A non-invasive magnetic resonance imaging approach for assessment of real-time microcirculation dynamics

Abstract

We present a novel, non-invasive magnetic resonance imaging (MRI) technique to assess real-time dynamic vasomodulation of the microvascular bed. Unlike existing perfusion imaging techniques, our method is sensitive only to blood volume and not flow velocity. Using graded gas challenges and a long-life, blood-pool T ₁-reducing agent gadofosveset, we can sensitively assess microvascular volume response in the liver, kidney cortex, and paraspinal muscle to vasoactive stimuli (i.e. hypercapnia, hypoxia, and hypercapnic hypoxia). Healthy adult rats were imaged on a 3 Tesla scanner and cycled through 10-minute gas intervals to elicit vasoconstriction followed by vasodilatation. Quantitative T ₁ relaxation time mapping was performed dynamically; heart rate and blood oxygen saturation were continuously monitored. Laser Doppler perfusion measurements confirmed MRI findings: dynamic changes in T ₁ corresponded with perfusion changes to graded gas challenges. Our new technique uncovered differential microvascular response to gas stimuli in different organs: for example, mild hypercapnia vasodilates the kidney cortex but constricts muscle vasculature. Finally, we present a gas challenge protocol that produces a consistent vasoactive response and can be used to assess vasomodulatory capacity. Our imaging approach to monitor real-time vasomodulation may be extended to other imaging modalities and is valuable for investigating diseases where microvascular health is compromised.

Figure 1

Stability of the blood-pool agent gadofosveset in the microvascular bed. Temporal evolution of mean T ₁ relaxation times and standard error in the aorta, liver, kidney, and paraspinal muscle post-administration of gadofosveset in the absence of gas challenge stimuli (N = 2).

Figure 2

Dynamic MR imaging of hypercapnia- and hypercapnic hypoxia-mediated vasomotion. The kidney is shown on T ₁-weighted images acquired at the end of successive 8–12 minute gas challenges in a rat subjected to the following sequence: normoxia (A) → hypercapnia (20% CO₂) (B) → hypercapnic hypoxia (20% CO₂ + 12% O₂) (C) → hypercapnia (5% CO₂) (D). Corresponding temporal evolution of changes in T ₁ relaxation times in the liver, kidney, and paraspinal muscle (E).

Figure 3

Dynamic MR imaging of differential microvascular response to hypercapnia and hypoxia. The kidney is shown on T ₁-weighted images acquired at the end of successive 8–12 minute gas challenges in a rat subjected to the following sequence: normoxia (A) → hypoxia (12% O₂) (B) → hypercapnia (20% CO₂) (C) → hypoxic hypercapnia (12% O₂ + 20% CO₂) (D). Corresponding temporal evolution of changes in T ₁ relaxation times in the liver, kidney, and paraspinal muscle (E).

Figure 4

Dynamic MR imaging of blunting effect of mild hypercapnia-mediated vasodilatation on subsequent vasoconstriction capacity. The kidney is shown on T ₁-weighted images acquired at the end of successive 8–12 minute gas challenges in a rat subjected to the following sequence: normoxia (A) → mild hypercapnia (5% CO₂) (B) → severe hypercapnia (20% CO₂) (C) → hypoxic hypercapnia (12% O₂ + 20% CO₂) (D). Corresponding temporal evolution of changes in T ₁ relaxation times in the liver, kidney, and paraspinal muscle (E).

Figure 5

MRI response to gas challenges. Mean changes in T ₁ relaxation times for different gas challenge regimes averaged across all animals (N = 4 per group) and standard error are shown for the liver, kidney cortex, and paraspinal muscle. Exception is 2% CO₂, which was studied in a single animal. Significant differences from baseline challenges are indicated (*P < 0.05).

Figure 6

Heart rate and blood oxygen saturation response to gas challenges. Mean changes in pulse oximeter measured blood oxygen saturation and heart rate for different gas challenge regimes averaged across all animals (N = 4 per group) and standard error are shown for the liver, kidney cortex, and paraspinal muscle. Exception is 2% CO₂, which was studied in a single animal. Significant differences from baseline challenges are indicated (*P < 0.05).

Figure 7

Laser Doppler perfusion response to gas challenges. Mean changes in blood perfusion units in the liver and kidney cortex for different gas challenge regimes averaged across all animals (N = 5 per group) and standard error are shown for the liver and kidney cortex. Exceptions are 5% CO₂ and 12% O₂, which were studied in a single animal. Significant differences from baseline challenges are indicated (*P < 0.05).

Figure 8

Real-time tissue perfusion response in the liver and kidney cortex. Dynamic laser Doppler perfusion recordings are shown for the following gas sequence: normoxia→ hypercapnia (20% CO₂) → hypercapnic hypoxia (20% CO₂ + 12% O₂) → hypercapnia (5% CO₂). Units are in relative blood perfusion units (BPU).