T2 measurements during hyperoxia in the lateral ventricles and peripheral CSF
Jolanda M Spijkerman1, Jill B de Vis1, Esben T Petersen2, Jeroen Hendrikse1, and Jaco J M Zwanenburg1

1Radiology, University Medical Center Utrecht, Utrecht, Netherlands, 2Centre for Functional and Diagnostic Imaging and Research, Danish Research Centre for Magnetic Resonance, Copenhagen University Hospital, Hvidovre, Denmark


The oxygen tension in CSF and tissue is coupled, but may be easier to measure in CSF. This work investigated the effect of O2 administration on the T2 of CSF. In five subjects, T2s were mapped in the lateral ventricles and the peripheral CSF during administration of a hyperoxic gas mixture (end-tidal O2 level: 500mmHg). In the lateral ventricles no T2 change was observed, in the periphery a significant T2 reduction was found under O2 administration (T2=1.54s) compared to baseline (T2=1.61s), showing the feasibility to use CSF relaxation parameter mapping to detect oxygen tension changes.


When a subject is exposed to hyperoxic breathing, the O2 concentration in blood, and potentially also in tissue, increases [1, 2]. The oxygen tension in CSF and tissue is coupled, but may be easier to measure in CSF due to the absence of e.g. exchange between extracellular and intracellular fluid [2, 3]. In Zaharchuk et al. [4] it is shown that administration of a hyperoxic gas mixture to a subject results in a T1 change in the third ventricle, the cortical sulci and the basilar cisterns. In this research it was investigated if this effect can also be measured using T2 mapping. If so, future work could investigate if T2 changes of CSF precede brain atrophy, since it is known that brain perfusion, and thus brain oxygenation, decreases in these patients [5]. In this case, T2 mapping of CSF could serve as an early marker for Alzheimer’s disease.


For this study five volunteers were included (24-34 years, 3 male), MR imaging was performed on a 3T Philips system, using an 8ch receive head coil. For T2 mapping of CSF an MLEV-based pulse sequence, described by Qin [6], was used. Effective echo times of 400, 800, 1600, 3200ms were used in all experiments for quantitative T2 mapping. Five repetitions of T2 mapping were performed: the first repetition was performed at baseline breathing after which a hyperoxic gas mixture was administered to the subject using a RespirActTM system (Thornhill Research Inc, Toronto, Canada), the target end-tidal O2 level was set at 500mmHg [7] (Figure 1). The acquisition duration of one T2 map was 1:22min, meaning that at the start of the final repetition, diffusion of O2 towards CSF had occurred during 4:06min.
In all scans a single slice was acquired with 3x3x6mm3 resolution, SENSE 2.3, TR 15s, FOV 240x240x6mm3 and a single shot 2D SE-EPI readout (TE = 53ms). Slice planning is shown in Figure 2A.
Two regions of interest (ROIs) were selected: one for the lateral ventricles and one for the peripheral CSF (Figure 2). For all scans the mean signal in each of the two ROIs was determined, and T2s were fitted over this mean signal, using a plain single exponential decay model. The image analysis was performed using Matlab (version 2015A, Mathworks). Statistical paired t-tests were used to compare the T2 values (p<0.05 was regarded significant).


In Table 1 the mean fitted T2s of the 5 volunteers for the different MLEV repetitions are shown. For the lateral ventricles, no significant differences in T2 were found between the baseline and the administration of hyperoxic gas. In peripheral CSF, the T2 values in the last three repetitions (during hyperoxic gas administration) were significantly different from the baseline and the second MLEV repetition (at the start of hyperoxic gas administration). Overall lower T2 values were measured in the periphery, compared to the lateral ventricles.


In the lateral ventricles no significant T2 changes were found during the administration of hyperoxic gas, but in peripheral CSF a significant T2 decay was found. These results are compliant with the work by Zaharchuk et al. [4]. This indicates that the oxygen tension of CSF in the lateral ventricles may be relatively independent of the blood oxygen tension, and that there might have been more O2 diffusion from blood and/or tissue towards peripheral CSF, compared to CSF in the lateral ventricles.
In this research no data was available of the third and fourth ventricle. Therefore, additional work is necessary to study possible T2 changes in these regions as well. Further work is also needed to obtain the relaxivity of O2 in CSF in order to be able to translate T2 changes to changes in oxygen tension.
The overall lower T2 values of peripheral CSF compared to that of the ventricles corresponds to previous work [6, 8], and is probably due to partial volume effects with brain tissue and its interstitial fluid (data not shown).


During hyperoxia a decrease in T2 of peripheral CSF was found while the T2 of ventricular CSF remained constant.


This work was supported by the European Research Council, ERC grant agreement n°337333.


1. Derdeyn, et al., Variability of cerebral blood volume and oxygen extraction: stages of cerebral haemodynamic impairment revisited. Brain, 2002. 125: p. 595-607.

2. Jarnum, et al., Cisternal fluid oxygen tension in man. Neurology, 1964(14): p. 703-707.

3. Bloor, et al., A study of cerebrospinal fluid oxygen tension preliminary experimental and clinical observation. Arch Neurol, 1961. 4(1): p. 37-46.

4. Zaharchuk, et al., Measurement of cerebrospinal fluid oxygen partial pressure in humans using MRI. Magn Reson Med, 2005. 54(1): p. 113-21.

5. Binnewijzend, et al., Distinct perfusion patterns in Alzheimer's disease, frontotemporal dementia and dementia with Lewy bodies. Eur Radiol, 2014. 24(9): p. 2326-33.

6. Qin, A simple approach for three-dimensional mapping of baseline cerebrospinal fluid volume fraction. Magn Reson Med, 2011. 65(2): p. 385-91.

7. Ito, et al., Non-invasive prospective targeting of arterial PCO2 in subjects at rest. J Physiol, 2008. 586(15): p. 3675-3682.

8. de Vis, et al., Cerebrospinal fluid volumetric MRI mapping as a simple measurement for evaluating brain atrophy. Eur Radiol, 2015.


Figure 1: Design of the experiment. The first T2 mapping repetition is performed at baseline breathing, during the remaining repetitions (starting at the second repetition) a hyperoxic gas mixture is administered continuously.

Figure 2: A; planning of the T2 mapping scans through the lateral ventricles. B; corresponding T2 weighted MLEV images for each effective echo time. C; the used ROI masks: the peripheral CSF (PER, red), and the lateral ventricles (LAT, yellow).

Table 1: Mean T2 values of CSF in the lateral ventricles and in the periphery, for the five consecutively acquired T2 maps (1:22min each). a) p<0.05 vs repetition 1, b) p<0.05 vs repetition 2, c) p<0.05 vs repetition 3, d) p<0.05 vs repetition 4, e) p<0.05 vs repetition 5

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)