A Novel Finding in Cerebral Blood Flow in Patients with Sickle Cell Disease Using Bold Functional MRI

Abstract

Background: Sickle cell disease (SCD) is an inherited blood disorder which is associated with acute and chronic pain, as well as other vascular complications such as arterial steno-occlusion, stroke, and silent cerebral infarction. Alterations of cerebral blood flow have been found to be associated with different painful stages in SCD, thus cerebral blood flow might be correlated with the disease severity determined by varied pain conditions in SCD. Blood oxygen level dependent (BOLD) imaging is a blood-related contrast technology in functional magnetic resonance imaging (fMRI). We have previously demonstrated the feasibility of using low frequency oscillation (LFO) in BOLD fMRI to track the hemodynamic features of the blood flow with maximum cross-correlation coefficient (MCCC) and the Delay Time (i.e., blood arrival time at different brain regions) (Tong et al., JCBFM. 2017). Positive MCCC value was found between the LFO BOLD signals extracted from the global mean (GMean, averaged signal across the whole brain) and the superior sagittal sinus (SSS) in our previous findings (Tong et al., JCBFM. 2018). In this study, we aim to use LFO BOLD signals to explore the cerebral hemodynamic features with MCCCs and their associations with hematological and disease severity related outcomes in patients with SCD. Methods: Ten subjects with SCD aged 14-43 years old (5 HbSS, 3 HbSC, 1 HbSB+, 1 HbB0) and 10 age- and sex-matched controls from a publicly available dataset Human Connectome Project (HCP) (Essen et al., Neuroimage, 2013) were included for data analysis. T1-weighted and resting-state BOLD fMRI (repetition time (TR)=1.05 sec) were collected using a 3T SIEMENS (Prisma) scanner. A BOLD fMRI dataset with TR of 0.72 sec from healthy subjects was used as control. All the BOLD fMRI data were preprocessed using FSL to correct the motion artifact. The SSS region was identified in the T1-weighted image using FSL. The corresponding SSS mask was then applied in the BOLD fMRI image to extract the BOLD fMRI signals in the SSS region. The GMean BOLD signal was averaged across the whole brain. BOLD fMRI Signals were demeaned and band-pass filtered (0.01-0.1 Hz) to acquire the LFO signals. Cross-correlation between the SSS-LFO signal and GMean-LFO signal was performed to calculate the MCCC and their corresponding Delay Time. Correlations between the MCCCs of the two LFO signals (from SSS and GMean) and each hematological index were further performed. Results: Consistent with the healthy control in our previous findings, half of the patients (3 HbSC, 1 HbSB+, and 1 HbSS) presented positive MCCC values. However, interestingly, negative MCCC values were found from another half of the patients (4 HbSS and 1 HbB0, Figure 1, p < 0.05 for all subjects). Notably, all patients with HbSC presented positive MCCC values that are similar to controls whereas 4 out 5 patients with HbSS (80%) were found to have negative MCCC values. More importantly, higher percentages of HbS and lower total Hgb (HbS(%): r=-0.73, p=0.03; Hgb(GM/dL): r=0.69, p=0.03) were found to be significantly associated with the negative MCCC values. Conclusions: We found a never-before-seen negative MCCC signature between the BOLD fMRI signals in the SSS and the GMean in a subset of patients with SCD compared with healthy controls. This might indicate a significant change in cerebral blood flow in a subset of patients with SS disease compared with SC and controls. Importantly, we further observed that this never-before-seen negative MCCC signature is highly associated with the percentage of the HbS and total Hgb that are known to be correlated with disease severity. This might be due to the microvascular shunting (Juttukonda et al., JCBFM, 2019) that exists in patients who have high percentage of HbS. This study is the first to associate BOLD fMRI derived hemodynamic metrics with disease severity related outcomes in SCD. This never-before-seen anti-correlation MCCC signature may have tremendous promise as a novel biomarker for SCD severity and differentiate pain characteristics in different sickle cell phenotypes, and further investigation in larger number of subjects is needed to understand variations between SCD phenotypes and its relationship to clinical outcomes. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal