2026 年 13 巻 p. 583-589
Retained medullary cord is a form of closed spinal dysraphism proposed by Pang et al. It is thought to result from the failed regression of the medullary cord, giving rise to a non-functional cord-like structure. This cord-like structure is continuous with the functional cord and tethers it, thereby causing symptoms. We previously reported five patients with saccular retained medullary cord, in whom a cystically dilated dural cul-de-sac protruded outside the spinal canal. This protrusion occurred through sacral spina bifida at or below approximately the S2-3 level, where the dural cul-de-sac is located, and through the sacral hiatus, forming a sacral subcutaneous meningocele sac. In these cases, the retained medullary cord terminated at the apex of the sac. We report a patient in whom the lower half of the sac lacked skin coverage and contained a neural placode-like structure, necessitating differentiation from an open myelomeningocele. The infant had no neurological deficits. Surgery was performed on the second day after birth. A non-functional cord-like structure was resected at the cranial end of the surgical field, excised en bloc with the skin lesion, and cord untethering was performed. Histopathological examination revealed glial fibrillary acidic protein-immunopositive neuroglial tissue containing an ependymal lined central canal-like structure. The retained medullary cord tissue was directly exposed at the site where the skin was missing. Patients previously diagnosed with terminal apical myelomeningocele based solely on neuroradiological findings are likely to have saccular retained medullary cord. In such cases, prophylactic cord untethering is an important surgical consideration in addition to sac repair.
Although primary neurulation is well understood, many aspects of secondary neurulation remain unclear. Secondary neurulation is generally considered to proceed as follows:1-5) it begins with mesenchymal to epithelial transformation within a pluripotent blastema called the caudal cell mass or caudal eminence (Figure 1a). The process then progresses from condensation of the mesenchyme into a solid medullary cord (condensation phase) (Figure 1b) to the formation of an intra-chordal lumen, i.e., a central canal, which is continuous with the central canal of the primary neural tube (cavitation and canalization phases) (Figure 1c). This is followed by partial regression of the cavitary medullary cord, leaving only the conus (below S1 and S2) and filum (regression phase) (Figure 1d).
(a-d) Schematic drawings illustrating normal secondary neurulation, including mesenchymal to epithelial transformation (a), condensation phase (b), cavitation and canalization phases (c), and regression phase (d).
(e-g) Various pathologies of a retained medullary cord (RMC) likely result from failure of regression of the medullary cord. These include the original RMC described by Pang et al.1,2) (e), an RMC of saccular type (saccular RMC) (f, g), an RMC extending to a skin-covered subcutaneous meningocele (previously reported by us6-8)) (f), and an RMC extending to a skin-uncovered meningocele requiring differentiation from a myelomeningocele (this case) (g). Please refer to the text for details.
GFAP: glial fibrillary acidic protein
Retained medullary cord (RMC) is a type of closed spinal dysraphism proposed by Pang et al.1,2) in 2011. It is considered to result from regression failure of the cavitary medullary cord, producing a redundant, non-functional, cord-like structure (C-LS) that is continuous with the functional spinal cord, tethers it, and causes symptoms (Figure 1e).1,2) Histopathological examination of the C-LS resected during surgery typically demonstrates an ependymal lined central canal-like structure (ELCC-LS) with surrounding glial fibrillary acidic protein (GFAP)-immunopositive neuroglial tissue (NGT) within fibro-collagenous tissues, corroborating its interpretation as a remnant of the cavitary medullary cord.1,2,4,5) Prophylactic untethering of the cord is recommended for RMC before symptom onset. The caudal end of this RMC usually terminates in the dural cul-de-sac.1,2,4,5) However, we previously reported five patients with RMC of saccular type (saccular RMC), in whom a cystically dilated dural cul-de-sac protruded outside the spinal canal to form a subcutaneous meningocele sac in the sacral region, with the RMC terminating at its apex (Figure 1f).6-8) Here, we report an additional patient in whom this meningocele sac was not covered by skin, requiring differentiation from open myelomeningocele (MMC), including terminal apical MMC (Figure 1g).9)
The baby's mother was a 29-year-old primigravida with an uncomplicated prenatal course. Labor commenced at 40+5 weeks of gestation but stalled due to weak contractions, necessitating an emergency cesarean section. The baby boy was born weighing 3,564 g, measuring 52 cm, with Apgar scores of 9 and 10. A dorsal midline sac resembling an MMC was observed in the sacral region, leading to his transfer to our hospital on the day of birth. Upon admission, physical examination of the sac revealed that its upper portion was skin-covered (Figure 2a), whereas the lower portion had a skin defect through which neural placode-like structures were visible. No leakage of cerebrospinal fluid (CSF) was observed, and the patient exhibited no neurological deficits.
(a) Photograph at birth showing a myelomeningocele-like dorsal midline sac in the sacral region. The lower part is not covered with skin, and a neural placode-like structure is noted. (b) Preoperative T1-weighted (slice thickness, 2 mm) and (c) three-dimensional heavily T2-weighted images (slice thickness, 0.97 mm) reveal a cystically dilated dural cul-de-sac, protruding outside the spinal canal and forming a meningocele sac in the sacral region, with a cord-like structure (C-LS) terminating at its apex. (d, e) Schematic drawing and (f) microscopic view of the intraoperative findings. Please refer to the text for details. (g) T2-weighted image (slice thickness, 2 mm) 2 weeks post-operation confirming successful surgery, including sac excision and cord untethering.
FCT: fibro-connective tissue
Magnetic resonance imaging (MRI), including a three-dimensional heavily T2-weighted image (3D-hT2WI) performed on the first day postpartum, showed a cystically dilated dural cul-de-sac protruding outside the spinal canal through a spina bifida below S3 and the sacral hiatus (Figure 2b and c, white arrows), forming a meningocele sac. The C-LS, continuous with the spinal cord, terminated at the apex of the sac (Figure 2b and c, blue arrows) and appeared to spread at its base (Figure 2c, red arrow). T1-weighted images demonstrated high signal intensity continuous with subcutaneous fat in the wall of the upper sac (Figure 2b, yellow arrows), which was absent in the lower part. No fatty acid signals were detected in the C-LS. There was no evidence of hydrocephalus or Chiari malformations.
On day two, cord untethering and sac repair surgery were performed. A skin incision was made from the midline of the upper sac, extending to the periphery of the uncovered skin lesion in the lower part (Figure 2d). After CSF drainage, the interior of the sac was inspected. The C-LS extended from the spinal canal outlet toward the skin-uncovered lesion and broadened at its base (Figure 2e and f). During intraoperative neurophysiological monitoring (IONM), electrical stimulation of the exposed C-LS, up to a maximum of 5 mA, did not induce compound muscle action potentials (CMAPs) in the external anal sphincter muscle, hamstring, or gastrocnemius. In addition to blood vessels, tissues presumed to be the vestigial nerve roots were observed on the surface of the C-LS (Figure 2f); however, electrical stimulation of this area failed to induce CMAPs. The C-LS was transected as close as possible to the spinal canal outlet, releasing the tethering, and the cranial stump was retracted toward the spinal canal. The severed caudal C-LS was excised en bloc with the sac. The spinal canal outlet was closed using a pedunculated flap created from the inner wall of the sac, forming a dural cul-de-sac. The dead space remaining after the sac resection was filled with a pedunculated subcutaneous fat flap, and the skin was closed in two layers: subcutaneous tissue and skin.
No de novo neurological symptoms developed postoperatively. MRI performed 2 weeks later confirmed the successful untethering of the cord (Figure 2g). Histopathological examination of the resected C-LS revealed that GFAP-immunopositive NGT surrounding the ELCC-LS occupied most of the cross-section (Figure 3a and b). Peripheral nerves, presumably a continuation of the vestigial nerve roots, were also observed around the C-LS (Figure 3a). Stratified squamous epithelium is not observed at the resection margins. In the skin-covered (Figure 3c, blue arrows) and skin-uncovered (Figure 3c, red arrows) sac walls, GFAP-immunopositive NGT–containing ELCC-LS was present, extending from the stalk of the RMC (Figure 3c and d). When the sac surface was not covered by skin tissue, the GFAP-immunopositive NGT–containing ELCC-LS was covered by fibro-collagenous tissues (Figure 3e-h); however, in some areas, it was directly exposed on the surface (Figure 3i and j).
(a, b) Photomicrographs of a cross-section of the cord-like structure (C-LS), with immunostaining for glial fibrillary acidic protein (GFAP) (a) and hematoxylin and eosin (H&E) staining (b). They show an ependymal lined central canal-like structure (ELCC-LS) surrounded by GFAP-immunopositive neuroglial tissue (NGT) within fibro-collagenous tissues (FCT). Peripheral nerves (PN) are noted around the C-LS.
(c, d) Vertical views of the sac, with immunostaining for GFAP (c) and H&E staining (d). The images reveal GFAP-immunopositive NGT, including ELCC-LS, extending from the C-LS to the skin-uncovered wall (red arrows). Blue arrows indicate the skin-covered wall. (e-j) In vertical sections of the skin-uncovered sac (red arrows) with H&E staining (e, f, and j) and GFAP immunostaining (g-i), the GFAP-immunopositive NGT containing ELCC-LS is covered by fibro-connective tissue (FCT) (e, f), but in some areas, it is directly exposed on the surface (i, j). The locations of the respective surgical specimens are shown in Figure 2d, e, and f. Higher-magnification views correspond to the areas indicated by dashed squares in the lower-magnification views.
Further careful follow-up, including monitoring for retethering, is required.
Although the diagnostic criteria for RMC have not been fully established, the following three criteria are generally considered important.1,2,4,5,10) First, morphological features of RMC associated with cord tethering are demonstrated on neuroimaging and in intraoperative findings.1,2,4,5,10) Second, detection of the exact border between the functional cord and the non-functional RMC with IONM is desirable during cord untethering. This is because non-functional RMC is indistinguishable from the functional cord on neuroimaging and in intraoperative findings.1,2,4,5,8,10-16) Third, histopathological examination of the resected C-LS typically shows an ELCC-LS with NGT.1,2,4,5,8,10-22) This case clearly fulfilled the first and third criteria. Regarding the second criterion, we did not accurately identify the electrophysiological border. However, the exposed C-LS was proven to be non-functional with IONM, leading to a diagnosis of RMC.
The mechanism by which RMC, originally a closed spinal dysraphism, presented as open spinal dysraphism in this patient remains speculative. Histopathological examination of the sac showed that RMC tissue was present within the subcutaneous tissue, extending from the RMC stalk. In areas where skin was absent, the RMC tissue was directly exposed on the surface. Kim et al.23) reported enlargement of extraspinal cysts associated with various types of spinal dysraphism due to CSF hydrodynamics. We also reported a patient with a saccular RMC in whom a subcutaneous meningocele sac enlarged over 3 months, between 8 and 11 months of age, presumably due to similar CSF hydrodynamics.7) Therefore, we speculate that, in this patient, the subcutaneous meningocele sac enlarged during the fetal period and that the skin over the lower part of the sac was deficient, directly exposing the RMC tissue and presenting a neural placode-like appearance.
The principal differential diagnosis of saccular RMC is MMC. Although RMC can be expected to have a good functional prognosis with appropriate cord untethering, MMC cannot be expected to have a favorable prognosis, even with cord untethering, including myeloplasty.24) Tortori-Donati et al.9) described cases with MRI morphology identical to that in this case as "terminal apical MMC." However, this neuroradiological diagnosis was made before the clinicopathological entity of RMC was proposed and lacked confirmation by intraoperative findings, including IONM, and histopathological examination.6)
We posit that the following features are important for differentiating saccular RMC from MMC (Table 1). First, in saccular RMC, the dural cul-de-sac is cystically dilated to form a meningocele sac. The sac usually protrudes outside the spinal canal through sacral spina bifida at or below approximately the S2-3 level, where the dural cul-de-sac is located, and through the sacral hiatus (Figure 1g). Conversely, MMC is caused by primary neurulation failure and is therefore located more cranially.25,26) Second, based on the intraoperative findings of five previous patients with saccular RMC, we believe that when an RMC terminates at the meningocele sac, its base broadens.6-8) In MMC, by contrast, the spinal cord transitions directly into a neural placode located on the surface of the sac, and nerve roots extending ventrally from this neural placode are frequently observed.25,26) To detect this subtle finding preoperatively, 3D-hT2WI with a thin slice thickness of approximately 1 mm, as shown in Figure 2c, is useful.25,26) Third, unlike patients with MMC, patients with RMC often have no neurological abnormalities.1,2) Fourth, RMC is rarely associated with Chiari malformation or hydrocephalus,1,2) which are commonly seen in MMC.24,27)
Differentiation between Saccular Retained Medullary Cord and Myelomenigocele
| Saccular retained medullary cord | Myelomenigocele | |
|---|---|---|
| Pathoembryology | Probably regression failure of medullary cord during secondary neurulation with cystic dilatation of dural cul-de-sac | Dysraphism during primary neurulation |
| Level of protruded sac | Spina bifida below S2 and sacral hiatus | Spina bifida at mainly lumbosacral region |
| Morphological findings | Cord-like structure ending at the sac apex | Cord being continuous to neural placode at the sac |
| Neurological deficit | Rare | Frequent |
| Complication of Chiari malformation and hydrocephalus | Rare | Frequent |
Another condition to consider in the differential diagnosis of a sacral mass, as in this case, is terminal myelocystocele (TMC). TMC is an unusual type of closed spinal dysraphism, thought to originate from the failed regression spectrum during secondary neurulation, similar to RMC, and is characterized by localized cystic dilatation of the terminal part of the central canal that then herniates through the spinal dysraphism in the sacral region.28,29) In this case, we believe TMC can be excluded because the central canal was not involved in the lesion.
Regarding the timing of surgery for RMC, because RMC is originally a closed spinal dysraphism, elective surgery is performed after the baby has grown, similar to that for spinal lipoma.1,2) However, in saccular RMC, early surgery is often performed immediately after birth, even when the sac surface is covered with skin, when the sac is too large to handle safely, or if there is a risk of rupture.6-8) In this case, in addition to these findings, skin defect and risk of infection were noted; therefore, early surgery within 48 h after birth was indicated, consistent with the timing of surgery for MMC.24)
During surgery for saccular RMC, reliable cord untethering is crucial in addition to sac repair.6-8) Untethering is recommended to be performed at the boundary between the functional and non-functional cords identified by IONM.1,2,13) However, identifying this boundary may require additional laminotomy or laminectomy, as in this patient, and this procedure may be invasive when surgery is performed immediately after birth.6-8) Therefore, in this patient, when the exposed C-LS was confirmed to be non-functional with IONM, we considered it sufficient to perform untethering as far cranially as possible within the exposed surgical field, without additional laminotomy as a general rule.6-8)
When no response is elicited by IONM, as in this case, it is not always possible to determine whether the absence of a response reflects truly non-functional tissue or results from technical factors or the limited testing sensitivity. Based on our experience with surgery for spina bifida resulting from secondary neurulation failure, including RMC cases, stimulation at levels up to 5 mA is sufficient for this assessment.6-8,10-12,15,20,22) In such cases, immediate verification of the device output is important and can be performed by stimulating a peripheral nerve accessible percutaneously outside the surgical field, such as the median nerve at the wrist.
However, in RMC, a dermal component, i.e., stratified squamous epithelium, may rarely migrate from the skin lesion into the stalk of the RMC. When this tissue remains cranial to the resection margin, it may subsequently form a dermoid within the spinal canal.8,11) Therefore, it is important to confirm the absence of dermal components by careful histopathological examination of the resection margin, and, when they are present, additional surgery should be indicated.8) This issue is analogous to whether an additional laminotomy should be performed during neonatal surgery to trace the stalk of limited dorsal myeloschisis to the site of cord attachment.30)
In conclusion, cases previously diagnosed as terminal apical MMC based only on neuroradiological findings are likely to be saccular RMC. During surgery, reliable cord untethering is important in addition to sac repair.
Author Koji Yoshimoto is one of the Editorial Board members of the Journal. This author was not involved in the paper-review or decision-making process for this paper.
We would like to thank Editage for the English language editing.
Morioka T, Inoue D, Kanata A, and Amano T contributed to the study design and project outline. Morioka T, Inoue D, Kanata A, Amano T, and Suzuki SO contributed to data acquisition, analysis, and manuscript drafting. Suzuki SO, Shimogawa T, Mukae N, and Yoshimoto K critically revised the manuscript.
All authors have no conflict of interest.
Written informed consent was obtained from the patient's parents for the publication of this case report.
