A 62-year-old male presented with a 2-year history of gait ataxia and sensory loss in both lower extremities. He reported that the symptoms first appeared in his left foot. He described that his left sole felt numb and that he could not feel shoes worn on his left foot. His gait became progressively ataxic, and his colleagues noted that he walked as if he was slightly intoxicated. Sensory symptoms spread slowly to his right sole, resulting in loss of cold and painful sensations 1 year previously. Radicular pain occurred on his left lower abdomen, inguinal area, and buttock around the time when he experienced symptoms in the right sole. He has maintained his daily routine with the occasional leisure activity of riding a mountain bike despite his sensory symptoms. However, the above-mentioned symptoms eventually resulted in him presenting to a local private clinic, where he was diagnosed with an intramedullary tumor and referred to our institution.

He had no medical history except for a spinal tumor and was taking analgesics for his radicular pain. He was alert and oriented in a neurologic examination, and a cranial nerve examination produced unremarkable results. His motor function was normal, including anal tone. Meanwhile, the pain and temperature sensations in the dorsum of his right foot and sole had diminished, while position and vibration sensations were absent in his left foot. The deep tendon reflex was normal, no pathologic reflexes. He reported paresthesia (electric shock pain) in the T9, 10, 11, 12, L1, and L2 dermatomes. He did not report bladder or bowel symptoms. Spine magnetic resonance imaging (MRI) revealed a partially enhanced intramedullary, well-demarcated mass at the T3 and T4 levels. Syrinxes in the rostral and caudal regions and intratumoral hemorrhage in the posterior region of the tumor were also noted (Fig. 1).

Surgical resection was advised, including a detailed explanation of possible postoperative morbidity with the worsening of preexisting symptoms and the advent of new symptoms. He understood and consented to surgical treatment. INM including somatosensory evoked potentials (SSEPs) and transcranial electric motor evoked potentials (tcMEPs) was performed as part of the routine care standard. INM records included tcMEPs and SSEPs of the upper and lower extremities, train of four (TOF), and spontaneous electromyography. TOF recordings were performed during nerve stimulation at the abductor pollicis brevis (APB), and with SSEPs during stimulation of the bilateral median and posterior tibial nerves and recorded at electrodes C3’-C4’ and C4’-C3’ using the international 10-20 system. The stimulation voltages were 20 mV and 30 mV at the upper and lower extremities, respectively, while the stimulation frequency and average rate were 2.9 Hz and 50 times, respectively. MEPs were measured during transcranial electrical stimulation at 400 V for five repetitive pulses at the C3 and C4 electrodes. MEPs were recorded in the bilateral APB, abductor digiti quinti, tibialis anterior, and abductor hallucis muscles (Xltek Protektor 32; Natus Medical, Pleasanton, WI, USA). The baseline MEPs and SSEPs were collected after anesthesia was induced using the following drugs: propofol at 100 µg/kg/minutes and remifentanil at 0.05-0.09 µg/kg/minutes with vecuronium at 0.3-0.9 mg (TOF: 2-3/4). The initial MEP recordings of the bilateral arms and legs were normal, whereas the initial SSEP recordings of both legs revealed discernable yet unstable waveforms. The recordings of the arms were normal. Completed loss of MEPs was noted after the laminectomy at T2-T4 and the laminoplasty at T5 and during epidural vessel cauterization. The operation was paused, and intravenous methylprednisolone (250 mg) was administered, and mean arterial blood pressure was controlled at a target of 105 mmHg. However, the loss of SSEPs in the bilateral legs was detected 10 minutes later, while MEPs and SSEPs in the bilateral arms remained normal. The aforementioned change in SSEPs and MEPs at the lower extremities occurred when the dural opening had not been initiated and the spinal cord had not been manipulated using the motor tract (Fig. 2).

The surgeon ceased the operation due to the serial aggravation of the INM results. MEPs and SSEPs remained lost in the bilateral legs until the end of the operation (Fig. 3). Postoperative spine MRI revealed that the hemorrhage in the syrinx in the caudal region of the tumor was enlarged (Fig. 1). Lower extremity motor function presented marked declines in the proximal and distal parts of the right lower extremity to medical research council (MRC) grade 4, and in the proximal and distal parts of the left lower extremity to MRC grades 3 and 2, respectively. Sensory loss was prominent below the bilateral knees. Despite the aggravation of symptoms, surgery was performed the next day to remove the tumor due to the presence of continuous tumor bleeding. MEPs were not measured in the lower extremities throughout the surgery. SSEP waveforms were discernible, yet unstable, in both lower extremities, and were lost completely after myelotomy and during tumor removal, and did not subsequently recover. The neurologic outcomes of the patient deteriorated after the second operation leading to complete paraplegia, and loss of the sensory modalities of pain, temperature, position, and vibration was observed below the T11 level. Sensory deficit was more severe on the right side (30%) than on the left side (50%). Urination and defecation function were also impacted, highlighting the need for clean intermittent catheterization and regular enemas. Subtle improvement in motor function to MRC grade 1 during right hip extension was noted during rehabilitation at 1 month after the second surgery. After the second operation, repeated MRI revealed gross removal of the tumor. Surgical pathology confirmed the intramedullary mass as ependymoma at World Health Organization grade II. Outpatient follow-up of the bilateral lower extremity motor grade revealed MRC grade 4 at 3 years after the operation, even with continued rehabilitation therapy.

Total microsurgical removal of the ependymoma remains the mainstay of therapy, and was first described as standard primary care by McCormick et al.³ Tumor recurrence was reported in 4.2% among 86.3% gross total resections in 102 surgical cases. Even though gross total resection is widely documented with an acceptable recurrence rate, aggressive resection often results in significant surgical complications, resulting in ependymoma having the highest postoperative morbidity rate among intramedullary tumors.^2,5 Poor prognosis is associated with intratumoral hemorrhage in the thoracic spine, poor preoperative neurologic functioning, and old age, which were somewhat concordant with the present case.^2,4 The risk factors for postoperative morbidity in ependymoma are a posterior approach to myelotomy that intrinsically damages the posterior column, vessel cauterization in the anterior and posterior regions of the tumor resulting in ischemic damage, and inevitable postlaminectomy syndrome. Since improvement in neurologic symptoms is uncommon after tumor removal, delaying the surgery until significant neurologic symptoms arise is not advocated even in patients with no or only insignificant preoperative symptoms.^1,2

A traumatic intramedullary spinal cord hemorrhage (hematomyelia) is a rare form of myelopathy that results from various underlying causes, including vascular malformation, anticoagulant therapy, bleeding disorders, and spinal cord tumor including metastatic ones.¹ Surgical manipulation led to acute intrasyringeal hemorrhage in the present case, resulting in the losses of MEPs and SSEPs and profound neurologic deterioration.

INM is a well-established tool for predicting postoperative neurologic deficit in spinal tumor resection.⁶ Previous studies have found various sensitivities (75-100%) and specificities (91-100%) of INM for predicting postoperative neurologic deficits after spinal tumor surgery according to tumor location and monitoring modality.^7,8 Although there is a consensus favoring INM in spinal tumor surgery, unresolved issues of false positives resulting in incomplete tumor resection and cost-effectiveness must be addressed.⁶ The present case demonstrated the complete and sudden loss of MEPs and SSEPs in the bilateral lower extremities, hence indicating neural damage. The value of INM lies in detecting real-time surgical damage correlated with functional deterioration. However, INM only has little value in reducing the rate of postoperative neurologic deficit, since detecting neural impairment does not always prevent the damage.⁹ The historical control study by Sala et al.¹⁰ found better outcomes in a group of patients who received INM after surgery. In that study, the intraoperative detection of neural damage could not prevent the approaching neurologic deficit, since the tumor had to be resected.¹⁰ If INM had not been conducted in the present case, the surgical procedure would not have been stopped. It would therefore have been difficult to determine the cause of neurologic decline after the surgery; that is, whether it was caused by the spontaneous bleeding or surgical intervention. This case indicates that INM is a reliable method in intramedullary spine tumor surgery that can sensitively detect surgical damage.