MRI using multi-timepoint arterial spin labeling: Acquisition, quantification, and clinical applications

Abstract

Accurate assessment of cerebral perfusion is vital for understanding the hemodynamic processes involved in various neurological disorders and guiding clinical decision-making. This guidelines article provides a comprehensive overview of quantitative perfusion imaging of the brain using multi-timepoint arterial spin labeling (ASL), along with recommendations for its acquisition and quantification. A major benefit of acquiring ASL data with multiple label durations and/or post-labeling delays (PLDs) is being able to account for the effect of variable arterial transit time (ATT) on quantitative perfusion values and additionally visualize the spatial pattern of ATT itself, providing valuable clinical insights. Although multi-timepoint data can be acquired in the same scan time as single-PLD data with comparable perfusion measurement precision, its acquisition and postprocessing presents challenges beyond single-PLD ASL, impeding widespread adoption. Building upon the 2015 ASL consensus article, this work highlights the protocol distinctions specific to multi-timepoint ASL and provides robust recommendations for acquiring high-quality data. Additionally, we propose an extended quantification model based on the 2015 consensus model and discuss relevant postprocessing options to enhance the analysis of multi-timepoint ASL data. Furthermore, we review the potential clinical applications where multi-timepoint ASL is expected to offer significant benefits. This article is part of a series published by the International Society for Magnetic Resonance in Medicine (ISMRM) Perfusion Study Group, aiming to guide and inspire the advancement and utilization of ASL beyond the scope of the 2015 consensus article.

Keywords: arterial spin labeling; arterial transit time; cerebral blood flow; multi‐timepoint; perfusion.

Figure 1:

Demonstrating the relative insensitivity to ATT of multi-timepoint PCASL data. The single-PLD data shows CBF underestimation in regions where the PLD is shorter than the ATT. When a sufficiently long PLD is used, the CBF estimated from single-PLD scans is relatively insensitive to ATT. However, by jointly fitting data at multiple PLDs to a kinetic signal model, ATT-insensitive CBF and ATT can be directly measured.

Figure 2:

Demonstrating changes in ATT with disease. The top row shows ATT maps at various slices for a patient with carotid artery occlusion, while the bottom row shows group average ATT maps from 9 healthy young subjects. The data was corrected for partial volume effects and represents estimates of pure GM ATT. Note the left-right symmetry and overall lower ATT in the healthy young subjects compared to the patient. ATTs in the ipsilateral hemisphere were on average ~1 s longer for the patient. Figure reproduced from ref with permission.

Figure 3:

Illustrating the timing principles of the recommended multi-timepoint PCASL protocol types. The label durations are depicted as orange and grey boxes, representing label and control conditions, respectively. The post-labeling delay is the time between the end of the label duration and the start of the image readout. Adapted from ref with permission.

Figure 4:

Demonstrating the time-encoded timing decoding process for a 4×3 Hadamard encoding pattern. Four encoded images (A,B,C,D, shown in green) are acquired from which three perfusion weighted images can be decoded, each with a different label duration and effective post-labeling delay. The label durations are depicted as orange and grey boxes, representing label and control conditions, respectively. The process of decoding each of the three perfusion weighted images are shown top-right, bottom-left and bottom-right. The equivalent label durations and post-labeling delays of each decoded perfusion weighted image for a sequential multi-timepoint protocol are shown as black boxes.

Figure 5:

Monte Carlo simulation CBF and ATT quantification errors for the protocols shown in Table 2 and two single-PLD protocols. (A) The quantification errors for the short ATT range, 0.5 – 1.8 s (single-PLD: LD = 1800 ms, PLD = 1800 ms, N_Ave/N_Seq = 37, scan duration 5:03 minutes); (B) the quantification errors for the extended ATT range, 0.5 – 2.5 s (single-PLD: LD = 1800 ms, PLD = 2500 ms, N_Ave/N_Seq = 31, scan duration 4:58 minutes). The RMSE was used as a general measure of error because it is a combination of both measurement bias and SD. The data was simulated and fit using the PCASL signal model in Equation 3 with $\alpha =0.85,{\alpha }_{BS}=1$, M_0a = 1, CBF = 50 mL/100g/min, T_1b = 1.65 s. White Gaussian noise was added to 2000 copies of each simulated control/label/encoded “image” before subtraction/decoding. The noise SD was $M_{0a}\bullet 1.4\bullet {10}^{-3}$, equivalent to a noise SD of 38% of the ASL signal or an SNR of 2.63 for a single control-label difference image with LD = PLD = 1.8 s, T_1b = 1.65 s, and ATT < 1.8 s. Model fitting used a non-linear least squares approach similar to that described in ref.

Figure 6:

Brain-wide ATT changes before and after Acetazolamide administration. Note that, although both cases have similar watershed borders, the ATT is generally shorter after Acetazolamide (B) than before (A). Figure reproduced from ref with permission.

Figure 7:

An example time-encoded multi-timepoint PCASL acquisition in a low-grade diffuse glioma (oligodendroglioma, WHO grade 2). Top row: The extent of the tumor can be seen in the T₂-FLAIR (hyperintense region). The T₁+Gd shows no enhancement. The calculated PCASL CBF and ATT maps indicate an area of high perfusion with short arrival time within the tumor region, in part due to large intravascular signal contributions present in this data which were not modelled here. This is consistent with the finding that oligodendroglioma tend to be highly perfused and have high blood volume with irregular, though non-leaky, vasculature. Bottom row: perfusion-weighted images at each of the acquired PLDs.