Hemodynamic response function description in patients with glioma

Stéren Chabert; Rodrigo Salas; Erika Cantor; Alejandro Veloz; Astrid Cancino; Matías González; Francisco Torres; Carlos Bennett

doi:10.1016/j.neurad.2023.10.001

Hemodynamic response function description in patients with glioma

J Neuroradiol. 2023 Oct 6:S0150-9861(23)00247-X. doi: 10.1016/j.neurad.2023.10.001. Online ahead of print.

Authors

Stéren Chabert¹, Rodrigo Salas², Erika Cantor³, Alejandro Veloz⁴, Astrid Cancino⁵, Matías González⁶, Francisco Torres⁷, Carlos Bennett⁶

Affiliations

¹ School of Biomedical Engineering, Universidad de Valparaiso, General Cruz 222, Valparaiso, Chile; Millennium Science Initiative Intelligent Healthcare Engineering, Santiago, Chile. Electronic address: steren.chabert@uv.cl.
² School of Biomedical Engineering, Universidad de Valparaiso, General Cruz 222, Valparaiso, Chile; Millennium Science Initiative Intelligent Healthcare Engineering, Santiago, Chile.
³ Institute of Statistics, Universidad de Valparaíso, Valparaíso, Chile.
⁴ School of Biomedical Engineering, Universidad de Valparaiso, General Cruz 222, Valparaiso, Chile.
⁵ Millennium Science Initiative Intelligent Healthcare Engineering, Santiago, Chile; Doctorado en Ciencias e Ingeniería para la Salud, Universidad de Valparaiso, Valparaiso, Chile.
⁶ Neurosurgery Department, Hospital Carlos van Buren, Valparaiso, Chile.
⁷ Millennium Science Initiative Intelligent Healthcare Engineering, Santiago, Chile; Radiology Department, Hospital Carlos van Buren, Valparaiso, Chile.

PMID: 37805126
DOI: 10.1016/j.neurad.2023.10.001

Abstract

Introduction: Functional magnetic resonance imaging is a powerful tool that has provided many insights into cognitive sciences. Yet, as its analysis is mostly based on the knowledge of an a priori canonical hemodynamic response function (HRF), its reliability in patients' applications has been questioned. There have been reports of neurovascular uncoupling in patients with glioma, but no specific description of the Hemodynamic Response Function (HRF) in glioma has been reported so far. The aim of this work is to describe the HRF in patients with glioma.

Methods: Forty patients were included. MR images were acquired on a 1.5T scanner. Activated clusters were identified using a fuzzy general linear model; HRFs were adjusted with a double-gamma function. Analyses were undertaken considering the tumor grade, age, sex, tumor location, and activated location.

Results: Differences are found in the occipital, limbic, insular, and sub-lobar areas, but not in the frontal, temporal, and parietal lobes. The presence of a glioma slows the time-to-peak and onset times by 5.2 and 3.8 % respectively; high-grade gliomas present 8.1 % smaller HRF widths than low-grade gliomas.

Discussion and conclusion: There is significant HRF variation due to the presence of glioma, but the magnitudes of the observed differences are small. Most processing pipelines should be robust enough for this magnitude of variation and little if any impact should be visible on functional maps. The differences that have been observed in the literature between functional mapping obtained with magnetic resonance vs. that obtained with direct electrostimulation during awake surgery are more probably due to the intrinsic difference in the mapping process: fMRI mapping detects all recruited areas while intra-surgical mapping indicates only the areas indispensable for the realization of a certain task. Surgical mapping might not be the gold standard to use when trying to validate the fMRI mapping process.

Keywords: Functional MRI; Glioma; Hemodynamic response function.