Diffuse glioma perfusion quantification with ASL and DSC: head-to-head comparison with 15O-H2O PET


Diffuse glioma perfusion quantification with ASL and DSC: head-to-head comparison with 15O-H2O PET

Petr, J.; Verburg, N.; Kuijer, J. P. A.; Koopman, T.; Keil, V. C.; Warnert, E. A. H.; Barkhof, F.; van den Hoff, J.; Boellaard, R.; de Witt Hamer, P. C.; J. M. M. Mutsaerts, H.

Background: Arterial spin-labeling (ASL) is a viable non-invasive alternative to dynamic susceptibility contrast (DSC) for measuring cerebral blood flow (CBF). While quantitative accuracy of ASL and DSC was compared in healthy volunteers and patients, a comparison to a reference standard in patients diagnosed with glioma is still missing.
Purpose: To probe the quantitative agreement between perfusion measurements based on ASL and DSC in comparison to the gold standard of PET in patients with glioma.
Materials and Methods: This secondary analysis included pre-surgical ASL and DSC perfusion measurements in participants diagnosed with grade 2-4 gliomas drawn from the prospective FRONTIER study (Dutch National Trial Register - NTR5354) and compared the two techniques with the gold-standard perfusion measurement 15O-H2O-PET. The quantitative comparison was performed both in normal-appearing tissue as well as in the tumor region. The mean, maximum, and voxel-wise perfusion values were with and without normalization to normal-appearing tissue. And finally, a qualitative analysis was performed on individual cases to help interpret the quantitative results.
Results: Eight patients (age 40.5 ± 17.0 years, 3 women) were analyzed. ASL showed better voxel-wise agreement with PET in the normal-appearing tissue than DSC (mean relative error of 26.8% vs. 33.8%). Within the tumor, cerebral blood flow (CBF) normalized contralateral gray matter showed similar tumor maximum values in both techniques - mean relative error of 23.2% for ASL and 22.0% for DSC. However, the mean relative error on a voxel-wise basis was better for ASL (27.2%) than for DSC (35.0%). Qualitatively, ASL tended to overestimate CBF in macrovessels, and DSC tended to overestimate CBF in non-enhancing tumors with small vessel diameters.
Conclusion: While neither ASL nor DSC can readily replace 15O-H2O-PET in tumor quantitative perfusion measurement, ASL CBF presents a viable non-invasive semi-quantitative alternative to DSC for glioma imaging.

  • Contribution to proceedings
    Deutsche Gesellschaft für Nuklearmedizin, 27.04.2022, Leipzig, Germany
  • Lecture (Conference)
    Deutsche Gesellschaft für Nuklearmedizin, 27.04.2022, Leipzig, Germany

Permalink: https://www.hzdr.de/publications/Publ-33840
Publ.-Id: 33840