Introduction

The concept of subarachnoid space dates back to the 17th century, when Humphrey Ridley – a British physician- first described the basal cisterns such as the cerebellomedullary, quadrigeminal, and olfactory cisterns. He provided evidence for the presence of the arachnoid membrane as a separate meningeal layer and detailed how this membrane enveloped different cerebral vessels and intracranial nerves [4, 7]. Until recently it was accepted that the subarachnoid space represented a continuous space filled with cerebrospinal fluid (CSF) [2, 9]. However, a recent publication authored by Mollgard et al. has revealed the existence of a mesothelial layer that partitions the subarachnoid space into two distinct compartments [5]. This layer, termed the subarachnoid lymphatic-like membrane (SLYM), acts as a semipermeable barrier, restricting the passage of molecules larger than 3 kilodaltons and containing immune cells involved in innate defense mechanisms.

The SLYM was primarily studied in murine models, where its anatomical and functional characteristics were characterized in detail. While the study suggested similar anatomical features in humans, these remain insufficiently explored [5]. This work aims to investigate imaging correlates of the SLYM in human patients, leveraging pathologic scenarios such as subarachnoid hemorrhage (SAH).

Given the fragility of the SLYM, histopathological assessment is challenging, as the layer is prone to fragmentation during dissection of cadaveric specimens. This is why it is hypothesized to have remained undiscovered until recently. In vivo imaging presents an alternative means of studying this structure. In vivo assessment with iodinated contrast or gadolinium-based contrast agents is not feasible due to their small molecular size [3, 8]. Acute aneurysmal subarachnoid hemorrhage (aSAH) provides an opportunity to use blood as a natural contrast agent, as the SLYM is hypothesized to be impermeable to cellular blood components [1, 6]. As such, the current study aims to assess whether imaging can be used to evaluate the compartmentalization of the subarachnoid space and indirectly deduce the anatomic reflection of the SLYM utilizing acute aSAH as a model.

Materials and methods

Patients’ selection

This retrospective single-center study was approved by the research ethics board with waiver of informed consent. All head CT scans with aSAH between January 2015 and June 31 2022 were identified by searching the departmental imaging database using specific keywords from CT reports (including “subarachnoid hemorrhage” and “aneurysm”). All adult (> 18-year-old) patients who presented with acute aSAH and had baseline routine plain head CT with multiplanar reformats were included. The cases with no sagittal reformats, poor visualization of the posterior fossa due to significant artifacts, or non-aneurysmal SAH (i.e. AVM, dAVF, Trauma) were excluded. Patient demographics, World Federation of Neurosurgical Societies grading system, aneurysm location, aneurysm size and type were retrieved using electronic medical records.

Image acquisition and analysis

All the standard-of-care head CTs were performed on 64-section CT scanners (Siemens Healthcare, Forchheim, Germany). Unenhanced CT scans covered the area from the skull base to the vertex with multiple reformats.

The initial unenhanced CT head was used to grade the subarachnoid hemorrhage using the modified Fisher scale (MFS). Additionally, the midline sagittal images were used to classify the distribution of subarachnoid blood within the basal cisterns into predominantly “Superficial” or predominantly “Deep” subarachnoid space based on the proximity of the hemorrhage to the dural or pial surfaces, respectively. Further classification was done based on the location in the interpeduncular, pre-pontine, pre-medullary cisterns and peri-spinal CSF space at the foramen magnum. The subarachnoid space within the cerebral convexities and fissures were classified into prominent or narrow, based on the degree of volume loss, and further assessment for compartmentalization was done only if the space is prominent (i.e. those demonstrating volume loss). CTA scans were used to assess the location, size, and shape of the aneurysms.

Statistical analysis

Statistical analysis was performed using SPSS (version 26.0) IBM Corp. Released 2019. IBM SPSS Statistics for Windows, Version 26.0. Armonk, NY: IBM Corp. (n.d.). [Computer software] and R (version 4.3.1). Descriptive statistics for patients’ demographic data were generated. The modified Fischer Scores were dichotomized into two groups; Low (Score 1–2), and High (Score 3–4). The relationship and distribution of SAH at different cisterns along the dural and pial surfaces with the degree of MFS was assessed using Chi-Square correlation test. Statistical significance was set at p-value < 0.05.

Results

A total of 97 patients with aSAH were included with an average age of 60.36 +/- 12.1 years (28–89) composed of 75 (77.3%) females and 22 (22.7%) males. The majority (72.2%) of subarachnoid hemorrhage cases had high MFS (3–4), with the remaining 27.8% having low MFS (1–2).

A total of 63 cases (64.9%) had the hemorrhage present in the “deep” juxta-pial subarachnoid space along the brainstem only, whereas no cases were seen only in the “superficial” juxta-dural subarachnoid space. Furthermore, 26 cases (26.8%) had hemorrhage in both the “superficial” and “deep” compartments along the brainstem. There were 8 cases (8.2%) with no subarachnoid hemorrhage along the brainstem but had hemorrhage along the cerebral convexities.

There is a significant association between the severity of aSAH - quantified by the MFS - and the distribution of the blood. Patients with higher MFS scores were roughly 7.6 times (p-value = 0.049) more likely to have hemorrhage at the “Superficial” juxta-dural subarachnoid compartment when compared to those with lower MFS scores.

At low MFS, hemorrhage present at the “Deep” juxta-pial subarachnoid space at any location along the brainstem is significantly associated with subarachnoid hemorrhage at other “Deep” juxta-pial locations (Table 1). For example, “Deep” juxta-pial aSAH at the pre-medullary cistern is significantly associated with juxta-pial blood at the prepontine (χ2 = 11.2; p-value = 0.001) and peri-spinal cisterns (χ2 = 5.3; p-value = 0.021). However, we found no cases at low MFS with “Superficially” distributed blood along the Juxta-dural subarachnoid space at the peri-spinal or pre-medullary cisterns and only 3 cases in the prepontine cistern but without statistically significant associations of aSAH with other locations (p value > 0.077, 0.964, 0.233). On the other hand, high mFS aSAH was found to distribute along the “superficial” or “deep” compartments or both. For example, juxta-pial peri-spinal aSAH was significantly associated with juxta-pial pre-medullary (χ2 = 10.1; p-value = 0.001) and juxta-dural peri-spinal and pre-medullary (both χ2 = 4.8; p-value = 0.028) aSAH. Similarly, juxta-dural prepontine aSAH was significantly associated with juxta-dural premedullary (χ2 = 10.3; p-value = 0.001) and peri-spinal (χ2 = 5.0; p-value = 0.025) aSAH.

Table 1 Relationship and association of hemorrhage at different location along the dural-adjacent and pial-adjacent subarachnoid space
Full size table

Discussion

The current study is the first in-vivo study to evaluate for the compartmentalization of the subarachnoid space, leveraging the distribution of blood products in the setting of acute aSAH. Our results suggest a potential imaging-based correlation to the recently discovered SLYM, which partitions the subarachnoid space into two distinct compartments. This correlation appears particularly evident in lower-grade aSAH (MFS 1–2), where blood functions as an internal contrast agent, highlighting the suspected location of this membrane. Conversely, in cases with more extensive hemorrhage (MFS 3–4), visualizing the partition layer becomes challenging. We hypothesize that this could be due to membrane displacement or rupture secondary to the high blood volume and elevated pressure within these confined spaces (Fig. 1).

Fig. 1
figure 1

Illustration depicting the compartmentalization of aneurysmal subarachnoid hemorrhage in high and low modified Fisher scale (MFS). (A) Patients with low MFS of 1–2 are more likely to have SAH compartmentalize towards the deeper or pial- adjacent subarachnoid space, which is hypothesized to be related to the SLYM acting as a barrier. (B) Patients with higher MFS of 3–4 are more likely to have SAH abutting both pial and dural surfaces, which can be explained by the rupture of the SLYM at higher pressures allowing the subarachnoid blood deep to the SLYM to distribute superficial to the SLYM. (C) As the SLYM is not directly visualized in CT head, another explanation for the apparent redistribution of subarachnoid hemorrhage towards the dural surface is the displacement of the SLYM towards the dura. (D) Midline sagittal CT brain demonstrating SAH along the pial surface of the premedullary cistern. In the prepontine cistern blood abuts the pial and dural surfaces. Axial CT brain at the level of foramen magnum/upper cervical cord shows semi circumferential SAH along the pial surface of the cord

Full size image

Future research could further advance our anatomical understanding by exploring the feasibility of measuring the distance between the suspected SLYM and adjacent structures, as well as quantifying the blood layer thickness. These metrics, though presently limited by the heterogeneous distribution of blood and variability in cisternal dimensions, hold the potential to provide more objective insights into subarachnoid compartmentalization. Such approaches could be particularly valuable in prospective studies utilizing advanced imaging techniques.

Identifying this membrane and understanding its role could have significant implications for various neurological disorders. Elucidating its influence on CSF circulation could shed light on related pathologies. For example, In neurodegenerative diseases like Alzheimer’s, where CSF flow is disrupted, the SLYM’s compartmentalizing function could help explain such abnormalities. If relevant, targeting the SLYM or its pathways may offer novel therapeutic avenues. Additionally, this discovery could pave the way for imaging-based tools to assess CSF dynamics and identify patients at risk for disorders like normal pressure hydrocephalus (NPH), ultimately guiding treatment decisions. Future investigations utilizing advanced imaging techniques with higher resolution are warranted to further validate these observations and refine our understanding of this newly described anatomical structure. Such advancements could pave the way for improved diagnostic accuracy and potentially inform future therapeutic strategies related to aSAH and other neurological conditions.

Conclusion

This study suggests an imaging correlate to the recently discovered SLYM membrane, potentially influencing aSAH compartmentalization, particularly in low-grade hemorrhage. While compartmentalization is limited in high-grade cases. This study builds on findings primarily derived from murine models. While imaging correlates provide indirect evidence for the SLYM, its exact anatomical reflections and function in humans require further validation through advanced imaging modalities and histopathological studies.