Reproducibility of cerebrovascular reactivity measures in children using BOLD MRI.

Abstract

To evaluate the reproducibility of cerebrovascular reactivity (CVR) measurements acquired in children using magnetic resonance imaging (MRI) in combination with a computer-controlled carbon dioxide (CO2 ) stimulus. Ten healthy children (age 16.1 ± 1.6 years) underwent CVR imaging on a 3T scanner using a blood-oxygen level-dependent (BOLD) MRI sequence. Targeted hypercapnia was induced during imaging with a CO2 gas challenge delivered using a specialized gas sequencer (RespirAct). A total of four BOLD scans were performed over 2 separate days to test within-day and between-day consistency of the data. CVR values were computed by correlating the relative change in BOLD signal in response to the CO2 stimulus delivered to the each subject. Intraclass correlation coefficients (ICCs) of within-day values show highly reproducible measures in both the gray matter (ICC = 0.857, P < 0.001) and white matter (ICC = 0.895, P < 0.001). Relatively lower between-day reproducibility was observed in both the gray matter (ICC = 0.776, P = 0.001) and white matter (ICC = 0.719, P = 0.004). Using a computer-controlled CO2 stimulus, we have demonstrated the reliability of BOLD-CVR measurements in pediatric subjects. Within-day and between-day metrics of reproducibility were comparable to adult data.