Spinal subependymoma mimicking an intradural extramedullary tumor: illustrative case Authors: Toshiaki Inomo, Toshiki Fukuoka, Masahito Hara, Ryuya Maejima, Risa Oishi, Koji Osuka, Tadashi Watanabe Published: 2026 ABSTRACT BACKGROUNDSpinal subependymomas are rare, benign tumors that typically arise within the spinal cord and usually appear as intramedullary lesions on MRI. The authors report a rare case of cervical spinal subependymoma with exophytic growth that mimicked an intradural extramedullary tumor on MRI.OBSERVATIONSA 53-year-old man presented with right-sided paresthesia. MRI revealed a spinal tumor causing marked spinal cord compression. The lesion was isointense on T1-weighted images, slightly hyperintense on T2-weighted images, and showed no gadolinium enhancement. A focal CSF cleft between the tumor and spinal cord led to a preoperative diagnosis of an intradural extramedullary tumor, and schwannoma was suspected. However, intraoperative pathological examination suggested a low-grade glioma, prompting a change in the surgical strategy. Subtotal resection was therefore performed to preserve neurological function. Histological examination subsequently confirmed the diagnosis of subependymoma.LESSONSSpinal subependymoma may rarely present with an apparent intradural extramedullary appearance on MRI. The presence of a CSF cleft and exophytic growth should not exclude an intramedullary origin. Awareness of this atypical presentation and avoidance of anchoring bias are essential for appropriate intraoperative decision-making. When subependymoma is suspected, maximal safe resection should be prioritized over aggressive tumor removal to preserve neurological function. https://thejns.org/doi/10.3171/CASE26728 Spinal subependymomas are rare, benign WHO grade 1 tumors that account for only a small proportion of spinal tumors.1 They are slow-growing lesions that most commonly arise in the cervical spinal cord and generally exhibit favorable long-term outcomes following surgical treatment.2 Because of their rarity and nonspecific clinical presentation, preoperative diagnosis remains challenging, and the definitive diagnosis is frequently established only after histopathological examination.3 Radiologically, spinal subependymomas typically appear as eccentrically located intramedullary tumors with minimal or no contrast enhancement and are therefore most commonly mistaken for other intramedullary tumors, particularly ependymomas and astrocytomas.2–6 An intradural extramedullary tumor-like appearance is exceedingly uncommon and may lead to diagnostic confusion as well as inappropriate intraoperative decision-making. Here, we report a rare case of cervical spinal subependymoma with marked exophytic growth that mimicked an intradural extramedullary tumor on MRI. We also discuss the radiological pitfalls, the impact of anchoring bias during surgery, and the optimal surgical strategy. Illustrative Case Clinical History A 53-year-old man presented to another hospital with right-sided paresthesia. MRI revealed a cervical spinal tumor, and he was referred to our institution for further evaluation and treatment. On admission, the patient complained only of right-sided paresthesia and had no motor weakness or gait disturbance. MRI demonstrated a dorsal tumor at the C1–2 level causing marked compression of the spinal cord. The lesion was isointense to the spinal cord on T1-weighted (T1W) images and slightly hyperintense on T2-weighted (T2W) images (Fig. 1A and B). No gadolinium enhancement was identified (Fig. 1C). The tumor was relatively well demarcated from the normal spinal cord, and a CSF cleft was focally observed between the lesion and the spinal cord (Fig. 1D and E). CT showed no calcification within the lesion (Fig. 1F). Based on these imaging findings, an intradural extramedullary tumor was suspected. Although the absence of contrast enhancement was atypical, schwannoma was considered the most likely diagnosis, and elective tumor resection was performed. FIG. 1.Preoperative images. A: Sagittal T1W MR image showing a lesion isointense to the spinal cord. B: Sagittal T2W MR image showing a lesion slightly hyperintense to the spinal cord. C: Gadolinium-enhanced sagittal T1W MR image showing no contrast enhancement. D and E: Axial T2W MR images showing a well-demarcated interface between the lesion and the normal spinal cord, with a CSF cleft (arrowhead) partially observed between them. F: Sagittal CT image showing no calcification within the lesion. Preoperative images. A: Sagittal T1W MR image showing a lesion isointense to the spinal cord. B: Sagittal T2W MR image showing a lesion slightly hyperintense to the spinal cord. C: Gadolinium-enhanced sagittal T1W MR image showing no contrast enhancement. D and E: Axial T2W MR images showing a well-demarcated interface between the lesion and the normal spinal cord, with a CSF cleft (arrowhead) partially observed between them. F: Sagittal CT image showing no calcification within the lesion. Surgical Findings Under general anesthesia, the patient was placed in the prone position. The occipital bone to the C2 spinous process was exposed. A suboccipital craniotomy and partial laminectomy of C1 and the upper portion of C2 were performed, followed by dural opening. The tumor was found beneath an intact arachnoid membrane and extensively covered the dorsal surface of the spinal cord, making it difficult to visualize its entire extent (Fig. 2A and B). Initially, a small portion of the lesion surface was submitted for intraoperative pathological examination. Internal decompression was continued, and once the tumor had become sufficiently mobile, a nerve root was identified emerging from the tumor (Fig. 2C–E). An intraoperative pathological examination suggested a low-grade glioma, which for the first time raised the possibility of an intramedullary origin of the lesion. Gross-total resection was abandoned, and the tumor was removed as much as possible while carefully preserving the interface with the normal spinal cord. As internal decompression progressed, the tumor became increasingly mobile, and continuity between the tumor and the spinal cord was observed at the ventral surface of the tumor (Fig. 2F and G). Ultimately, a small portion of the tumor was intentionally left in situ (Fig. 2H). FIG. 2.Intraoperative findings. A: The tumor was located just beneath the arachnoid membrane. B: The tumor extensively covered the dorsal aspect of the spinal cord. C: Internal decompression was performed. D: After mobilization of the tumor (blue arrowhead), the normal spinal cord (orange arrowhead) was identified beneath the lesion. E: A nerve root (green arrowhead) was observed arising from the tumor (blue arrowhead). The orange arrowhead indicates the normal spinal cord. F: After debulking the tumor, a plane of continuity between the rostral aspect of the tumor and the spinal cord (white arrowhead) was identified. G: A plane of continuity between the caudal aspect of the tumor and the spinal cord (white arrowhead) was also identified. H: Final surgical view. A small portion of the tumor was intentionally left behind. Intraoperative findings. A: The tumor was located just beneath the arachnoid membrane. B: The tumor extensively covered the dorsal aspect of the spinal cord. C: Internal decompression was performed. D: After mobilization of the tumor (blue arrowhead), the normal spinal cord (orange arrowhead) was identified beneath the lesion. E: A nerve root (green arrowhead) was observed arising from the tumor (blue arrowhead). The orange arrowhead indicates the normal spinal cord. F: After debulking the tumor, a plane of continuity between the rostral aspect of the tumor and the spinal cord (white arrowhead) was identified. G: A plane of continuity between the caudal aspect of the tumor and the spinal cord (white arrowhead) was also identified. H: Final surgical view. A small portion of the tumor was intentionally left behind. Histopathological Findings H&E staining demonstrated scattered clusters of small cells within a fibrillary background accompanied by microcystic changes (Fig. 3A and B). Neither perivascular pseudorosettes nor true ependymal rosettes, characteristic histological features of ependymoma, were observed. Nuclear atypia and mitotic activity were absent. Immunohistochemical analysis showed diffuse positivity for glial fibrillary acidic protein (GFAP), and the Ki-67 labeling index was approximately 1% (Fig. 3C and D). Stainings for both epithelial membrane antigen (EMA) and SRY-related HMG-box gene 10 (SOX10) were negative (Fig. 3E and F), making extramedullary tumors such as meningioma and schwannoma, which had been suspected preoperatively, less likely. Staining for oligodendrocyte transcription factor 2 (OLIG2) was also negative (Fig. 3G), making an astrocytic tumor less likely. Based on these findings, the tumor was diagnosed as a spinal subependymoma. FIG. 3.Pathological findings. A and B: H&E staining showed scattered clusters of small cells within a fibrillary background accompanied by microcystic changes. Original magnification ×40 (A) and ×100 (B). C: Immunohistochemical staining showed diffuse positivity for GFAP. Original magnification ×40. D: The Ki-67 labeling index was approximately 1%. Original magnification ×40. E: Immunohistochemical staining was negative for EMA. Original magnification ×40. F: Immunohistochemical staining was negative for SOX10. Original magnification ×40. G: Immunohistochemical staining was negative for OLIG2. Original magnification ×40. Pathological findings. A and B: H&E staining showed scattered clusters of small cells within a fibrillary background accompanied by microcystic changes. Original magnification ×40 (A) and ×100 (B). C: Immunohistochemical staining showed diffuse positivity for GFAP. Original magnification ×40. D: The Ki-67 labeling index was approximately 1%. Original magnification ×40. E: Immunohistochemical staining was negative for EMA. Original magnification ×40. F: Immunohistochemical staining was negative for SOX10. Original magnification ×40. G: Immunohistochemical staining was negative for OLIG2. Original magnification ×40. Postoperative Course The patient developed right upper extremity weakness immediately after surgery, with a manual muscle test (MMT) grade of 3. The weakness improved to MMT grade 4 on postoperative day 1 and gradually improved with rehabilitation. The patient was discharged home on postoperative day 12 with a modified Rankin Scale (mRS) score of 2 and returned to work 2 months after surgery. At the 1-year follow-up, his mRS score had improved to 1 and his muscle strength had recovered to nearly the preoperative level, although right-sided paresthesia persisted. No adjuvant therapy was administered, and follow-up MRI has demonstrated no evidence of tumor progression (Fig. 4). One limitation of the present case is that the extent and volume of residual tumor are difficult to assess accurately because its signal intensity is relatively similar to that of the normal spinal cord, and it shows no contrast enhancement on MRI. We plan to perform follow-up MRI every 6–12 months and carefully monitor the patient for any clinical or radiological changes. FIG. 4.A: Sagittal T2W MR image obtained immediately after surgery. B and C:Sagittal (B) and axial (C) T2W MR images obtained at the 1-year follow-up. No tumor recurrence was observed. A: Sagittal T2W MR image obtained immediately after surgery. B and C:Sagittal (B) and axial (C) T2W MR images obtained at the 1-year follow-up. No tumor recurrence was observed. Informed Consent The necessary informed consent was obtained in this study. Discussion Observations Subependymomas are rare, benign WHO grade 1 tumors that account for less than 1% of all CNS tumors and represent only a small proportion of spinal ependymal tumors.1,7,8 They most commonly arise in the fourth and lateral ventricles, whereas spinal involvement is uncommon and predominantly occurs in the cervical and cervicothoracic regions.3,4 Owing to their indolent growth and nonspecific imaging findings, preoperative diagnosis remains challenging,6 and it is not uncommon for subependymoma to be diagnosed only after resection and subsequent histopathological examination.8 Radiologically, spinal subependymomas typically present as eccentrically located intramedullary tumors that are iso- to hypointense on T1W imaging, hyperintense on T2W imaging, and demonstrate minimal or absent gadolinium enhancement.4,5 As a result, they are often preoperatively mistaken for other intramedullary tumors, particularly ependymomas and astrocytomas.6 Toi et al. proposed that their subpial growth pattern results in a steep fusiform expansion of the spinal cord, termed the “bamboo leaf sign,” which may help distinguish subependymoma from ependymoma and astrocytoma.9 Overall, spinal subependymomas are generally recognized radiologically as intramedullary tumors. However, our case demonstrated none of these characteristic findings and instead closely mimicked an intradural extramedullary tumor. The lesion caused marked ventral displacement of the spinal cord, and an apparent CSF cleft was identified between the tumor and the cord, leading us to suspect schwannoma preoperatively (Fig. 5). FIG. 5.Schematic illustration of the lesion. The subependymoma, an intramedullary tumor, extended into the subarachnoid space with exophytic growth, resulting in an appearance mimicking an intradural extramedullary tumor. Schematic illustration of the lesion. The subependymoma, an intramedullary tumor, extended into the subarachnoid space with exophytic growth, resulting in an appearance mimicking an intradural extramedullary tumor. Only a few reports have described spinal subependymomas with an exophytic or apparently extramedullary appearance. Do et al. reported a cervical subependymoma that appeared completely extramedullary on MRI and emphasized that subependymoma should remain in the differential diagnosis of spinal masses with an extramedullary appearance.10 Similar to their case, our patient had a dorsal cervical lesion with minimal enhancement and marked exophytic growth. However, our case is unique in that an apparent CSF cleft was identified preoperatively, further strengthening the impression of an intradural extramedullary lesion. We speculate that, rather than following its typical subpial growth pattern, the tumor expanded predominantly toward the subarachnoid space, thereby creating an apparent CSF cleft between the lesion and the spinal cord and mimicking an intradural extramedullary tumor. Another important lesson from this case is the influence of anchoring bias on intraoperative decision-making. Because preoperative imaging strongly suggested an intradural extramedullary tumor, the intraoperative findings were initially interpreted within this diagnostic framework. Furthermore, because of the large size of the tumor, adequate mobilization and reflection of the lesion were difficult. Consequently, the point of attachment to the spinal cord could not be adequately identified (Fig. 2B). Furthermore, the finding of a nerve root emerging from the tumor further reinforced the erroneous impression that the lesion was a schwannoma arising from a nerve root (Fig. 2E). The possibility of an intramedullary origin was not recognized until the intraoperative pathological examination suggested a low-grade glioma. In such a situation, inadvertent manipulation could have resulted in injury to the normal neural tissue. This case highlights that excessive reliance on preoperative imaging may bias intraoperative interpretation and delay recognition of the true pathology, underscoring the importance of maintaining diagnostic flexibility throughout surgery. The present case also has important implications for surgical management. Subependymomas are benign and extremely slow-growing tumors with a low proliferative potential and a very low risk of recurrence. Several studies have shown favorable long-term outcomes even after subtotal resection.2,4,11 Soleiman et al. reviewed 105 reported cases and demonstrated that postoperative neurological deterioration occurred in more than half of patients undergoing gross-total resection, whereas malignant transformation has never been reported.2 Similarly, Jain et al. reported durable tumor control after subtotal resection in a series of subependymomas.4 Therefore, maximal safe resection rather than aggressive gross-total resection should be considered the preferred surgical strategy. In our patient, gross-total resection was abandoned once an intramedullary origin was suspected, and a small residual tumor was intentionally left to preserve neurological function. Although transient postoperative weakness developed, the patient recovered to nearly his preoperative neurological status and has shown no radiological progression during follow-up. This case highlights the importance of continuously reassessing the diagnosis and surgical strategy during surgery. Once a benign intramedullary tumor such as subependymoma is suspected, preservation of neurological function should take precedence over aggressive gross-total resection. Lessons Spinal subependymomas can rarely exhibit marked exophytic growth and mimic intradural extramedullary tumors on MRI. The presence of a CSF cleft does not necessarily indicate an extramedullary origin. Surgeons should remain aware of this possibility and avoid anchoring bias during intraoperative decision-making. Once an intramedullary origin is suspected, maximal safe resection should be prioritized over aggressive tumor removal to preserve neurological function. Disclosures The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author Contributions Conception and design: Inomo, Maejima. Acquisition of data: Inomo, Fukuoka, Oishi. Analysis and interpretation of data: Inomo, Fukuoka, Oishi. Drafting the article: Inomo, Osuka. Critically revising the article: Fukuoka, Hara, Oishi, Osuka, Watanabe. Reviewed submitted version of manuscript: Inomo, Oishi. Approved the final version of the manuscript on behalf of all authors: Inomo. Administrative/technical/material support: Fukuoka. Study supervision: Hara, Watanabe. Correspondence Toshiaki Inomo: Aichi Medical University, Nagakute, Aichi, Japan. inomo.toshiaki.495@mail.aichi-med-u.ac.jp.