Thoracic root–related intradural extramedullary cavernoma presenting with subarachnoid hemorrhage: illustrative case

Abstract

BACKGROUND

Just 5% of all cavernomas are located in the spine. Thoracic root–related subtypes are the rarest, with a total of 14 cases reported in the literature to date. Among them, only 4 presented with subarachnoid hemorrhage (SAH).

OBSERVATIONS

A 65-year-old female presented after an ictus of headache with no neurological deficits. Computed tomography (CT) demonstrated sulcal SAH, with the remainder of the workup nondiagnostic for etiology. Three weeks later, she re-presented with acute thoracic back pain and thoracic myelopathy. CT and magnetic resonance imaging suggested dubiously a T9–10 disc herniation with spinal cord compression. Surgical decompression and resection were then planned. Intraoperative ultrasound (IUS) demonstrated an intradural extramedullary lesion, confirmed to be cavernoma. Complete resection was achieved, and the patient was discharged a few days postoperatively to inpatient rehabilitation.

LESSONS

Although spine imaging is deemed to be low yield in the evaluation of cryptogenic SAH, algorithms can be revisited even in the absence of spine-related symptoms. Surgeons can be prepared to change the initial surgical plan, especially when preoperative imaging is unclear. IUS is a powerful tool to assess the thecal sac after its exposure and to help guide this decision, as in this rare entity.

Cavernomas are vascular malformations composed of flimsy, thin-walled sinusoidal vessels with no intervening parenchyma, typically angiographically occult. Just 5% of all cavernomas occur in the spinal canal. Among the vascular spinal pathologies, they comprise 3% to 16% of cases, and among these, a mere 3% are intradural extramedullary.^1–8

Magnetic resonance image (MRI) is the mainstay diagnostic imaging modality for these lesions, but unlike their intracranial counterparts, the typical “popcorn” imaging appearance is not often described for spinal cavernomas. T1 and T2 sequence signal patterns are impacted by the age of hemoglobin breakdown products and the amount of gadolinium enhancement. Generally, these lesions are noted to have a hemosiderin rim, suggesting the pathology.^6,9–11

The first description of an intradural extramedullary cavernoma dates back to 1903 in an autopsy report, and the first resection was detailed by Hirsch et al. in 1965, in a young patient experiencing sphincter dysfunction.^3,12,13 Overall, 72 surgically treated cases have been reported in the literature.^12,13

In the current work, we present a thoracic root–related cavernoma whose clinical presentation included cortical subarachnoid hemorrhage (SAH) initially with no clear etiology, followed by acute midthoracic back pain and myelopathy. This progression ultimately led to surgery, which confirmed the diagnosis.

Illustrative Case

A 65-year-old female with a past medical history of obesity, hypertension, and fibromyalgia presented with an atraumatic, sudden-onset severe thunderclap headache in the setting of elevated blood pressure without neurological deficits. Computed tomography (CT) of the brain revealed a mild SAH within the convexity sulcus, and MRI redemonstrated the same nonaneurysmal pattern (Fig. 1). She was initially admitted to the neurointensive care unit and underwent further cerebrovascular workup, including CT angiography of the head and neck, as well as a diagnostic cerebral angiography, which was negative for intracranial vascular pathology. Furthermore, the patient also experienced a partial simple seizure, responsive to antiepileptic therapy. Hypotheses of reversible cerebral vasoconstriction syndrome or posterior reversible encephalopathy syndrome were then established, and the patient was treated accordingly. Over the course of her 2-week hospital stay, the patient’s symptoms resolved and she was discharged home.

FIG. 1.

Axial CT demonstrated a convexity sulcal pattern of SAH (A), with clear basal cisterns (B). Axial fluid-attenuated inversion recovery (FLAIR) sequences demonstrated the same pattern (C), also showing a small amount of hemorrhage in the interpeduncular cistern (D), with no other new findings as compared to CT.

Approximately 1 week later, she experienced severe thoracic back pain with radiation to the bilateral lower extremities in a nondermatomal distribution, prompting her to return to the emergency department. On physical examination, she had brisk patellar reflexes (grade 3+) with bilateral clonus. Additionally, she also had urinary retention, and a Foley catheter was placed. Given the myelopathic signs, thoracic spine MRI was performed, demonstrating a T9–10 disc herniation with spinal cord signal change; however, an intradural lesion could not be completely excluded (Fig. 2).

FIG. 2.

Sagittal and axial T2-weighted MRI sequences portrayed a lesion contiguous to the T9–10 disc (white arrowhead, A), resulting in a picture more suggestive of an extraaxial origin (white arrow, B) with cord edema. Postgadolinium sequences showed just mild peripheral enhancement (C and D).

Upon review of the imaging and clinical presentation, surgical intervention was recommended. The patient was placed prone on a Jackson table. After completing the subperiosteal dissection, T9–10 laminectomies were performed. We devised two surgical strategies based on the intraoperative ultrasound (IUS) evaluation: 1) if it revealed an extradural lesion, we planned to perform a transpedicular approach with instrumentation for resection of the lesion; and 2) if it demonstrated an intradural lesion, we would proceed with a standard midline durotomy for resection.

Next, utilizing the IUS, a lesion was observed, although it was noted to be intradural with a clear border adjacent to the spinal cord (intradural extramedullary compartment). We then performed a midline durotomy and intradural dissection, identifying a berry-like extramedullary lesion originating from the left T10 exiting nerve root. We performed a gross-total resection with nerve root sacrifice (Fig. 3).

FIG. 3.

IUS after thoracic laminectomy confirmed that the lesion (hypoechoic) was intradural, using a small cottonoid (black dotted arrow, A) over the dura to reference its exact site. After the durotomy, a small subarachnoid-subpial hematoma (white arrow, B) was seen just above the lesion level, revealing a hemosiderin layer covering the cord (C) after its removal. Further dissection revealed the cavernoma (Cav) on the left side of the cord (D), and with progression of its piecemeal resection (E), it became evident that its origin related to the left T10 root (white asterisk). Gross-total resection was achieved, sacrificing the involved root (F). Postoperative sagittal T2-weighted (G), postgadolinium axial T1-weighted (H), and axial T2-weighted (I) MRI sequences showed no residual lesion.

The patient had an outstanding postoperative course with resolution of her back pain and sphincter dysfunction and was discharged home on postoperative day 6. Postoperative MRI demonstrated no residual lesion (Fig. 3G–I).

Subsequent pathology confirmed the final diagnosis of thoracic root–related cavernous malformation (Fig. 4). In retrospect, we could attribute the initial nonaneurysmal SAH to a prior hemorrhage from this lesion.

FIG. 4.

Photomicrographs of the surgical specimen, compatible with cavernous malformation. Hematoxylin and eosin–stained samples showed macrophages and fibrin clusters (A), with multiple large and hyalinized vessels (B). Trichrome stain highlighted the pathological vessels in dense blue (C), and CD163-expressing histiocytes were markedly demonstrated on immunohistochemistry (D). Original magnification ×10.

Patient Informed Consent

The necessary patient informed consent was obtained in this study.

Discussion

Observations

In a comprehensive review of all previously documented surgically treated intradural extramedullary cavernomas, we found that only 14 were root-related thoracic variants (Table 1).^1–4,6–15 Remarkably, within this rare category, only 4 presented with SAH, and the remainder presented with subacute-chronic radiculopathy, myelopathy, and/or back pain.^12,13

TABLE 1.

Case reports of thoracic root–related extramedullary cavernomas

Case No.	Authors & Year	Sex, Age (yrs)	Level	Isolated SAH (other signs)
1	Heimberger et al., 19823	M, 24	T2–3	Yes
2	Pagni et al., 199013	M, 46	T12–L1	No (radiculopathy)
3	Mastronardi et al., 199113	F, 49	T4	No (myelopathy for 6 mos)
4	Mori et al., 19913	M, 65	T1	Yes
5	Sharma et al., 19925	M, 43	T5	Yes
6	Moreno et al., 199513	M, 63	T12–L1	No (back pain & sphincter dysfunction for 7 yrs)
7	Rao et al., 199713	F, 35	T12	No (radiculopathy for 3 yrs)
8	Er et al., 200713	M, 67	T12–L2	No (radiculopathy for 4 mos)
9	Jin et al., 201113	M, 55	T12–L1	No (headaches & dizziness for 3 mos)
10	Tao et al., 201413	M, 45	T3–4	Yes
11	Shi et al., 201713	M, 73	T11–12	No (myelopathy for 5 yrs)
12	Ziechmann et al., 201813	M, 55	T3–4	No (myelopathy & radiculopathy for 1 mo)
13	Vicenty et al., 20199	M, 56	T2	No (myelopathy for 1 yr)
14	Tawil et al., 202312	F, 55	T11–12	No (myelopathy & back pain for 1 yr)

Approximately 15% of intracranial SAH cases ultimately remain cryptogenic.^16,17 Scrutinizing the etiology of a nonperimesencephalic SAH with an initial negative angiography, Andaluz et al.¹⁷ mentioned the role of spinal CT angiography as a next step, but advised regarding the lack of evidence. Germans et al.¹⁸ specifically investigated the yield of spinal imaging for nonaneurysmal nonperimesencephalic SAH in a multicenter prospective study, concluding that the yield and clinical relevance of spinal MRI in this context are low and thus would not be recommended. In 2018, Sadigh et al.¹⁹ performed a meta-analysis on 240 patients with an angiographically negative atraumatic SAH to evaluate the role of cervical MRI, demonstrating an estimated pooled diagnostic yield of just 1.3%. Based on their findings, they concluded that cervical MRI was not recommended in these cases, although one caveat was considered for proceeding otherwise: the presence of clinical findings suggesting a spinal etiology. In our case, although the SAH could be attributed to the thoracic cavernoma, the patient did not present with spinal symptoms initially.

Studies on IUS for spinal procedures were initiated in the 1980s.²⁰ Its application to traumatic, degenerative, and oncological cases has been consistently described and then was consolidated as a guidance for surgical strategy.^20,21 Harel et al.²² published a surgical case series encompassing spinal tumors, thoracic disc herniations, and Chiari malformations. IUS altered the course of the operation in 63% of these cases, entailing either an extension of the exposure or additional steps to enhance resection or ensure optimal neural decompression based on IUS assessments.

Numerous cases of MRI-related misdiagnoses or imprecise definitions have been published.^1,23,24 Even with gadolinium-enhanced sequences, it can be challenging to distinguish between a degenerative condition such as a disc herniation and a neoplastic entity. This difficulty arises because of varying patterns of enhancement influenced by local inflammatory responses, potentially leading to an inaccurate diagnosis. MRI signal intensity of different pathologies may present a dynamic switch of classic patterns along their natural history, further complicating this matter. Cases of sequestrated discs masquerading as tumors and vice versa have been described.^23,25 Different MRI approaches and sequences have been developed and proposed to overcome these diagnostic hindrances.

In the setting of disc herniations, coronal imaging has been proposed as a powerful tool to differentiate schwannomas from a mimicking tumor disc, an entity defined as a sequestrated disc fragment that was misdiagnosed as a tumor based on neuroimaging criteria.²⁵

This contemporary landscape presents another important angle to be discussed, regarding the inconclusiveness of preoperative spinal MRI about the precise location of the pathology. Radiological reads in this case stamped a T9–10 disc herniation as the conclusion, but our own analysis was not that unequivocal. Retrospectively looking at these images, one might point a possible hemosiderin rim and an intradural configuration for the lesion at one cut or another, but the acute presentation of back pain with compressive myelopathy due to a lesion contiguous to the disc space and the overall appearance pointed otherwise, along with the obvious epidemiological argument. We considered these limitations intrinsic to our current MRI technique. But our uncertainty granted a stepwise plan of starting with a laminectomy, performing IUS, and also directly exploring the anatomy to make an intraprocedural decision of the next steps of our approach.

Lessons

Although spinal imaging is conventionally deemed as low yield in cryptogenic SAH, diagnostic algorithms can be rethought as the knowledge builds up even in the absence of spine-related clinical findings. Surgeons can be prepared to adapt their initial surgical plan, establishing a stepwise approach, particularly when preoperative imaging is unclear. IUS is a powerful tool to assess the thecal sac after its exposure, enabling real-time decision-making as regards how to manage next steps. These points underscore the importance of continually refining our diagnostic and therapeutic strategies in dealing with rare entities, such as the one we described herein.

Author Contributions

Conception and design: Martins Coelho Junior, Toop. Acquisition of data: Martins Coelho Junior, Toop. Analysis and interpretation of data: Chakravarthy, Martins Coelho Junior. Drafting the article: Martins Coelho Junior. Critically revising the article: Chakravarthy. Reviewed submitted version of manuscript: Chakravarthy, Kobalka. Approved the final version of the manuscript on behalf of all authors: Chakravarthy. Statistical analysis: Chakravarthy. Administrative/technical/material support: Chakravarthy. Study supervision: Chakravarthy.