Figures and data
Clinical characteristics of all patients
Dynamic patterns of cerebrospinal fluid (CSF) drainage and signal changes across brain regions following intrathecal contrast administration.
a-c Three anatomical planes for region of interest (ROI) placement. ROIs were placed in the skull bone marrow (SBM) (red circle), subarachnoid space (SAS) (blue circle) and cortex (yellow circle) adjacent to three anatomical landmarks: the superior sagittal sinus (SSS) (first column), lateral fissure (second column) and cisterna magna (third column). The venous sinus is indicated by the lavender circle. White boxes in the top row images indicate regions that are magnified in subsequent rows. d-o Sequential MRI images show ROIs at different time points following intrathecal gadodiamide administration: baseline (0 h), 4.5 hours (4.5 h), 15 hours (15 h), and 39 hours (39 h). Percentage change (PC) values are labeled to illustrate signal variations over time. p-x Percentage change in signal unit ratio from baseline to post-contrast administration in the SBM (p-r), SAS (s-u), and cortex (v-x) adjacent to the SSS, lateral fissure and cisterna magna. Data are presented as mean ± standard error of the mean. Statistically significant differences between time points, identified using linear mixed model analysis, are indicated by asterisks: * p < 0.05, *** p < 0.001, **** p < 0.0001. y High-magnification image identifying key structures with arrowheads: skull (orange), SBM (red), SAS (blue), and dura mater (green). z Hypothesized patterns of SBM drainage. There is potential bidirectional communication between CSF in the SBM and SAS. Drainage in SBM near the SSS is likely relatively rapid, mainly flowing into the SAS. Conversely, near the cisterna magna, CSF in the SAS primarily drains into the SBM. Intermediate regions, such as those near the lateral fissure, may exhibit characteristics of balanced bidirectional drainage. Given the unbalanced drainage near the SSS and cisterna magna, we hypothesize that there might be an alternative drainage pathway in between skull and brain, which could expedite the flow of CSF from areas neighboring the SBM adjacent to the cisterna magna to those near the SSS. Moreover, the early peak enrichment at 4.5 hours in SBM suggests that this pathway may be very rapid.
Influencing factors related to drainage function in skull bone marrow (SBM) near the superior sagittal sinus (SSS).
Female (a), patients with hypertension (b) or diabetes mellitus (c) exhibit lower percentage changes of signal unit ratio from baseline to 4.5 hours in SBM near the SSS, compared to male and patients without these conditions. Age (d) and the Pittsburgh Sleep Quality Index (PSQI) total scores (e) are positively correlated with the percentage change, while the Telephone Montreal Cognitive Assessment (T-MoCA) total scores (f) was negatively correlated with the percentage change. Neutrophil percentage (g) is positively related, while monocyte (h) and lymphocyte percentages (i) are negatively correlated with the percentage change. Each plot displays the sample size, the fitted regression line, and the respective Spearman or Pearson correlation coefficients (ρ or R), with corresponding P-values.
Correlations between sleep quality and its sub-items and percentage change of signal unit ratio from baseline to 4.5 hours in SBM near the SSS.
Correlations between cognitive function and its sub-items and percentage change of signal unit ratio from baseline to 4.5 hours in SMB near the SSS.
Mediation analysis evaluating whether the association of sleep quality with cognitive performance is mediated by the drainage function of the skull bone marrow (SBM) adjacent to the superior sagittal sinus (SSS).
The direct effect of Pittsburgh Sleep Quality Index (PSQI) score on Telephone Montreal Cognitive Assessment (T-MoCA) score is denoted by c’ (β = -0.1478). The indirect effect through the drainage function of SBM adjacent to the SSS is captured by path a (β = 0.0125*) to path b (β = -7.3881**), resulting in an indirect effect (ab) of β = -0.0971*. The intermediate effect explains 38% of the total effect, highlighting the significant mediating influence of CSF drainage efficiency on cognitive outcomes relative to sleep quality. Asterisks denote levels of statistical significance with covariates considered: * p < 0.05, ** p < 0.01.
Tracer enrichment in skull bone marrow (SBM) among different diseases.
This figure presents the percentage changes in the signal unit ratio from baseline to 4.5, 15, and 39 hours post cerebrospinal fluid (CSF) tracer administration across different SBM regions in patients with encephalitis (a-c, n = 15), neurodegenerative disease (d-f, n = 19), and peripheral neuropathy (g-i, n = 45). Error bars denote the standard error of the mean. Statistically significant differences between pairs of time points, as identified through linear mixed model analysis, are marked with stars: * p < 0.05, ** p < 0.01, **** p < 0.0001.
Correlation between tracer enrichment in skull bone marrow (SBM) and subarachnoid space (SAS).
This figure elucidates the relationship between the percentage changes (PC) in the signal unit ratio of SBM and SAS from baseline to 4.5, 15, and 39 hours post cerebrospinal fluid (CSF) tracer administration in three anatomical regions: the superior sagittal sinus (a-c), the lateral fissure (d-f), and the cisterna magna (g-i). Each panel displays the sample size, a fitted regression line, and the Spearman correlation coefficient (ρ) alongside the associated P-value.
Tracer enrichment in venous sinus and nasal turbinates.
a Percentage change in the signal unit ratio from baseline to 4.5, 15, and 39 hours post-cerebrospinal fluid (CSF) tracer administration in venous sinus. b-d Percentage change in the signal unit ratio from baseline to 4.5, 15, and 39 hours post-CSF tracer administration in the superior turbinate (b), middle turbinate (c), and inferior turbinate (d). Data are presented as mean ± standard error of the mean. Statistically significant differences between time points, identified using linear mixed model analysis, are indicated by asterisks: * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.
Correlation between tracer enrichment in skull bone marrow (SBM) and venous sinus (VS).
This figure illustrates the relationship between the percentage changes (PC) in the signal unit ratio of SBM and the VS from baseline to 4.5, 15, and 39 hours post-cerebrospinal fluid (CSF) tracer administration in three anatomical regions: the superior sagittal sinus (SSS) (a-c), the lateral fissure (d-f), and the cisterna magna (g-i). Each panel displays the sample size, a fitted regression line with 95% confidence intervals (dotted lines), and the Spearman correlation coefficient (ρ) alongside the corresponding P-value, indicating the strength and statistical significance of the correlations observed.
Correlation between tracer enrichment in skull bone marrow (SBM) and cortex.
This figure illustrates the relationship between the percentage changes (PC) in the signal unit ratio of SBM and the cortex from baseline to 4.5, 15, and 39 hours post-cerebrospinal fluid (CSF) tracer administration in three anatomical regions: the superior sagittal sinus (SSS) (a-c), the lateral fissure (d-f), and the cisterna magna (g-i). Each panel displays the sample size, a fitted regression line with 95% confidence intervals (dotted lines), and the Spearman correlation coefficient (ρ) alongside the corresponding P-value, indicating the strength and statistical significance of the observed correlations.
Influencing factors related to drainage function in SMB near the SSS in the subgroup analysis of patients with peripheral neuropathy.
In the subgroup analysis of patients with peripheral neuropathy (n=45), no difference of the percentage changes of signal unit ratio from baseline to 4.5 hours in skull bone marrow (SBM) near the superior sagittal sinus (SSS) was found in female and male patients (a). Patients with hypertension (b) or diabetes mellitus (c) exhibit lower percentage changes compared to male and patients without these conditions. Age (d) and the Pittsburgh Sleep Quality Index (PSQI) total scores (e, n=34) are positively correlated with the percentage change, while the Telephone Montreal Cognitive Assessment (T-MoCA) total scores (f, n = 34) was negatively correlated with the percentage change. Each plot displays the sample size, the fitted regression line, and the respective Spearman or Pearson correlation coefficients (ρ or R), with corresponding P-values.
