ORIGINAL ARTICLE

What is the dorsal median sulcus of the spinal cord? Interestfor surgical approach of intramedullary tumors

Timothee Jacquesson • Nathalie Streichenberger •

Marc Sindou • Patrick Mertens • Emile Simon

Received: 6 March 2013 / Accepted: 19 August 2013 / Published online: 31 August 2013

� European Union 2013

Abstract

Purpose For intramedullary tumor (IMT) surgery, a bal-

ance has to be found between aggressively resecting the

tumor and respecting all the sensory and motor pathways.

The most common surgical approach is through the dorsal

median sulcus (DMS) of the spinal cord. However, the

precise organization of the meningeal sheats in the DMS

remains obscure in the otherwise well-described anatomy

of the spinal cord. A better understanding of this archi-

tecture may be of benefit to IMT surgeon to spare the spinal

cord.

Methods Three spinal cords were studied. The organiza-

tion of the spinal cord meninges in the DMS was described

via macroscopic, microsurgical and optical microscopic

views. A micro dissection of the DMS was also performed.

Results No macroscopic morphological abnormalities

were observed. With the operative magnifying lens, the

dura was opened, the arachnoid was removed and the pia

mater was cut to access the DMS. The histological study

showed that the DMS was composed of a thin rim of

capillary-carrying connective tissue extending from the pia

mater and covering the entire DMS. There was no true

space between the dorsal columns, no arachnoid or cross-

ing axons either.

Conclusion Our work indicates that the DMS is not a

sulcus but a thin blade of collagen extending from the pia

mater. Its location is given by tiny vessels coming from the

surface towards the deep. Thus, the surgical corridor has to

follow the DMS as closely as possible to prevent damage to

the spinal cord during midline IMT removal.

Keywords Spinal cord � Meninges � Dorsal media

sulcus � Intramedullary tumors

Introduction

The earliest example of spinal cord surgery may date to

1,549, when the French army surgeon Ambroise Pare

excised bony splinters compressing a spinal cord and

applied traction for spinal dislocations. Later, in 1887,

Horsley described the first removal of a spinal meningioma

[7]. Then, at the beginning of the twentieth century,

Cushing reported the first myelotomy to expose an inop-

erable intramedullary tumor (IMT) [3], and this was

quickly followed by the first cases of IMT removal [4, 10].

Today, IMT surgery, although still delicate, is performed

daily in neurosurgical centers. Excepting lateral or exo-

phitic tumors, the surgical approach of IMT is generally

done via the dorsal median sulcus (DMS), situated between

the two dorsal columns [5, 6]. IMT surgery often neces-

sitates a careful balance between aggressive resection of

the complete tumor and the respect of sensory-motor

pathways.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s00276-013-1194-1) contains supplementarymaterial, which is available to authorized users.

T. Jacquesson (&) � M. Sindou � P. Mertens � E. Simon

Department of Neurosurgery, Hopital Neurologique Pierre

Wertheimer, 59 Bd Pinel, 69677 Bron, Cedex, France

e-mail: timothee.jacquesson@neurochirurgie.fr

T. Jacquesson � P. Mertens � E. Simon

Department of Anatomy, University of Lyon1,

8 Avenue Rockefeller, 69003 Lyon, France

N. Streichenberger

Department of Pathology, Groupement Hospitalier Est,

59 Bd Pinel, 69677 Bron, Cedex, France

123

Surg Radiol Anat (2014) 36:345–351

DOI 10.1007/s00276-013-1194-1

Human spinal cord anatomy has long been well estab-

lished [2, 8, 21, 24]. The spinal meninges are quite similar

to those of the brain. Indeed, they include three main layers

enclosing the spinal central nervous system from the

foramen magnum to the sacrum. The outermost layer is the

dura mater, a single cylindrical sleeve enclosing the spinal

cord and its roots all along the spinal canal. Next is the

outer parietal arachnoid mater, which projects a series of

highly perforated intermediate sheets that envelope nerve

roots and vessels. These sheets are impermeable to cere-

brospinal fluid (CSF) and form the subarachnoid space.

The innermost layer is the pia mater, which covers the

entire spinal cord surface and supports arteries and nerves.

The term leptomeninges refers to the arachnoid and the pia

mater together, but excludes the dura mater [16, 23].

Despite the fact that spinal cord anatomy is generally

well known, and that IMT surgery has existed for quite a

long time, the precise architecture of the leptomeninges in

the DMS is currently poorly understood. The brain men-

ingeal barrier has been studied to understand blood-CSF

exchanges or tumor growth [13]. Furthermore, the ultra-

structure of cranial or spinal meninges has also been

recently studied, but the DMS or the anterior median sulcus

(AMS) has never been described at a histological scale [14,

15, 22, 23].

The DMS has been described differently in different

ways by different authors: a partition made of pia mater

[21]; a small intermediate sulcus that may be detectable on

the posterior surface [17]; a distinct median raphe, over

which the very tortuous posterior spinal vein runs [1]; or a

thin fibrous structure that dorsally separates the spinal cord

in two halves [19]. Also, the existence of a true deep sulcus

with meninges or a virtual space crossed by axons between

the dorsal columns is not formally established. In the

present study, we thus specifically studied the meninges in

the DMS to clarify knowledge of this anatomical feature

and to provide useful information for IMT removal.

Materials and methods

Samples

Three fresh cadavers were used in the Laboratory of

Anatomy of Lyon. One cervical spinal cord was used for

preliminary tests. Red latex was injected in the arteries of

another cervical spinal cord to provide a more complete

macroscopic view. Finally, an entire spinal cord was

sampled. The delay after death was inferior to 30 days and

formalin preparation was done on all bodies. To access the

spinal cord, the bodies were placed in a prone position and

posterior midline skin incisions from the occipital protu-

berance to the sacrum were done. Then, the paravertebral

muscles were retracted to expose the posterior arches and

to allow complete wide laminectomy. After sharp section

of the dural sac from C1 to S1 levels, the meninges were

carefully released from the posterior longitudinal ligament.

Then, the spinal cord with all the meninges and roots were

removed ‘‘en bloc’’. The samples were placed in a 4 %

formaldehyde bath for 15 days at room temperature to

obtain additional formalin fixation.

Macroscopic and operative study

First, orthogonal transversal sections were performed with

ultra fine razor blades at the cervical, thoracic and lumbar

levels, corresponding to the C2, C6, T4, T8 and L1 nerve

root exits. Using micro-surgical instruments, the dura

mater and parietal arachnoid layer were opened and fixed

with pins on a rigid plate. Then, under a 109 magnification

operating microscope, the microsurgical dissection of the

meninges and the DMS was performed, using scissors,

tweezers and spatulas to carefully separate the anatomic

structures. Secondly, another transversal section was per-

formed to precisely include the DMS opening site.

Histological study

Eight mm thick samples were included in a 30 9 25 9

10 mm box for a 4 % formaldehyde bath at room tem-

perature. After paraffin embedding and microtome sec-

tioning 10 microns thin, the samples were spread on a slide

and stained with Luxol Fast Blue—PAS. Analyses were

then performed by the pathologist using an optical micro-

scope at high magnification (209, 409 and 1009).

Latex injection

One cadaver head was cut at the C7 level, irrigated with

warm water to remove blood and fixed with formalin via

the neck vessels. Red latex was then injected in the internal

carotid and vertebral arteries (Aerographe Magenta acrylic

Fig. 1 Macroscopic view of the C7 spinal cord. Note the butterfly-

shaped gray matter

346 Surg Radiol Anat (2014) 36:345–351

123

ink pb15.2 Pebeo). After formalin fixation, the cervical

spinal cord was removed, thus providing a better view of

arterial supply.

Results

Macroscopic views

No macroscopic morphological abnormalities such as root

duplications, tumors or vascular malformations were

observed. Comfortable access to the posterior face of the

spinal cord was attained by pinning the dura open. Rootlets

and roots merged regularly on both sides. The spinal cord

measured 1 cm diameter in average and presented the

typical cervical and lumbar enlargements. The roots of the

cauda equina emerged normally from the terminal conus.

The denticulate ligament was stretched between the lateral

columns and the dura mater. The arachnoid mater covered

the entire spinal cord and hid the pia mater and vessels. The

tortuous posterior spinal vein was discernable along the

posterior midline surface and two dorsolateral arteries sent

Fig. 2 Microsurgical views. a Dissection of the arachnoid and the pia

maters using microsurgical scissors. b Presence of fibrous septa

between dorsal columns. c Separation of the posterior columns to

open the DMS. d Presence of multiple tiny vessels coming from the

surface and going deeply along the DMS. They could lead the surgeon

to find the DMS

Surg Radiol Anat (2014) 36:345–351 347

123

small perforating arteries laterally and toward the midline.

The gray matter, with its characteristic butterfly shape, was

visible at the inferior end of the sample, surrounded by the

white matter (Fig. 1).

Operative views

Using the operating microscope, the arachnoid bridging

tissues were cut to release the dura mater from the pia

mater. The deep arachnoid layer covering the pia mater

was then gently removed to preserve the underlying

structures. These later comprised a large number of very

small vessels covering the surface of the spinal cord. After

switching the spinal vein, a small trench, more or less

visible according to the location, was exposed exactly in

the middle of the posterior face of the spinal cord. Formalin

fixation maintains tissue consistency very close to those

in vivo. The pia mater, which carries all the vessels, was

then cut, thus allowing the spontaneous separation of the

two dorsal columns (Fig. 2a).

If it is not clearly seen, the pia mater may be incised

longitudinally and medially using the emergence of the

posterior roots, the two dorsolateral arteries or the bulging

of the dorsal columns as reference points. In our samples,

no true septum could be identified in the DMS, but there

were many small fibrous bridges connecting the two dorsal

columns (Fig. 2b). The dissection was conducted smoothly

to spare the white matter laterally (Fig. 2c). A deep vas-

cularized area, the ependyma, was ultimately reached.

Sulco commissural vessels from the anterior spinal artery

at the bottom of this last area were also visible. Tiny

arteries or veins on the walls of the DMS plunged deeply

coming from the surface (Fig. 2d). This meningeal archi-

tecture in the DMS was consistent throughout the spinal

cord, but less evident at the thoracic level. Indeed, DMS

dissection seemed easier at the cervical and lumbar

enlargements, where we found more landmarks or vessels.

Histological study

Via histological examination with an optical microscope,

we found that the DMS presented as a single layer between

the dorsal columns (Fig. 3). In the AMS by contrast, there

was a real open space surrounded by pia and arachnoid

maters. Under higher magnification, we identified a thin

rim of conjunctive tissue that covered the columns

throughout the DMS, but did not cover the ependyma. No

arachnoid or subarachnoid space was found. In our sam-

ples, the DMS appeared to be an extension of the pia mater

comprising a narrow layer of collagen and small capillaries

(Fig. 4). Laterally, the white matter contained axons for

sensory pathways but none crossed the DMS. There were

no neuronal cell bodies, but oligodendrocytes and

astrocytes were present. There were also degenerative,

amylaceous bodies (stained red). The ‘‘dorsal intermedio-

lateral sulci’’ appeared as empty spaces also covered by pia

like the DMS.

The surgical pathway followed the DMS, but some

white fibers were damaged and, in several samples, the

dissection exceeded the ependyma, reaching the AMS.

The histological architecture of the DMS was preserved,

albeit with slight variations, along the entire spinal cord.

At the cervical and lumbosacral levels, we noted that in

some samples, a layer of pia mater entered partially into

the DMS. At the thoracic level, the DMS was less rec-

ognizable and there were several dorsolateral sulci with

septa formed from the pia mater as in the DMS (Fig. 5).

The septa at this last level could not be followed

throughout the DMS, which suggests that they were

discontinuous.

Discussion

The history of spinal meninges studies starts at end of the

nineteenth century with the works of Key and Retzius [9].

Different authors have referred to the DMS as ‘‘the septum

arachnoidale’’, ‘‘septum intermedium’’, or ‘‘septum lep-

tomeningeum dorsale’’. Samii described the DMS as a

septum along the posterior midline, attached to the pia

mater of the spinal cord. For the author, the DMS runs from

the cervical to the lower thoracic region [19]. Weller

described a substantial layer of subpial collagen separating

the pia mater from the glia limitans on the surface of the

spinal cord. He went on to say that this layer of subpial

collagen attaches to the dura via the dentate ligament and

Fig. 3 Optical Microscopic view of C6 spinal cord axial section.

Zoom 920. An example of spinal cord sample as it was embedded

with its dura sleeve, rootlets and arachnoid sheets. (DMS dorso medial

sulcus, Ar arachnoid, DM dura mater, R rootlet)

348 Surg Radiol Anat (2014) 36:345–351

123

may play a role in stabilizing the cord within the sub-

arachnoid CSF, in contrast to the brain [23]. Brotchi

described a sulcus that can sometimes be identified only on

the basis of the convergence of vessels toward the midline.

Surgically, he stated that these variously sized vessels

running vertically over the dorsal columns must be dis-

sected and mobilized laterally to expose the posterior sul-

cus, with attempts to spare all of the thinnest arterial or

venous vessels in the sulcocommissural region. The dorsal

columns are then carefully retracted and opened progres-

sively over the entire length of the solid portion of the

tumor, ‘‘as if they were pages of a book’’ [1].

The other main spinal sulcus, the AMS has been

described as a real sulcus lined by the pia mater until its

bottom and which separates cortico-spinal tracts from each

side [21]. As arachnoid sheets were never found into the

DMS or the AMS, it explains that spinal meningiomas are

never growing from these ‘‘sulcus’’ in the few reported

series [18, 20]. Moreover, the vascularization of the spinal

cord has been already precisely explored. Indeed, the two

dorsal spinal arteries provide dorsal medial arteries which

flow into the DMS and send branches in proprioceptive

columns of Goll until the gray commissure [12]. An

important dorso-median vein follows the path of the DMS

Fig. 4 Optical microscopic views of the C6 spinal cord DMS.

Zoom 940 (a), 9100 (b), 9200 (c), 9400 (d). The DMS appears as

an extension of the pia mater consisting of a thin layer of capillary-

carrying collagen. (SDM dorso medial sulcus, PM pia mater, Ar

arachnoid, V vessel, AB amylaceous body)

Surg Radiol Anat (2014) 36:345–351 349

123

but appears often sinuous or discontinuous. So, this dorsal

vein is not a fiable landmark of the dorsal midline. Finally,

because of the somatotopic organization inside the dorsal

column described by Kahler, the sensory fibers coming

from the lumbosacral area are the more medial fibers in the

funiculus. Also, they are preferentially exposed to vessel

damages during the surgical approach through the DMS.

Thus, this vascular lesion could be responsible for the

lumbosacral proprioceptive symptoms.

In the present work, we shed new light on the anatomy

of the spinal meninges and provide potentially useful

information for preoperative planning in IMT removal. Our

macroscopic, operative and histological study was

designed to mimic the point of view of the surgeon and the

pathologist. We performed micro-dissection of the DMS

with visual and tactile information close to that encoun-

tered in the most frequent approach to IMT surgery.

Indeed, our findings suggest that the surgical path has to be

precisely through the DMS; the gentle spacing of the dorsal

columns is vital to sparing the white matter and nerve

functions. Our assessment of the entire spinal cord con-

firmed that the architecture of the DMS is quite similar at

the cervical and lumbosacral levels. Some authors report

that the anterior spinal artery is not continuous and there

exists a critical narrow zone for the vascularization in the

midthoracic region [11]. We noticed that, at this last level,

the organization of the meninges was less homogeneous.

This may be due to poorer blood supply in the DMS. In

contrast, the cervical and lumbosacral levels had the same

meninges anatomy. In a few samples, we found that the pia

mater even fed the DMS superficially.

Cut artifacts can affect the analysis of histological

samples. To best manage this issue, we used ultra fine razor

blades to achieve the most orthogonal cutting plane as

possible. Luxol Fast Blue was used in our study to differ-

entiate white matter, gray matter, leptomeninges or blood

vessels. We also tried staining with silver or periodic acid-

Schiff but these provided no additional information.

Fig. 5 Optical microscopic views. a, b A small piece of pia mater

extending into the DMS in D12 spinal cord (orange narrow);

Zoom 920 and 9100. c The DMS is not homogeneous. The ‘‘dorsal

intermediolateral sulci’’ between dorsal columns have ‘‘DMS like’’

structures in D8 spinal cord (orange narrow); Zoom 940. d C6 spinal

cord after DMS dissection showing that the adjacent white fibers of

the dorsal columns can be endangered by surgical opening of the

DMS (orange narrow); Zoom 912 (color figure online)

350 Surg Radiol Anat (2014) 36:345–351

123

Obviously, the conditions encountered during surgery will

be more complex than those of our research setting. Indeed,

the anatomical changes induced by the IMT itself may

make its surgical removal more complex. Micro-surgical

dissection is furthermore surgeon-dependent; the ability to

save white fibers grows with the surgeon’s experience.

Finally, we do wish to underline that our study was per-

formed on only a few samples and thus the significance of

our conclusions is limited. Our protocol will need to be

repeated in larger studies to confirm the present findings.

Conclusions

Unlike the variable organization of the meninges in the

DMS described in the literature, our work indicates that the

DMS is not a sulcus but a septum or a raphe which appears

as a thin blade of capillary-carrying collagen that extends

from the pia mater, tightly separates the dorsal columns

and goes deeply to—but excludes—the ependyma. Fur-

thermore, we found no arachnoid, no true space between

the dorsal columns and no crossing axons. The tiny vessels

coming from the surface has to be considered as Ariane’s

threads finding the way for the surgeon to the ‘‘dorsal

median septum’’. Thus, the surgical corridor in midline

IMT removal has to follow the DMS as closely as possible

to prevent damage to the spinal cord.

Acknowledgments We thank Prof Sindou who was at the origin of

this work. We thank the technical staff of the pathology laboratory for

their preparation of samples prior to histological analyses. We thank

K. Erwin for proofreading the English article.

Conflict of interest The authors declare that they have no conflict

of interest.

References

1. Brotchi J (2002) Intrinsic spinal cord tumor resection. Neuro-

surgery 50:1059–1063

2. Cajal SR (1894) Les nouvelles idees sur la structure du systeme

nerveux chez l’homme et chez les vertebres. Paris

3. Cushing H (1905) The special field of neurosurgery. Bulletin of

the Johns Hopkins Hospital, Baltimore, pp 77–87

4. Elsberg CA (1912) Surgery of intramedullary affections of the

spinal cord: anatomical basis and technique. JAMA 59:1532–

1536

5. Fischer G, Brotchi J (1994) Les tumeurs intramedullaires. Societe

de Neuro-Chirurgie de Langue Francaise, 45eme Congres Ann-

uel. Angers, 12-15 juin 1994. Neurochirurgie 40(Suppl 1):1–108

6. Fischer G, Mansuy L (1980) Total removal of intramedullary

ependymomas: follow-up study of 16 cases. Surg Neurol 14:243–249

7. Gowers R, Horsley V (1888) A case of tumour of the spinal cord.

Removal; recovery. Med Chir Trans 53:377–428

8. Hirschfeld L (1865) Traite et iconographie du systeme nerveux et

des organes des sens de l’homme. Paris Masson

9. Key EAH, Retzius MG (1876) Studien der Anatomie des

Nervensystems und des Bindegewebes. Stockholm Samson

Wallin, Stockholm

10. Krause F (1908) Erfahrungen bei 26 operativen Fallen von

Ruckenmarkstumoren mit Projektionen. Dtsch Z Nervenheilkd

36:106–113

11. Lazorthes (1973) Vascularisation et circulation de la moelle

epiniere. Paris Masson

12. Lazorthes G, Gouaze A, Zadeh JO, Santini JJ, Lazorthes Y,

Burdin P (1971) Arterial vascularization of the spinal cord.

Recent studies of the anastomotic substitution pathways. J Neu-

rosurg 35:253–262

13. Lescanne E, Velut S, Lefrancq T, Destrieux C (2002) The internal

acoustic meatus and its meningeal layers: a microanatomical

study. J Neurosurg 97:1191–1197

14. Mertens P, Guenot M, Hermier M, Jouvet A, Tournut P, Froment

JL, Sindou M, Carret JP (2000) Radiologic anatomy of the spinal

dorsal horn at the cervical level (anatomic-MRI correlations).

Surg Radiol Anat 22:81–88

15. Nauta HJ, Dolan E, Yasargil MG (1983) Microsurgical anatomy

of spinal subarachnoid space. Surg Neurol 19:431–437

16. Nicholas DS, Weller RO (1988) The fine anatomy of the human

spinal meninges. A light and scanning electron microscopy study.

J Neurosurg 69:276–282

17. Nieuwenhuys R (1978) The human central nervous system.

A Synopsis and Atlas. Springer, Heidelberg

18. Roux FX, Nataf F, Pinaudeau M, Borne G, Devaux B, Meder JF

(1996) Intraspinal meningiomas: review of 54 cases with dis-

cussion of poor prognosis factors and modern therapeutic man-

agement. Surg Neurol 46:458–463

19. Samii M, Keklamp J (2007) Surgery of spinal tumors. Springer,

Baltimore, p 526

20. Sandalcioglu IE, Hunold A, Muller O, Bassiouni H, Stolke D,

Asgari S (2008) Spinal meningiomas: critical review of 131

surgically treated patients. Eur Spine J 17:1035–1041

21. Testut L (1911) Traite d’Anatomie Humaine. Systeme nerveux

central. Octove Doin, Paris

22. Vandenabeele F, Creemers J, Lambrichts I (1996) Ultrastructure

of the human spinal arachnoid mater and dura mater. J Anat

189(Pt 2):417–430

23. Weller RO (2005) Microscopic morphology and histology of the

human meninges. Morphologie 89:22–34

24. Dejerine J (1980) Anatomie des centres nerveux. Paris Masson

Surg Radiol Anat (2014) 36:345–351 351

123