First in vivo visualization of the human subarachnoid space and brain cortex via optical coherence tomography.

The present work explores optical coherence tomography (OCT) as a suitable in vivo neuroimaging modality of the subarachnoid space (SAS). Patients (n = 26) with frontolateral craniotomy were recruited. The temporal and frontal arachnoid mater and adjacent anatomical structures were scanned using microscope-integrated three-dimensional OCT, (iOCT). Analysis revealed a detailed depiction of the SAS (76.9%) with delineation of the internal microanatomical structures such as the arachnoid barrier cell membrane (ABCM; 96.2%), trabecular system (50.2%), internal blood vessels (96.2%), pia mater (26.9%) and the brain cortex (96.2%). Orthogonal distance measuring was possible. The SAS showed a mean depth of 570 µm frontotemporal. The ABCM showed a mean depth of 74 µm frontotemporal. These results indicate that OCT provides a dynamic, non-invasive tool for real-time imaging of the SAS and adjacent anatomical structures at micrometer spatial resolution. Further studies are necessary to evaluate the value of OCT during microsurgical procedures.

Introduction

The subarachnoid space (SAS) is the cavity between the arachnoid barrier cell membrane (ABCM) and the pia mater, filled with cerebrospinal fluid (CSF). The microstructural composition is considered a basis for the understanding of physiological and further pathophysiological functions of the brain.[1] Although representing the largest continuous intrathekal compartment, the SAS of the cerebral convexity, in particular, hides from in vivo visualization due to technical limitations. Conventional imaging tools like magnetic resonance imaging (MRI) and ultrasound (US) - currently the most established tool for dynamic sectional imaging of the brain - are llmited here due to their lack of spatial resolution.[2-4]

In contrast, optical coherence tomography (OCT) shows an exceedingly high maximal spatial resolution of 1–15 µm, which approaches the resolution of conventional histopathology.[5] OCT imaging depends on the detection of back scattered near-infrared light. It is harmless to biological tissue.[6] Due to these physical properties light microscope integration is possible. Therefore, immediate three-dimensional (3D), sectional scanning of tissue during microsurgical procedures is feasible. Medical applications already span the fields of neurology, cardiology, dermatology and ophthalmology.[7-10] In ophthalmology it is regularly implemented in vitroretinal surgical setups.[11]

In neuroimaging polarization sensitive OCT has proven its potential for delineation of white and grey matter in rat brain[12] and fibre tracts in the human brain in post mortem conditions.[13] In the field of neurosurgery, OCT could be used as a device for ‘optic biopsy’.[14,15] Ex vivo solid tumors, diffusely invaded brain tissue and the normal brain in human gliomas could be differentiated.[16] In vivo, during glioma surgery, residual tumors could be depicted in the mouse[17] and human brain.[18]

Ex vivo, in a porcine brain with released CSF, OCT imaging could delineate the sulcal arachnoid trabeculae and arachnoid blood vessels,[19] but the study could not depict intact sulcal or gyral SAS.

Due to light translucent tissue properties in combination with a penetrating depth of 4000 µm and suitable spatial resolution[20] we hypothesized that OCT is a useful tool for the visualization of the microstructural composition of the human SAS in vivo.

Materials and methods

Participants

Patients (n = 26; female = 13, male = 13) with indication for frontolateral craniotomy were included. Diagnoses ranged from intracranial tumors (n = 13), unruptured cerebral aneurysms (n = 11) and arteriovenous malformations (n = 2). Patients were positioned supine or semi-supine. The study was approved by the local ethics committee (no. 3012-2016). All participants gave written informed consent.

Optical coherence tomography

Intraoperative OCT was enabled by connecting the OCT camera (OptoMedical Technologies GmbH, Lübeck, Germany) to the port of a surgical microscope optimized for light transmission in the near-infrared spectral range (HS Hi-R1000G, Haag-Streit Surgical GmbH, Wedel, Germany). OCT scanning was displayed on internal and external monitors as well as a transparent overlay in the field of view of the surgeon (head-up display). OCT volume scans and corresponding light microscopic pictures were further stored for post-procedural analysis. The OCT camera was optimized for the optics of the operating microscope. The scan head of the OCT device was connected with an optical light fibre cable and electronic cable to the OCT camera head. The OCT camera consisted of a light source at a central wavelength of 840 nm with a spectral bandwidth of 55 nm, a spectral domain detector with a speed of 15,000 A-scans/second. The detector speed resulted in an OCT speed of 10 frames/second. A two-axis scanner enabled 3D volume scans. Internal motorized reference optics allowed for different working distances ranging from 200 to 480 mm. The optical window depth was 4.2 mm in air and approximately 3.1 mm in brain tissue with an index of refraction of 1.34. The measured axial spatial resolution was approximately 10 µm in air and approximately 7.5 µm in tissue. The lateral scan width depended on the magnification of the microscope and varied from 5 mm up to 37 mm. The lateral resolution ranged from 23 µm up to 47 µm at a working distance of 232 mm depending on the microscope magnification. The system was CE certified (conformity with the European Union guidelines) for intraoperative documentation of tissue structures. For a detailed description, see Gallwas and colleagues.[21] The OCT setting is demonstrated in Figure 1.

Figure 1.

The OCT-camera (OptoMedical Technologies GmbH, Lübeck, Germany) is connected to the port of a surgical microscope (HS Hi-R1000G, Haag-Streit Surgical GmbH, Wedel, Germany), which is optimized for light transmission in near-infrared spectral range. OCT scanning is displayed on internal and external monitors and as a transparent overlay in the field of view of the surgeons (head-up display). Photograph with permission of OptoMedical.

iOCT, microscope-integrated three-dimensional OCT.

Image acquisition

The region of interest (ROI) was defined by the surgeon, according to the surgical approach. A rectangular scan angle to the surface of the ROI and the highest microscope magnification was intended. For each ROI the corresponding light microscopic picture and 3D OCT volume scan were recorded and stored with a data sheet defining: the date and time of scanning, scan angle, number of A-scans, number of B-scans, height in pixels, ranges of volume scan in mm, zoom of microscope and focus of microscope.

Anatomical analysis and distance measurements

Microscopic pictures of the scanned site and corresponding OCT volume scans were sorted according to image quality and richness of anatomical details. The aspect ratio of the selected OCT scans was adjusted according to the microscope magnification. For anatomical analysis the data sets were then screened by two independent neurosurgeons with experience in OCT imaging. For distance measurement an exemplary A-scan was selected from the volume scan. A rectangular measurement angle parallel to the optic path and measurement under maximum augmentation was maintained in any case. Whenever feasible, the center of the probe was selected to reduce scattering effects. The final measurement site was marked. All steps were done using ImageJ2 and Fiji software.[22,23] Measuring was not adjusted according to the index of refraction.

Results

Anatomical structures of the SAS

A detailed depiction of the microanatomical structures of the SAS and adjacent brain tissue was possible in most of the volume scans. The SAS (76.9%), ABCM (96.2%), trabecular system (66.7%), internal blood vessels (96.2%), pia mater (26.9%) and the brain cortex (96.2%) could be delineated, (Table 1, Figure 2; see supplemental material for 3D volume scan of the frontal SAS, S1).

Table 1.

Screening of anatomical structures.

	Volume scans		CSF in situ		CSF released
	n	%	n	%	n	%
Cases	26		12	46.2%	14	53.9%
SAS	20	76.9%	12	100.0%	12	85.7%
ABCM	25	96.2%	11	91.7%	0	0.0%
Trabecular system	13	50.2%	13	66.7%	9	60%
Blood vessels	25	96.2%	12	100.0%	12	85.7%
Pia mater	7	26.9%	3	25.0%	4	26.7%
Brain cortex	25	96.2%	11	91.7%	14	93.3%

Number of volume scans in which the corresponding anatomical structure could be delineated overall, before and after CSF release and percentage share. In CSF-released cases the perivascular SAS was screened.

ABCM, arachnoid barrier cell membrane; CSF, cerebrospinal fluid; SAS, subarachnoid space.

Figure 2.

Microscopic view of the frontal lobe and sylvian fissure. Orange line indicates the region of scan. B: OCT-scan of the region depicting the gyral SAS. C: Enlarged excerpt demonstrating details of the SAS. D: Schematic drawing of microstructures: 1. ABCM, 2. Trabecular system, 3. Internal blood vessels, 4. Pia mater and 5. Brain cortex.

CSF, cerebrospinal fluid; OCT, optical coherence tomography; SAS, subarachnoid space.

When the CSF was released microanatomical structures could still be depicted but were less detailed. The perivascular and gyral SAS (85.7%), trabecular system (60%), internal blood vessels (85.7%), pia mater (26.7%) and the brain cortex (93.33%) could be delineated. The ABCM (0%) could not be delineated, since a clear differentiation to the underlying trabecular system (60%) was not given (Figure 3, Table 1).

Figure 3.

Above light microscopic view of the frontal lobe after CSF release. The orange line indicates the site of OCT-scan. The black arrow demonstrates the site of CSF release - suprasylvian incision of the ABCM. Below OCT-scan of the region visualizing the collapsed SAS, with adjacent internal blood vessels. Red rectangle shows enlarged details of the OCT-Scan.

CSF, cerebrospinal fluid; OCT, optical coherence tomography; SAS, subarachnoid space.

OCT scanning showed that subarachnoid haemorrhage (SAH) could be depicted (Figure 4).

Figure 4.

Above light microscopic picture of the parietal lobe with SAH associated with the surgical procedure after CSF release. The orange line demonstrates the scanning site. Below OCT-Scan with depiction of perivascular SAS. The ABCM does not show a clear differentiation to the trabecular system, internal blood vessels and brain cortex. Red rectangle shows enlarged details of the OCT-Scan.

CSF, cerebrospinal fluid; OCT, optical coherence tomography; SAH, subarachnoid haemorrhage.

Orthogonal distance measurement of the SAS and ABCM

Orthogonal distance measuring was possible. The gyral SAS showed a mean depth of 570 µm frontotemporal, 463 µm frontal and 676 µm temporal. The sylvian SAS showed, as anatomically expected, a higher mean depth of 1204 µm. See Figure 5 for the measurement technique and Figure 6 for the values of the frontal and temporal SAS depth.

Figure 5.

Above light microscopic picture of the frontal lobe. The horizontal orange line demonstrates the scanning site. Below OCT-scan of corresponding region. The red rectangle shows enlarged details of the OCT-Scan and demonstrates the measurement technique. The vertical yellow line indicates the measurement site of the depth of the SAS and the ABCM.

ABCM, arachnoid barrier cell membrane; OCT, optical coherence tomography; SAS, subarachnoid space.

Figure 6.

Depth of the frontal and temporal SAS.

SAS, subarachnoid space.

The ABCM showed a mean depth of 81 µm frontotemporal, 46 µm frontal and 119 µm temporal (Figure 7). The ABCM at the sylvian fissure showed a higher depth (Figure 8).

Figure 7.

Depth of the frontal and temporal ABCM.

ABCM, arachnoid barrier cell membrane.

Figure 8.

Above light microscopic image of the sylvian fissure. The horizontal orange line demonstrates the scanning site. Below OCT-Scan demonstrating the higher depth of the ABCM at the sylvian fissure.

ABCM, arachnoid barrier cell membrane.

Discussion

Ultrasound based methods-based methods are the gold standard for in vivo dynamic sectional imaging of the central nervous system, but lack the spatial resolution to delineate the microstructural composition of the SAS.[2-4] Still imaging modalities are needed to display intact tissue and pathological alterations in the context of observational research and as intraoperative imaging tools. The present study reveals that microscope-integrated OCT might be capable of closing this neuroimaging gap. In addition, the technique showed to be a valuable tool for the first orthograde distance measurements of microanatomical structures of the SAS.

Albeit, in vivo conditions are constrained by tissue movements of several millimetres in amplitude which follow the respiratory and arterial cycle.[16] Microscope-integrated OCT offered a feasible technique to display the architecture of the intact SAS approaching the spatial resolution of histological analysis. Delineation of the ABCM, trabecula system, internal blood vessels and human brain cortex was robust, whereas differentiation of the pia mater was only possible in 26.9% of cases. Furthermore, SAH as a pathological condition could be clearly depicted. We demonstrated that SAH inhibited the delineation of trabeculae, ABCM, pia mater or minor internal blood vessels. The only prior study examining the SAS composition with OCT was carried out in CSF-released porcine brain ex vivo and reported a similar depiction of the ABCM, trabecular system, internal blood vessels and brain cortex, whereas differentiation of the pia mater was not possible. Post mortem conditions and CSF release with a collapsed SAS could here merely provide intragyral scanning.[19] Our data set demonstrated that under CSF-released conditions, in comparison with intact tissue, a clear differentiation between the ABCM and trabecular system was not possible; on the contrary ABCM depth could easily be overestimated by clinging trabeculae. CSF release furthermore, led to blood vessels which were adjacent to the human brain cortex. This resulted in pronounced optic shadows (Figure 3). Clinging as opposed to free trabeculae might minimize scattered radiation, enhancing optical density. In this respect, future experiments are necessary for clarification.

In animal studies, OCT recently proved to be a valuable and diverse neurophysiological imaging tool in domains like haemodynamic imaging,[24-27] functional imaging[28,29] and molecular imaging.[30] In experimental setups OCT could not only differentiate grey and white matter[12] but furthermore, healthy and tumor-invaded brain tissue.[16] These experimental results could be translated to intraoperative imaging in vivo.[17,18] Our mere delineation of the intact microanatomical architecture of the human SAS in vivo and the knowledge of the above mentioned optic effects are valuable state planning criteria for such conceivable future applications in clinical settings.

Microscope-integrated OCT offered the possibility for depth measuring of the human SAS in vivo. For the first time, depth measurements of these anatomical structures were possible. Technical accuracy was defined to 7.5 µm. Due to the small number of participants, conditions after craniotomy with dural opening and the presence of brain-related pathologies, our results need reevaluation to define matters like age or sex which might influence the depth of these structures. As anatomically expected, we could demonstrate higher SAS depths at the level of the Sylvian fissure in relation to the frontotemporal SAS. Unexpectedly, we could demonstrate higher depths of the ABCM at the Sylvian fissure in comparison with the frontotemporal ABCM depth.

In previous studies, OCT showed a remarkable axial spatial resolution.[18] Since it depends on light, image acquisition is rapid and versatile implementations via fibre optic catheters are simple.[31] In our study, we demonstrated that OCT offered a clear and robust delineation of the microanatomical composition of the intact human SAS and adjacent brain cortex. Further studies are necessary to evaluate the clinical relevance as an additional intraoperative tool during microsurgical procedures.