3803

Estimation of arterial blood T1 during hyperoxic gas-mixture breathing

Erin K Englund¹, Zachary B Rodgers¹, Hyunyeol Lee¹, and Felix W Wehrli¹

¹Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States

Synopsis

Hyperoxia gas mixture breathing causes additional oxygen to be dissolved in blood plasma thereby shortening the T₁ of arterial blood (T_1a). Here, we propose a method estimate T1a during hyperoxia gas breathing by solving a perfusion quantification model for T_1a given CBF measured during hyperoxia via phase contrast (PC) of the feeding arteries in the neck. The ratio between PC-based total CBF and ASL-based regional CBF was first determined at baseline, and then applied to normalize the expected ASL-based CBF during hyperoxia. T_1a was estimated as 1.46±0.09 s during hyperoxia gas mixture breathing, in reasonable agreement with previous reports.

Background & Motivation

Hyperoxia gas mixture breathing, defined as fractional oxygen in inspired air (FiO₂) > 21%, is an isometabolic stimulus that increases the oxygen carrying capacity of blood through additional oxygen dissolved in blood plasma. Oxygen itself is paramagnetic; therefore the increased presence of dissolved oxygen shortens the longitudinal relaxation time (T₁) of blood¹ which will impact the accuracy of perfusion quantification based on arterial spin labeling (ASL) because models rely on blood T₁. For example cerebral blood flow (CBF) can be calculated as²:

$$$CBF=\frac{\Delta M}{M_0}\frac{\lambda}{2\alpha}\frac{\frac{1}{T_{1a}}}{exp\left(\frac{-w}{T_{1a}}\right)-exp\left(\frac{-\left(w+\tau\right)}{T_{1a}}\right)}$$$ [1]

where ΔM is the tag-control difference, M₀ the equilibrium magnetization, λ the blood-tissue partition coefficient, α the labeling efficiency, w the duration of the post-labeling delay, τ the labeling duration, and T_1a the transverse relaxation time of arterial blood. Previous studies have investigated the dependence of T_1a on FiO₂ ³. However, use of the previously measured T_1a values requires precise knowledge of FiO₂, and thus precise control of the gas delivered to the subject, potentially impeding the use of hyperoxia as a stimulus. If T_1a could be estimated without prior knowledge of FiO₂, hyperoxia challenges could be performed more easily, without the need for expensive and uncomfortable equipment (e.g. hyperoxic gas could be delivered through nasal cannula). Here, we present a method to estimate T₁ of arterial blood based on simultaneous measurement of total CBF from phase contrast (PC) data acquired in feeding arteries (CBF_FLOW) and as the average of ASL-derived perfusion (CBF_PERF) in grey matter. Assuming CBF_FLOW and CBF_PERF changes due to hyperoxia are the same, measured CBF_FLOW during hyperoxia can be used to estimate CBF_PERF, and [Eq 1] can be solved for T_1a.

Methods

ASL and PC were measured using an interleaved background-suppressed 3D-GRASE-based pCASL⁴ and PC sequence at 3T. Data were acquired in five subjects throughout 5 minutes of baseline and 5 minutes of hyperoxia gas mixture breathing (FiO₂ = 100%). At baseline, CBF_PERF was computed in grey matter from the mean of perfusion weighted images acquired during the resting periods from [Eq 1] with τ=1.52 s, w=1.5 s, α=0.71 ⁴, and T_1a during normoxia of 1.65 s⁵. CBF_FLOW was also computed at baseline as the integral of velocity over the cross-sectional area of the four feeding arteries: right and left internal carotid and vertebral arteries divided by brain mass estimated from a T₁-weighted MPRAGE acquisition. The ratio between global CBF measured by CBF_FLOW and regional CBF measured by CBF_PERF was determined at baseline. That ratio was applied to CBF_FLOW measurements during hyperoxia to estimate the expected CBF_PERF and subsequently the T_1a that yielded the expected CBF_PERF was calculated (see Figure 1 for details).

Results

Figure 2 shows the average CBF_FLOW and CBF_PERF responses at baseline and hyperoxia. Here, CBF_PERF was calculated using T_1a=1.65 s for both baseline and hyperoxia states. In this subject, there is a slight reduction in CBF_FLOW during hyperoxia (this was not observed for all subjects) however the amount that CBF is reduced is overestimated from the ASL data relative to the PC data (CBF_FLOW = 5% reduction, CBF_PERF = 19% reduction). Averaged over the five subjects, T_1a was estimated as 1.46±0.09 s during 100% FiO₂. Figure 3 shows the corrected perfusion map based on the estimated T_1a. Table 1 shows results for measured CBF_PERF and CBF_FLOW at baseline and hyperoxia, as well as the estimated T_1a during hyperoxia.

Discussion & Conclusion

The approach used here is similar to that of Aslan, et al’s study⁶ to determine ASL labeling efficiency.

The estimated T₁ values are in reasonable agreement with those in a recent study by Siero, et al⁷ that estimated T₁ during hyperoxia based on flow changes during a hypercapnia stimulus and hypercapnic-hyperoxic stimulus, positing that the flow changes should be equivalent (driven by CO₂). Those authors found T_1a=1.49±0.9 s. Both Siero, et al’s results⁷ and those presented herein are higher than a previously cited value of 1.38 s³.

Although a slight reduction of blood flow is expected during hyperoxia³, this response was not observed in every subject, potentially due to changes in respiratory rate. However, regardless of the underlying physiologic response there will still be an increase in dissolved oxygen, shortening T_1a, and causing perfusion to be underestimated using T_1a of blood at baseline, as is seen in these results.

Ongoing studies are investigating the estimates of T_1a at various FiO₂ and will attempt to compare results to measured T_1a values. Estimation of T_1a during hyperoxia would be particularly useful in applications of hyperoxia-based calibrated fMRI, in which flow changes are typically assumed based on group measurements⁸.

Acknowledgements

The project described was supported by the National Center for Research Resources, Grant UL1RR024134, and is now at the National Center for Advancing Translational Sciences, Grant UL1TR000003. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. Also supported in part by the Institute for Translational Medicine and Therapeutics’ (ITMAT) Transdisciplinary Program in Translational Medicine and Therapeutics, and NIH grant T32 HL007954.

References

1. Tadamura E, Hatabu H, Li W, Prasad PV, Edelman RR. Effect of oxygen inhalation on relaxation times in various tissues. J Magn Reson Imaging. 1997;7:220–225.

2. Wang DJ, Zhang Y, Wolf RL, Roc AC, Alsop DC, Detre JA. Amplitude-modulated continuous arterial spin-labeling 3.0-T perfusion MR imaging with a single coil: Feasibility study. Radiology. 2005;235:218–228.

3. Bulte DP, Chiarelli PA, Wise RG, Jezzard P. Cerebral perfusion response to hyperoxia. JCBFM. 2006;27:69–75.

4. Fernández-Seara MA, Rodgers ZB, Englund EK, Wehrli FW. Calibrated bold fMRI with an optimized ASL-BOLD dual-acquisition sequence. NeuroImage. 2016;142:474–482.

5. Alsop DC, Detre JA, Golay X, Günther M, Hendrikse J, Hernandez-Garcia L, Lu H, MacIntosh BJ, Parkes LM, Smits M, van Osch MJP, Wang DJ, Wong EC, Zaharchuk G. Recommended implementation of arterial spin-labeled perfusion MRI for clinical applications: A consensus of the ISMRM perfusion study group and the European consortium for ASL in dementia. Magn Reson Med. 2014;73:102–116.

6. Aslan S, Xu F, Wang PL, Uh J, Yezhuvath US, van Osch M, Lu H. Estimation of labeling efficiency in pseudocontinuous arterial spin labeling. Magn Reson Med. 2010;63:765–771.

7. Siero JCW, Strother MK, Faraco CC, Hoogduin H, Hendrikse J, Donahue MJ. In vivo quantification of hyperoxic arterial blood water T1. NMR Biomed. 2015;28:1518–1525.

8. Chiarelli PA, Bulte DP, Wise R, Gallichan D, Jezzard P. A calibration method for quantitative BOLD fMRI based on hyperoxia. NeuroImage. 2007;37:808–820.

Figures

Figure 1. Steps for estimation of T₁ are as follows:

1. Quantify CBF_FLOW at baseline from PC data by integrating velocity in inflow vessels (internal carotid and vertebral arteries) over their cross-sectional area and dividing by brain mass (estimated from T₁-weighted MPRAGE).

2. Quantify CBF_PERF at baseline from ASL data assuming T_1a = 1.65 s.

3. Determine ratio between CBF_FLOW and CBF_PERF at baseline.

4. Measure CBF_FLOW during hyperoxia (HO), and apply ratio from step 3 to estimate CBF_PERF during hyperoxia.

5. Calculate average ΔM from ASL data during hyperoxia.

6. Solve perfusion quantification equation for T_1a during hyperoxia.

Figure 2. Data from a representative subject during baseline (left column) and hyperoxia (right column). Based on CBF_FLOW, there is a slight reduction in total CBF, however the observed response from CBF_PERF is much greater due to the shortening of T_1a.

Figure 3. Perfusion maps at baseline (A), during hyperoxia assuming T_1a = 1.65 s (B), and during hyperoxia using the calculated T_1a (T_1a=1.47 s in this subject) (C). Prior to correction for the shortening of T_1a, there is a larger than expected perfusion decrease observed during hyperoxia.

Table 1. Baseline CBF_FLOW and CBF_PERF, along with hyperoxia CBF_FLOW and estimated T_1a for each subject. Note that the ratio between CBF_FLOW and CBF_PERF is variable for each subject at baseline. Additionally, subjects exhibited different physiologic responses to the hyperoxia stimulus, ranging from a decrease (as expected with hyperoxia), to no change or an increase in CBF. Regardless of the physiologic response, the T_1a was estimated in a relatively tight range of plausible values.

Proc. Intl. Soc. Mag. Reson. Med. 25 (2017)

3803