Image characteristics of retained medullary cord in secondary neurulation arrest: an observational study Authors: Seul Bi Lee, Seunghyun Lee, Yeon Jin Cho, Young Hun Choi, Jung-Eun Cheon, Kyung Hyun Kim, Ji Yeoun Lee Published: 2024 ABSTRACT Retained medullary cord (RMC) is a recently defined term denoting closed spinal dysraphism arising from the failure of regression in secondary neurulation. Despite the acknowledgment of this condition, there needs to be more literature elucidating the radiologic manifestations of RMC. This study aimed to describe the MR imaging findings of RMC. A retrospective analysis was conducted on spinal MRI scans of pediatric patients with surgically confirmed RMC, from January 2014 to August 2023. The investigation focused on evaluating the following image characteristics: level and morphology of the cord-like structure (C-LS), the signal intensity in the far distal C-LS, nerve root-like structures originating from the C-LS, intradural fatty lesions, and arachnoid cysts. This study included 38 patients (19 girls, mean age 7.3 months) who showed a low-lying cord in all cases. The morphology of C-LS was either smooth tapered (50%) or hourglass with fusiform cystic dilatation (50%). The C-LS exhibited aberrant T2 hypointense signal compared to the juxta-proximal level. T2 hypointense signal vestigial nerves emanating from the C-LS and intradural fatty masses were observed (89.5%). Sacral arachnoid cysts in extradural location were identified in eight patients. The characteristic MR features of RMC revealed an extremely low-lying distal C-LS with smooth tapering or hourglass-shaped cystic dilatation of the caudal part, accompanied by an intradural fatty stalk and the aberrant signal vestigial nerve. This study suggests those radiologic findings can be RMC instead of the previously called terminal syrinx with low-lying conus. Introduction The secondary neurulation process forms the distal spinal cord through the cavitation and differentiation of the caudal eminence, a mass of mesenchymal cells1,2. The impaired secondary neurulation process sometimes results in closed spinal dysraphism, such as myelocystocele, caudal- or transitional-type spinal lipoma, thickened filum terminale, and a retained medullary cord (RMC)3–5. RMC is a recently defined term, an elongated cord-like structure (C-LS) with non-function that begins at the distal spinal cord and continues toward the cul-de-sac, previously called “low-lying conus” or “terminal syrinx with low-lying conus”3,6. It has been shown that non-functional and functional cords cannot be distinguished with the naked eye and MRI. In addition, it is often preoperatively indistinguishable from other regression failures, such as lipomyelomeningocele, terminal myelocystocele, or thickened filum, due to sharing histopathological characteristics, i.e., only the different amounts of fatty tissue attached to the conus medullaris7–10. Electrophysiological stimulation evoking the compound muscle action potentials during the surgery is the only helpful method to distinguish a non-functioning C-LS, confirming the presence or absence of a non-functional conus6,11,12. Several reports have been published on histopathological and electrophysiological characteristics, such as central canal-like ependymal-lined lumen with surrounding neuroglial tissues and the border between the conus and non-functioning RMC13–15. However, existing literature lacks a detailed description of the preoperative MR image of RMC. Therefore, this observational study aimed to explore the preoperative MR findings of RMC and investigate radiological findings for the differential diagnosis among the ambiguous secondary neurulation disease spectrum. Results Thirty-eight patients (19 girls, mean age of 7.3 months) were reviewed, and the most common cutaneous marker was a sacral dimple (52.6%). Gluteal fold deviation (21.1%), skin tag (18.4%), hemangioma (5.3%), and cigarette burn scar (2.6%) were also noted. Most patients (60.5%) showed no significant clinical symptom, but 13 patients had polyradiculopathy, confirmed by nerve conduction test and electromyography, and two patients had decreased bladder compliance. Associated diseases were noted in eight patients (21.1%) as follows: an imperforate anus (10.5%), cloacal anomaly (5.3%), and urogenital anomaly (5.3%). The features of the RMC on spine MRI are summarized in Table 1. All patients had extremely low distal C-LS, terminating more distal than S1/2 in 92.1% of patients. Regarding C-LS morphology, the 19 patients had smooth tapering into the caudal end (Fig. 1), and others showed hourglass-shaped fusiform cystic dilatation (Fig. 2). The C-LS exhibited aberrant T2 hypointense signal compared to the juxta-proximal level (94.7%). Syringohydromelia above the C-LS was noted in 8 patients (21.1%), which showed a long syrinx with more than three vertebral body levels. A fine vestigial nerve from the C-LS with T2 hypointense signal was frequently demonstrated (89.5%). These fine, delicate vestigial nerves showed unusual anatomical course (78.9%), attached fatty masses (78.9%), or vestigial nerve root extended into the extradural subcutaneous fat (89.5%). Sacral arachnoid cysts were found in eight patients with extradural locations (Fig. 3). Six patients had sacral deformities, and seven patients had dermal tract without extension into the intradural space. Table 1Morphological features of retained medullary cord on spine MRI.CharacteristicsNumber (%)Redundant cord-like structure (C-LS), ambiguous conus* Low-lying location below L5/S13 (7.9) Low-lying location below S1/235 (92.1) Morphology Smooth tapering into the caudal end19 (50.0) Hourglass with fusiform cystic dilatation19 (50.0) Aberrant T2 hypointense signal36 (94.7) Syringohydromelia8 (21.1)Traversing nerve root within the thecal sac Fine vestigial nerve with T2 hypointense signal**34 (89.5) Fine vestigial nerve with unusual anatomical course30 (78.9) Attached fatty stalk, extending to extradural subcutaneous fat34 (89.5)Arachnoid cyst8 (21.1)Sacral bone/subcutaneous fat layer Sacral deformity6 (15.8) Dermal tract without extension into the intradural space7 (18.4)*A redundant cord-like structure means the ambiguous form, not radiologically conus medullaris, from the evident distal spinal cord and extending to the dural cul-de-sac. **It shows abnormally low signal intensity compared to other nerve roots on T2-weighted axial images, which was retrospectively confirmed as cases identified through intraoperative monitoring.Fig. 1Retained medullary cord with a smooth distal cord-like structure (A) A 10-month-old girl with a sacral dimple had a distal cord-like structure (C-LS) (white arrows) extending into the cul-de-sac. T1-weighted (left) and T2-weighted (right) sagittal images showed the syringohydromelia (asterisk) above the C-LS and attached to the intrathecal fatty tissue (black arrows). (B, C, D) Each T2-weighted (left) and T1-weighted (right) axial images revealed the C-LS (white arrows) with terminal syrinx and T2 hypointense signal fine vestigial nerve root (arrowheads) juxta distal portion of the C-LS. The most distal portion attaches to intrathecal fat tissue (black arrows). (E) The retained medullary cord (RMC) was observed below the S3 level, and multiple small nerve roots on the right side were seen exiting into the neural foramen, and the terminal part of the RMC was a thickened filum terminale. After stimulating the nerve roots, the caudal most two nerve roots on the right were thought to be vestigial roots and were observed to be nonfunctioning roots, so sectioning was performed. (F) Hematoxylin and eosin (left) and immunohistochemical glial fibrillary acidic protein (right) staining also revealed the ependymal lining (black dotted circle) in the C-LS cutting edge with glioneuronal tissue (brown); therefore, the non-functioning RMC was confirmed.Fig. 2Retained medullary cord with a fusiform cystic dilatation of distal cord-like structure. (A) A 9-month-old girl with a sacral dimple received T2-weighted (right) and T1-weighted (left) MR imaging study. Initially, the radiologists had interpreted that the image showed true conus medullaris at L3-4 level accompanied by a filar cyst (asterisk) attached to the fatty tissue (back arrows) at the thecal sac. However, this could be interpreted as an ambiguous conus, called a distal cord-like structure (C-LS) (white arrows), with an hourglass-shaped central cystic dilatation (asterisk) smoothly tapering into a proximal normal cord and a distal cul-de-sac. (B, C, D) Each T2-weighted (left) and T1-weighted (right) axial images revealed the C-LS (white arrow) with central cystic dilatation (asterisk) and T2 hypointense signal fine vestigial nerve root (arrowhead) juxta distal portion of the C-LS. The most distal portion attaches to intrathecal fat tissue (black arrow). Intraoperative neurophysiological stimulation and histopathologic analysis confirmed the retained medullary cord with a non-functioning vestigial nerve root.Fig. 3Retained medullary cord with arachnoid cyst. (A) 5-month-old girl with a gluteal fold deviation represented a distal cord-like structure (C-LS) (white arrows) without syringohydromelia and an extradural cystic lesion (asterisks) attached to the fatty tissue (black arrows) at the thecal sac on a T1-weighted (left) and T2-weighted (right) sagittal image. (B, C, D) Axial T2-weighted images showed the shifted C-LS (white arrows) due to an elongated-shaped arachnoid cyst (asterisk) in the right extradural space, accompanied by T2 hypointense signal component (arrowheads) in the distal C-LS, which was not chemical shift artifact, considering no definite fat tissue at this level. (E) After vertical incision and S1-2 laminectomy, the upper portion of the cyst was noted in the caudal epidural space (arrow). After nerve root stimulation was done to find the caudal end of the functioning cord, untethering was done. (F) Hematoxylin and eosin (left) and immunohistochemical glial fibrillary acidic protein (right) staining also revealed the ependymal lining (black dotted circle) in the C-LS with glioneuronal tissue (brown); therefore, the non-functioning RMC was confirmed. Table 1Morphological features of retained medullary cord on spine MRI.CharacteristicsNumber (%)Redundant cord-like structure (C-LS), ambiguous conus* Low-lying location below L5/S13 (7.9) Low-lying location below S1/235 (92.1) Morphology Smooth tapering into the caudal end19 (50.0) Hourglass with fusiform cystic dilatation19 (50.0) Aberrant T2 hypointense signal36 (94.7) Syringohydromelia8 (21.1)Traversing nerve root within the thecal sac Fine vestigial nerve with T2 hypointense signal**34 (89.5) Fine vestigial nerve with unusual anatomical course30 (78.9) Attached fatty stalk, extending to extradural subcutaneous fat34 (89.5)Arachnoid cyst8 (21.1)Sacral bone/subcutaneous fat layer Sacral deformity6 (15.8) Dermal tract without extension into the intradural space7 (18.4)*A redundant cord-like structure means the ambiguous form, not radiologically conus medullaris, from the evident distal spinal cord and extending to the dural cul-de-sac. **It shows abnormally low signal intensity compared to other nerve roots on T2-weighted axial images, which was retrospectively confirmed as cases identified through intraoperative monitoring. Morphological features of retained medullary cord on spine MRI. *A redundant cord-like structure means the ambiguous form, not radiologically conus medullaris, from the evident distal spinal cord and extending to the dural cul-de-sac. **It shows abnormally low signal intensity compared to other nerve roots on T2-weighted axial images, which was retrospectively confirmed as cases identified through intraoperative monitoring. Retained medullary cord with a smooth distal cord-like structure (A) A 10-month-old girl with a sacral dimple had a distal cord-like structure (C-LS) (white arrows) extending into the cul-de-sac. T1-weighted (left) and T2-weighted (right) sagittal images showed the syringohydromelia (asterisk) above the C-LS and attached to the intrathecal fatty tissue (black arrows). (B, C, D) Each T2-weighted (left) and T1-weighted (right) axial images revealed the C-LS (white arrows) with terminal syrinx and T2 hypointense signal fine vestigial nerve root (arrowheads) juxta distal portion of the C-LS. The most distal portion attaches to intrathecal fat tissue (black arrows). (E) The retained medullary cord (RMC) was observed below the S3 level, and multiple small nerve roots on the right side were seen exiting into the neural foramen, and the terminal part of the RMC was a thickened filum terminale. After stimulating the nerve roots, the caudal most two nerve roots on the right were thought to be vestigial roots and were observed to be nonfunctioning roots, so sectioning was performed. (F) Hematoxylin and eosin (left) and immunohistochemical glial fibrillary acidic protein (right) staining also revealed the ependymal lining (black dotted circle) in the C-LS cutting edge with glioneuronal tissue (brown); therefore, the non-functioning RMC was confirmed. Retained medullary cord with a fusiform cystic dilatation of distal cord-like structure. (A) A 9-month-old girl with a sacral dimple received T2-weighted (right) and T1-weighted (left) MR imaging study. Initially, the radiologists had interpreted that the image showed true conus medullaris at L3-4 level accompanied by a filar cyst (asterisk) attached to the fatty tissue (back arrows) at the thecal sac. However, this could be interpreted as an ambiguous conus, called a distal cord-like structure (C-LS) (white arrows), with an hourglass-shaped central cystic dilatation (asterisk) smoothly tapering into a proximal normal cord and a distal cul-de-sac. (B, C, D) Each T2-weighted (left) and T1-weighted (right) axial images revealed the C-LS (white arrow) with central cystic dilatation (asterisk) and T2 hypointense signal fine vestigial nerve root (arrowhead) juxta distal portion of the C-LS. The most distal portion attaches to intrathecal fat tissue (black arrow). Intraoperative neurophysiological stimulation and histopathologic analysis confirmed the retained medullary cord with a non-functioning vestigial nerve root. Retained medullary cord with arachnoid cyst. (A) 5-month-old girl with a gluteal fold deviation represented a distal cord-like structure (C-LS) (white arrows) without syringohydromelia and an extradural cystic lesion (asterisks) attached to the fatty tissue (black arrows) at the thecal sac on a T1-weighted (left) and T2-weighted (right) sagittal image. (B, C, D) Axial T2-weighted images showed the shifted C-LS (white arrows) due to an elongated-shaped arachnoid cyst (asterisk) in the right extradural space, accompanied by T2 hypointense signal component (arrowheads) in the distal C-LS, which was not chemical shift artifact, considering no definite fat tissue at this level. (E) After vertical incision and S1-2 laminectomy, the upper portion of the cyst was noted in the caudal epidural space (arrow). After nerve root stimulation was done to find the caudal end of the functioning cord, untethering was done. (F) Hematoxylin and eosin (left) and immunohistochemical glial fibrillary acidic protein (right) staining also revealed the ependymal lining (black dotted circle) in the C-LS with glioneuronal tissue (brown); therefore, the non-functioning RMC was confirmed. Discussion The characteristic MR features of RMC revealed an extremely low-lying distal C-LS with smooth tapering or hourglass-shaped cystic dilatation of the caudal part, accompanied by an intradural fatty stalk. The far distal C-LS and the fine vestigial nerve from the C-LS showed T2 hypointense signal, presumed a non-functioning nerve root, as confirmed by intraoperative electrophysiological monitoring. The RMC is a newly defined clinicopathologic entity for closed spinal dysraphism that is thought to originate from the near-complete arrest of apoptosis in the final or regression phase of secondary neurulation3,6. The clinical symptoms may include all those found in tethered cord syndrome3. The surgeon must find the non-functioning medullary cord that does not respond to electrical stimulation, cut it safely, and untether the RMC completely. However, the functioning cord and non-functioning medullary cord are not clearly distinguished in preoperative MRI, and they can only be confirmed by intraoperative neurophysiological monitoring6,7. A previous report suggested that RMCs can be broadly divided into non-cystic forms with a smooth, tapering distal cord and cystic forms with an hourglass-shaped distal cord, based on the distal cord morphology17. Therefore, the imaging findings in this study align with the classification outlined in earlier reports on distal C-LS based on the pathophysiology of RMC. Some RMC cases have varying amounts of fat in the distal part of the low-lying conus, which must be differentiated from some types of lumbosacral lipomatous malformation despite the gray zone. In addition, when the arrest of regression occurs relatively early, this may result in the ‘low-lying conus with distal luminal dilatation’ because the secondary neural tube transiently has a lumen3. Previous case series reported that RMC can be associated with other secondary neurulation disorders such as caudal lipomas, limited dorsal myeloschisis, congenital dermal sinuses, and subcutaneous meningoceles6,8,9,12. This study presented several ancillary findings suggestive of RMCs based on preoperative MRI in addition to the characteristic distal cord-like structure. First, intradural fatty lesions revealed an indistinct border with a distal C-LS and communication with the extradural fat caudally through the dural cul-de-sac opening. Similar characteristics between those secondary neurulation impairments were intradural fatty lesions with the indistinct border attached to low-lying conus medullaris in cases of not presenting the subcutaneous adipose bump tissue. Various amounts of adipose tissue could be explained as a spectrum of regression failures during late secondary neurulation regarding the timing and severity of apoptotic loss13,14,18,19. Second, a characteristic finding distinguishing RMCs from lipomas is the presence of vestigial nerve roots emanating from the caudal cord, possible to see on the MR images. One of the critical features of RMCs might be the predominance of the neuroglial core with an ependyma-lined central canal and vestigial nerve roots, often in a chaotic arrangement reflecting non-functionality3,18. The vestigial nerve roots showed the characteristic T2 hypointense signal on MRI. T2 hypointense signal observed in the far distal C-LS might be explained by one of the chemical shift artifacts discernible at the fat-fluid interface resulting from fatty tissue mixed with the C-LS20. However, in most cases in this study, those vestigial nerve roots exhibited T2 hypointense signal even in the absence of any association with fatty lesions; thus, one of the characteristic findings about RMC. High-resolution, heavily T2-weighted images can provide a more detailed evaluation of RMC and its associated pathology. Recently, the quality of high-resolution 3D T2-weighted images has improved with the development of novel deep-learning reconstruction or acceleration techniques, and they have been used in various diseases. In the future, these 3D high-resolution, heavily T2-weighted images will be beneficial in identifying the distal C-LS structure and its associated vestigial nerve roots or in differentiating non-functioning conus, and studies will be necessary to apply them to various diseases of congenital spinal anomalies. Third, RMCs were frequently accompanied by cystic lesions such as arachnoid cysts, which might be explained by a defect in the caudal axonal mesoderm derived from the caudal cell mass and caudal buds21,22. It might be confused with a filar cyst, one of the cystic structures observed beneath the conus. Filar cysts, however, manifested at the initiation of the filum terminale just below the conus and occupied a midline position. In contrast, arachnoid cysts associated with RMC are frequently observed at the sacral level. Finally, the imperforate anus was the most associated anomaly in the RMC patients. It might also be explained as the malformations of the caudal cell mass and caudal buds21,22. RMCs exhibit broad-spectrum and indistinct MRI features, which may overlap with other secondary neurulation arrests. Recent studies have proposed that disorders such as RMC, caudal- or transitional-type spinal lipoma, and thickened filum terminale may constitute a continuous spectrum of regression failures during late secondary neural tube formation19. These reports suggest that the manifestation of such pathologies is contingent upon the timing and severity of apoptosis failure18. Therefore, there may be controversy over inclusion criteria in this study, whether the definition of RMC disease itself is narrow or broad spectrum. In this study, we retrospectively collected MR image data with a small cohort that was strictly defined and based on the intraoperative data to identify characteristic imaging findings. This study described several MR characteristics of RMC that may have resulted from an arrest of secondary neurulation late in the cavitation stage of medullary cord development just before its degeneration. Going into the field knowing the possibility of RMC through preoperative imaging may help select the incision site, allowing for limited bone exposure and only amputation of the junction between the functional and non-functional cords. This study suggests those radiologic findings can be RMC instead of the previously called terminal syrinx with low-lying conus. Materials and methods This retrospective study was approved by the Seoul National University Hospital Institutional Review Board (IRB no. 2302-118-1407). The institutional review board waived the requirement of informed consent for the study. All methods followed the relevant guidelines and regulations that conform to the Declaration of Helsinki. Study population Using our institution’s clinical data repository system, we identified 425 pediatric patients who underwent untethering surgery. Based on medical records, 147 patients with closed spinal dysraphism who did not have a subcutaneous mass through physical examination were selected. Surgical and histopathological records were reviewed to select 40 patients with identified nonfunctioning distal C-LS, which was identified by intraoperative electrophysiological stimulation and glial tissue. This institution uses the term ‘possible RMC’ for lesions with low-lying conus, which cannot be confirmed electrophysiologically or histologically for reasons such as limited surgical field. In this study, the presumptive diagnosis was excluded to reveal typical imaging features for RMC. Two patients with RMC were excluded from the study because of poor-quality MR images that were difficult to interpret. Finally, thirty-eight patients enrolled in this study were used to evaluate the pre-operative MR imaging. A detailed patient selection flowchart is shown in Fig. 4. Fig. 4Patient Selection Flow Chart. Note. lipomyelomeningocele, LMMC; myelomeningocele, MMC; LDM, limited dorsal myeloschisis. We identified and included 38 patients with intraoperative electrophysiologically and histopathologically confirmed retained medullary cord (RMC), with retrospective reviewing through the clinical data repository system, for image analysis of preoperative performed MRI. Patient Selection Flow Chart. Note. lipomyelomeningocele, LMMC; myelomeningocele, MMC; LDM, limited dorsal myeloschisis. We identified and included 38 patients with intraoperative electrophysiologically and histopathologically confirmed retained medullary cord (RMC), with retrospective reviewing through the clinical data repository system, for image analysis of preoperative performed MRI. MR interpretation MRI was performed using the following parameters: repetition time/echo time, 537–750/–13 ms (T1-weighted) and 3000–4470/93–117 ms (T2-weighted); section thickness, 3 mm (sagittal plane) and 5 mm (axial plane); a field of view, 160–180 × 160–180 mm (sagittal plane), 100–120 × 80–120 mm (axial plane); matrix, 448–512 × 224–256 (sagittal plane) and 256–320 × 128–192 (axial plane). Two pediatric radiologists retrospectively reviewed the MR images by consensus (S. L. and S. B. L., with 11 and 8 years of experience, respectively). The distal C-LS structure means the ambiguous indistinct conus medullaris-like structure from the evident distal spinal cord and extending to the dural cul-de-sac. We investigated this confusing structure, C-LS, previously interpreted as a low-lying cord with terminal syrinx or filar lipoma on MRI as follows: the level and morphology of C-LS, the aberrant signal intensity of far distal C-LS, the nerve root within the sac from the distal C-LS, attached fatty stalk extending to the extradural subcutaneous fat layer, and ancillary findings such as arachnoid cyst, sacral bone, and subcutaneous fat layer abnormalities. Statistical analysis Statistical analyses were primarily descriptive because of the small sample size. Categorical variables were summarized using frequencies and percentages, whereas continuous variables were summarized using means and standard deviations. Statistical significance was set at a P value of less than 0.05. Data analyses were performed using the commercially available statistical software MedCalc, version 20.218.