Feasibility of noninvasive quantitative measurements of intrarenal R2′ in humans using an asymmetric spin echo echo planar imaging sequence

Xiaodong Zhang, Yudong Zhang, Xuedong Yang, Xiaoying Wang, Hongyu An, Jue Zhang, Jing FangNMR 2012 · DOI

TL;DR

This paper introduces a rapid, non-invasive MRI technique (asymmetric spin echo EPI) to quantify intrarenal R2' in humans, a marker of tissue oxygenation. The method is shown to be reproducible and sensitive to changes in kidney oxygenation induced by furosemide, offering a promising tool for clinical assessment of renal health.

What Problem Does This Paper Address?

The kidney, particularly its medulla, is highly susceptible to injury due to its naturally low oxygen levels and delicate balance between oxygen supply and demand. Understanding and monitoring renal oxygenation is crucial for diagnosing and treating various kidney diseases. While invasive methods like microelectrodes exist for animal models, they are not suitable for human clinical use. Previous non-invasive MRI approaches, like BOLD (Blood Oxygenation Level-Dependent) contrast using gradient echo sequences to measure R2, have been used. However, R2 is a composite measure of both irreversible (R2) and reversible (R2') relaxation rates. R2' is more directly and linearly related to tissue oxygenation, making it a more specific indicator. The challenge has been to reliably and rapidly measure R2' non-invasively in the human kidney, which is prone to motion artifacts from breathing and cardiac activity.

Methods

The researchers employed an asymmetric spin echo (ASE) single-shot echo planar imaging (EPI) sequence. This sequence is designed to be fast, acquiring a full image volume in approximately 20 seconds, which can be performed within a single breath-hold. The core principle relies on the fact that the ASE signal decay is primarily governed by R2' when certain timing parameters are met. By analyzing the signal decay over different echo timings (ΔTE), R2' can be calculated using a linear fit.

A key aspect of their methodology was investigating the impact of intravascular signal contribution on R2' measurements. This was achieved by applying different flow dephasing gradients (b-values of 0, 40, and 80 s/mm²). These gradients are designed to suppress signals from moving blood, thereby isolating the tissue signal. The researchers hypothesized that using these gradients would improve the accuracy and reproducibility of R2' measurements by minimizing the influence of blood oxygenation.

Reproducibility was assessed by repeating scans on healthy subjects over three consecutive days. To evaluate the sensitivity to oxygenation changes, a loop diuretic, furosemide, was administered intravenously to a separate group of subjects. Furosemide is known to increase oxygen tension in the renal medulla. R2' measurements were taken before and after furosemide administration.

Datasets

Not discussed in the paper. The study involved human subjects.

Key Insights and Ideas

  • R2' as a direct oxygenation marker: The paper reinforces the idea that R2' is a more direct and sensitive indicator of tissue oxygenation compared to R2*, which is a combination of R2 and R2'.
  • ASE-EPI for rapid renal R2' measurement: The study successfully demonstrates that the ASE-EPI sequence can acquire intrarenal R2' data in humans within a clinically feasible timeframe (20 seconds), making it suitable for breath-hold acquisitions.
  • Flow dephasing gradients are crucial: The use of flow dephasing gradients (b=40 and 80 s/mm²) significantly improved the reproducibility of R2' measurements, particularly in the medullary region, by effectively suppressing intravascular signals. This suggests that these gradients are essential for accurate tissue R2' estimation.
  • Medulla is more hypoxic: Consistent with existing knowledge, the study found that baseline R2' values were higher in the renal medulla compared to the cortex, indicating a relatively hypoxic state in the medulla.
  • Furosemide's effect on medullary oxygenation: The administration of furosemide led to a significant decrease in medullary R2', confirming an increase in oxygen tension in this region, as predicted. This validates the method's ability to detect physiological changes in renal oxygenation.
  • Cortex R2' measurement challenges: The cortex showed higher inter-scan variability (CV) and less consistent responses to furosemide, suggesting potential challenges in accurately measuring cortical R2' with the current parameters, possibly due to partial volume effects or susceptibility artifacts.

Key Results

  • Reproducibility: The coefficient of variation (CV) for intrarenal R2' measurements was significantly reduced in the presence of flow dephasing gradients (9.68±3.58% with b=40 s/mm² and 10.50±3.62% with b=80 s/mm²) compared to the absence of gradients (21.31±4.52% with b=0 s/mm²).
  • Furosemide effect: A significant reduction in medullary R2' (p<0.05) was observed after furosemide administration when using flow dephasing gradients (b=40 and 80 s/mm²). No significant changes were observed in the cortex.
  • Gradient effectiveness: R2' measurements did not differ significantly between b=40 s/mm² and b=80 s/mm² scans, indicating that even moderate flow dephasing gradients are sufficient to minimize intravascular signal contributions.
  • Baseline differences: Significant differences in R2' between the cortex and medulla were observed across all days and gradient conditions, with the medulla consistently showing higher R2' values.

Limitations and Future Work

  • Motion artifacts: While the sequence is rapid, subjects who could not hold their breath for the full 20 seconds experienced motion artifacts, impacting image quality and SNR. This is particularly relevant for patients who may have difficulty with breath-holding.
  • Spatial resolution: The ASE-EPI sequence suffers from relatively low spatial resolution (2.5 × 3.75 × 6 mm³ voxel size), which may lead to diminished contrast between the medulla and cortex and increased measurement variation in the cortex due to partial volume effects.
  • Intravascular signal suppression: While flow dephasing gradients were effective, the paper notes that the method may underestimate R2' if the oxygen extraction fraction (OEF) is below 0.30, suggesting that longer ΔTE might be needed for very low OEF conditions.
  • Future directions: The authors suggest optimizing the imaging protocol to improve spatial resolution, reduce measurement variations, and potentially shorten acquisition time (e.g., to <15 seconds). They also propose exploring segmented ASE-EPI and parallel imaging techniques to enhance spatial resolution at the cost of acquisition time. Further refinement of parameters is needed for accurate cortical R2' studies.

Summary updated

Abstract

The purpose of this study was to demonstrate the feasibility of an asymmetric spin echo (ASE) single-shot echo planar imaging (EPI) sequence for the noninvasive quantitative measurement of intrarenal R(2) ' in humans within 20 s. The reproducibility of R(2) ' measurements with the ASE-EPI sequence was assessed in nine healthy young subjects in repeated studies conducted over three consecutive days. Moreover, we also evaluated whether the ASE-EPI sequence-measured R(2) ' reflected the intrarenal oxygenation changes induced by furosemide in another group of normal human subjects (n = 10). Different flow attenuation gradients (b = 0, 40 and 80 s/mm(2) ) were utilized to examine the impact of the intravascular signal contribution on the estimation of intrarenal R(2) '. In the absence of flow dephasing gradients (b = 0 s/mm(2) ), the computed coefficient of variation (CV) of R(2) ' was 21.31 ± 4.52%, and the estimated R(2) ' value decreased slightly, but not statistically significantly (p > 0.05), after the administration of furosemide in the medullary region. However, CV of R(2) ' was much smaller in the presence of flow dephasing gradients (9.68 ± 3.58% with b = 40 s/mm(2) and 10.50 ± 3.62% with b = 80 s/mm(2) ). Moreover, a significant reduction in R(2) ' in the renal medulla was obtained (p < 0.05 for both b = 40 s/mm(2) and b = 80 s/mm(2) ) after the administration of furosemide, reflecting an increase in oxygen tension in the medullary region. In addition, R(2) ' measurements did not differ between the b = 40 s/mm(2) and b = 80 s/mm(2) scans, suggesting that small diffusion gradients were sufficient to minimize the intravascular signal contribution. In summary, we have demonstrated that renal R(2) ' can be obtained rapidly using an ASE-EPI sequence. The measurement was highly reproducible and reflected the expected intrarenal oxygenation changes induced by furosemide.
BibTeX
@inproceedings{zhang2012feasibility,
  author = {Xiaodong Zhang and Yudong Zhang and Xuedong Yang and Xiaoying Wang and Hongyu An and Jue Zhang and Jing Fang},
  title = {{Feasibility of noninvasive quantitative measurements of intrarenal <i>R</i><sub>2</sub>′ in humans using an asymmetric spin echo echo planar imaging sequence}},
  booktitle = {NMR},
  year = {2012},
  doi = {10.1002/nbm.2823},
}
Figure