High grade intracranial gliomas exhibit widespread impaired cerebrovascular reactivity

Christiaan Hendrik Bas van Niftrik¹, Marco Piccirelli², Jan-Karl Burkhardt¹, Athina Pangalu², Antonio Valavanis², Michael Weller³, Oliver Bozinov¹, Luca Regli¹, and Jorn Fierstra¹

¹Department of Neurosurgery, University Hospital Zurich, Zurich, Switzerland, ²Department of Neuroradiology, University Hospital Zurich, Zurich, Switzerland, ³Department of Neurology, University Hospital Zurich, Zurich, Switzerland

Synopsis

Neurovascular uncoupling (false negative BOLD activation) can be found in patients with high grade gliomas. An underlying mechanism could be impaired cerebrovascular reactivity. We determined overall cerebrovascular reactivity (CVR) as well as the perifocal CVR and used a healthy control group. We applied an automated tumor masking with determination of CVR in 7 consecutive rings of 3 mm. We found an overall impaired CVR as well as significantly impaired intratumoral CVR and perifocal up to 12 mm. No such trend was found on the contralateral hemisphere, after flipping of the tumor mask.

Introduction

In functional MRI, neurovascular coupling is the mechanism by which a neuronal induced vascular response leads to subsequent blood oxygen-level-dependent (BOLD) MRI signal increase. In pre-surgical planning of intracranial high grade gliomas, false negative task based fMRI activation, termed neurovascular uncoupling, can be found potentially leading to undesirable resection of eloquent areas. Neurovascular uncoupling may be identified with BOLD derived cerebrovascular reactivity (CVR), where impaired CVR results in decreased fMRI task-based activation in eloquent areas. CVR has not been spatially quantified for high grade intracranial gliomas.

Method

Ten subjects (mean age 55±18) with high grade gliomas (WHO grade III or IV) underwent a standardized hypercapnic challenge during 3T BOLD MRI acquisitions. The tumor was masked using an automated subtraction analysis based on (non)contrast enhanced T1-weighted and FLAIR T2 volumes. Voxels with FLAIR hyperintensities and a positive subtraction value were determined to be the tumor outline. Around the tumor CVR was determined in 7 concentric rings with an individual expansion of 3mm. The ring analysis is corrected to only include grey and white matter voxels with a combined likelihood greater than 0.9 (SPM12 segmentation). As comparison a similar contralateral analysis was done. Furthermore, 10 healthy controls were included.

Results

Whole brain CVR of high grade glioma subjects was significantly impaired compared to healthy controls (0.11±0.10 versus 0.28±0.8, p<0.01) Intratumoral CVR was 0.04±0.07. Figure1 shows the overall impairment in both hemispheres for one patient. Perifocal CVR exhibited an increasing CVR with every consecutive concentric ring expansion up to 12mm perifocally, however, remained significantly impaired as to CVR measured in controls. No such trend was observed in the contralateral hemisphere. Figure 2 shows the comparison averaged on 10 patients of the tumor periphery with the contralateral tissue.

Conclusion

Whole brain CVR of subjects with high grade gliomas is impaired, with a more severe CVR impairment seen in tumor tissue and the first 12mm perifocally. No such trend was found on the contralateral side. BOLD-derived quantitative CVR measurements may provide a potential tool to assess neurovascular uncoupling in high grade glioma.

Acknowledgements

No acknowledgement found.

References

Ben Bashat, D., et al. (2012). "Hemodynamic response imaging: a potential tool for the assessment of angiogenesis in brain tumors." PLoS One 7(11): e49416.

Hsu, Y. Y., et al. (2004). "Blood oxygenation level-dependent MRI of cerebral gliomas during breath holding." J Magn Reson Imaging 19(2): 160-167.

Pillai, J. J. and D. J. Mikulis (2015). "Cerebrovascular Reactivity Mapping: An Evolving Standard for Clinical Functional Imaging." AJNR Am J Neuroradiol 36(1): 7-13.

Figures

CVR ring analysis of patient with gliomblastoma. A) Contrast enhanced T1-weighted image. B) FLAIR-T2. C) Ring Analysis illustrated on the T1-weighted image. Different colors represent different consecutive expanding 3mm-layers surrounding the tumor. D) BOLD-CVR of this patient. CVR colormap-thresholds between -0.6% and 0.6% BOLD signal-change/mmHg CO2. E) healthy control BOLD-CVR.

A) Peritumoral CVR linear increase of the BOLD-signal-change/mmHg CO2. Up to 12 mm peritumoral, the CVR differs significantly from mean CVR. B) On the contralateral Hemisphere, no upwards trend is found. This means that the particular brain voxels included within the rings do not confound our results.

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

4176