Thomas T. Lee

Diagnostic and surgical management of spinal dural arteriovenous fistulas.

OBJECTIVE A retrospective review was conducted to compare magnetic resonance (MR) and conventional spinal angiographic images and to investigate the outcome of our treatment protocol for patients with spinal dural arteriovenous fistulas (DAVFs). MATERIALS AND METHODS Nine patients with a diagnosis of DAVF based on clinical myelopathy and preoperative MR imaging (MRI) and MR angiography (MRA) findings were treated at our institution by the senior author (BAG). All nine patients initially presented with progressive myelopathy. Preoperative MRI revealed T2-weighted signal abnormalities in all patients, and MRA was diagnostic in all patients. Each patient underwent a laminectomy and ligation of the arterialized draining vein. Selective spinal angiograms were used to confirm the level of fistula immediately before the surgical procedure was performed and to document complete obliteration after clip ligation of the medullary draining vein. Follow-up MRI and MRA were performed approximately 2 months postoperatively. RESULTS MRI T2-weighted signal hyperintensity improved after surgery in all nine patients. Postoperatively, progression of motor weakness and gait difficulty was halted and some improvement was observed in all patients. No patient was neurologically normal, however. To date, there has been no clinical or MRA evidence of recurrence in any patient. CONCLUSION Preoperative MRA and intraoperative spinal x-ray angiography present as an effective combination for diagnosing and intraoperatively confirming DAVF. Both T1-weighted enhancement and T2-weighted signal hyperintensity on MR images improved after the obliteration of the DAVFs and correlated with clinical improvement in all nine patients. MRA provides adequate visualization and localization of spinal DAVFs and may serve as a useful noninvasive tool for diagnosing and following patients with spinal DAVFs in the future.

Surgical treatment of post-traumatic myelopathy associated with syringomyelia

Study Design. Retrospective review. Objective. Evaluate the clinical outcome of surgical intervention for post-traumatic syringomyelia. Introduction. Progressive post-traumatic cystic myelopathy (PPCM), or syringomyelia, can occur after spinal cord injury. The authors present their surgical treatment protocol and treatment outcome of a series of patients with post-traumatic syringomyelia. Methods. The medical records of 53 patients with PPCM undergoing surgical treatment were reviewed. Laminectomies and intraoperative ultrasonography were performed. For patients with no focal tethering and only a confluent cyst on ultrasonography, a syringosubarachnoid shunt (stent) was inserted. For patients with both tethering and a confluent cord cyst, an untethering procedure was performed first. When a cyst showed significant size reduction (>50%) after untethering, no shunt was placed. When the cyst size persisted on ultrasonographic images, a short syringosubarachnoid shunt was used. The mean follow-up was 23.9 months for the 45 patients available for follow-up (range 12–102 months). Results. The interval between the causative event and the operation was from 5 months to 37 years (mean 6.5 years). Pain was the most frequent manifestation, followed by motor deterioration and spasticity. Postoperative improvements in >50% of the patients were noted in those presenting with worsening motor function or spasticity. In 19 of 28 patients with associated tethered spinal cord, untethering alone caused significant collapse of the cyst. Postoperative MRI demonstrated cyst collapse in 95% of the patients with untethering alone and 93% of the patients with a syringosubarachnoid shunt. Conclusion. Post-traumatic syringomyelia can occur with or without cord tethering. Untethering alone for patients with cord tethering and cyst formation can reduce cyst size and alleviate the symptoms and signs of syringomyelia in the majority of these cases. Untethering with expansion of subarachnoid space with an expansile duraplasty may be a more physiologic way of treating a tethered cord with associated syringomyelia, i.e., treating the cause rather than the result.

Treatment of Stable Burst Fracture of the Atlas (jefferson Fracture) With Rigid Cervical Collar

Study Design. A retrospective review of a clinical series. Objective. To evaluate the use of a rigid cervical collar alone as the treatment for stable Jefferson fracture, and to devise an algorithm for treatment of Jefferson fracture with or without an associated cervical injury. Summary of Background Data. The traditional treatment for Jefferson fracture, if there is no indication for surgery, is immobilization by halo vesst. Because halo vest placement is associated with intracranial infection and a significant degree if patient discomfort, slightly less rigid forms of external immobilization may be useful for the treatment of stable Jefferson fractures. No standard protocol calling for the use of one form of stabilization device has been reported. Materials. The medical records and radiographs of 16 consecutive patients with Jefferson fracture during a 2‐year period were reviewed. Each patient underwent a complete cervical radiograph series and a computed tomographic scan. The mean C1 lateral mass displacement was 1.8 mm. Cervical spine radiographs, including lateral flexion‐extension views were obtained 10 to 12 weeks after injury before the removal of an external immobilization device. Results. Of these 16 patients, 1 sustained a complete injury, and 7 sustained an incomplete injury. Eight patients were neurologically intact. Twelve patients sustained a stable Jefferson fracture and were treated with a rigid cervical collar (Miami‐J collar [Jerome Medical, Moorestown, NJ]) alone from 10 to 12 weeks. The patient sustaining the complete neurologic injury died of multisystem trauma. All 15 live patients showed no instability on their follow‐up plain radiographs before the removal of an external stabilization device. Six patients underwent further plain radiographs approximately 1 year after the fracture and similarly demonstrated no instability. Conclusions. Isolated stable burst fracture of the atlas can be treated effectively with a rigid cervical collar alone for 10 to 12 weeks with good neurologic recovery and segmental stability. Unstable Jefferson fractures with concurrent unstable fracture of other cervical vertebrae, especially C2, requires surgical stabilization.

Unique Clinical Presentation of Pediatric Shunt Malfunction

Introduction: A cerebrospinal fluid (CSF) shunt is the primary treatment for most etiologies of hydrocephalus in the pediatric population. Malfunction of the shunt may present with unique symptoms and signs. This retrospective review investigates the presenting signs and symptoms of pediatric patients with shunt malfunction. Clinical Material and Method: One-hundred-and-thirty CSF diversion procedures were performed at two affiliated pediatric hospitals over a 2-year period. Seventy consecutive cases of CSF shunt revision were reviewed. These 70 operations were performed on 65 patients. Their medical records and radiographic studies were reviewed, and supplemented with a telephone interview to obtain a minimum of 3 months follow-up. Results: The 65 patients’ age ranged from 3 months to 16 years. The original etiology of the hydrocephalus was Chiari II malformation in 17, idiopathic in 15 and intraventricular hemorrhage in 10, neoplasm in 8 patients and meningitis in 5 patients. The most frequent presenting symptoms were headache (39 admissions), nausea/vomiting (28) and drowsiness (21). Seven Chiari patients (41%) presented with neck pain, 2 (12%) presented with lower cranial nerve palsy, and 2 (12%) presented with symptomatic syrinx, complaints not reported by non-Chiari patients (p < 0.01, χ2 analysis). Four myelodysplastic patients presented with a new-onset or recurrent seizure episode, which was significantly more frequent than in nonmyelodysplastic patients (p < 0.05, χ2 analysis). On examination, increased head circumference was noted in 17 patients. Parinaud’s syndrome was noted more prominently in patients with a history of intracranial neoplasm (4 of 8 cases) than in patients with nonneoplastic diseases (2 of 62 cases; p < 0.05, χ2 analysis). Other interesting presenting signs were pseudocyst (2), syringomyelia (2), hemiparesis (2) and Parkinson-like rigidity (2). Conclusion: Pediatric shunt malfunction generally presents with headache, nausea/vomiting, altered mental status, increased head circumference and bulging fontanelle. Other less frequent but unique presenting signs and symptoms, such as neck pain, syringomyelia and lower cranial nerve palsy in the myelodysplastic population, and Parinaud’s syndrome in patients with a history of intracranial neoplasm are frequently associated with shunt malfunction and should prompt a radiographic workup.

Surgical Treatment of Spinal Ependymoma and Post-Operative Radiotherapy

SummaryIntroduction. With the advances in microsurgical and monitoring techniques, spinal ependymomas are gross totally resected more frequently. The use of adjuvant radiotherapy has become questionable with gross total resection and its role for residual neoplasm need to be redefined. A retrospective analysis of a series of patients was carried out to investigate our clinical outcome and selected use of postoperative radiotherapy. Clinical materials and methods Between July 1990 and May 1995, nineteen patients [M:F=12:7; age range: 21 to 71 years] with a spinal ependymoma were treated at University of Miami by the senior author. (BAG)Pre-operative MRI diagnosed the intraspinal tumor, and pathology reports demonstrated that each patient had a histologically confirmed ependymoma. At the time of diagnosis, the most common symptoms presented were pain (in 16 patients=84.2%). The pattern of progression of clinical symptoms was directly related to the location of the tumor. Each patient had an MRI immediately after surgery, approximately 6 months post-operatively, and then annually. Results All 19 patients underwent intradural microsurgical exploration with an attempted gross total resection (achieved in 16 patients=79%) of the ependymoma through a posterior approach. Direct neural tissue stimulation halted further resection in 2 patients with questionable tumor margins. Radiation therapy was employed as a surgical adjunct in 3 patients (15.8%) because of possible residual tumor.All patients were followed up postoperatively for an average of 50.6 months (range 6 months to 6 years). All patients are surviving to date. Surgical resection of these tumors led to significant alleviation of pre-operative symptoms. There has been no radiographic evidence of tumor recurrence or growth in any patient to date. Conclusion. Surgical resection of spinal ependymoma leads to significant improvement of pre-operative symptom. Surgical removal alone, with an attempt to grossly resect the tumor, should be the treatment of choice, with careful clinical and radiographic follow-up. Radiation therapy should only be considered as a surgical adjunct where gross total resection is not achieved.

Diffuse axonal injury (DAI) is not associated with elevated intracranial pressure (ICP).

SummaryObjective. Traditionally, intracranial pressure (ICP) monitoring has been utilized in all patients with severe head injury (Glasgow coma score of 3–8). Ventriculostomy placement, however, does carry a 4 to 10 percent complication rate consisting mostly of hematoma and infection. The authors propose that a subgroup of patients presenting with severe head trauma and diffuse axonal injury without associated mass lesion, do not need ICP monitoring. Additionally, the monitoring data from ICP, MAP, and CPP for a comparison severe head injury group, and subgroups of DAI would be presented. Materials and methods. Thirty-six patients sustaining blunt head trauma and fitting our strict clinical and radiographic diagnosis of DAI were enrolled in our study. Inclusion criteria were severe head injury patients who did not regain consciousness after the initial impact, and whose CT scan demonstrated characteristic punctate hemorrhages of <10 mm diameter at the greywhite junction, basal ganglia, corpus callosum, upper brainstem, or a combination of the above. Patients with significant mass lesions and documented anoxia were excluded. Their intracranial pressure (ICP) and cerebral perfusion pressure (CPP) were compared to a control group of 36 consecutive patients with severe non-penetrating non-operative head injury, using the Analysis for Variance method. Results. Eighteen (50.0%), six (16.7%), and twelve (33.3%) patients had types I, II, and III DAI, respectively. The admission Glasgow Coma Score (GCS) was higher for types I and II than for type III DAI. ICP was monitored from 23 to 165 hours, with a mean ICP for 36 patients of 11.70 mmHg (SEM=75) and a range from 4.3 to 17.3 mmHg. Of all ICP recordings, of which 89.7% (2421/2698) were ≤20 mmHg. Average mean arterial pressure (MAP) was 96.08 mmHg (SEM=1.69), and 94.6% (2038/2154) of all MAP readings were greater than 80 mmHg. Average cerebral perfusion pressure (CPP) was 85.16 mmHg (SEM=1.68), and 90.1% (1941/2154) of all CPP readings were greater than 70 mmHg. This is compared to the control group mean ICP, MAP, and CPP of 16.84 mmHg (p=0.000021), 92.80 mmHg (p=0.18), and 76.49 mmHg (p=0.0012). No treatment for sustained elevated ICP>20 mmHg was needed for DAI patients except in two; one with extensive intraventricular and subarachnoid hemorrhage who developed communicating hydrocephalus, and another with ventriculitis requiring intrathecal and intravenous antibiotic treatments. Two complications, one from a catheter tract hematoma, and another with Staph epidermidis ventriculitis, were encountered.All patients, except type III DAI, generally demonstrated marked clinical improvement with time. The outcome, as measured by Glasgow Coma Score (GCS) and Glasgow Outcome Score (GOS) was similarly better with types I and II than type III DAI. Conclusion. The authors conclude that ICP elevation in DAI patients without associated mass lesions is not as prevalent as other severe head injured patients, therefore ICP monitoring may not be as critical. The presence of an ICP monitoring device may contribute to increased morbidity. Of key importance, however, is an accurate clinical history and interpretation of the CT scan.

Safety and Stability of Open-door Cervical Expansive Laminoplasty

Cervical expansive laminoplasty has been utilized for over 20 years. This retrospective analysis investigated the safety and incidence of postoperative instability of patients undergoing a modified expansive laminoplasty. One hundred five patients underwent a modified cervical expansive laminoplasty and had at least a 6-month follow-up. These 105 patients were followed for a mean of 18.6 months (range 6-89 months). All patients remained in a rigid cervical collar for 8-12 weeks after the laminoplasty. Postoperative cervical plain radiographs were obtained on postoperative day 1, an average of 9.6 weeks postoperatively (range 8-12 weeks) and an average of 10.1 months (range 6-12 months) postoperatively. Thirty-five patients underwent further radiographs >18 months postoperatively. Laminoplasty was performed in 82 patients with progressive cervical spondylotic myelopathy, 4 patients with ossification of the posterior longitudinal ligament, 7 patients for posterior approach to a cervical neoplasm, and 12 patients for early posttraumatic decompression. The canal/vertebral body ratio showed a significant increase from 0.78 to 1.02 (paired t test, p < 0.001). Postoperatively, no incidence of graft dislodgement or segmental instability was diagnosed in any patient. Modified open-door expansive laminoplasty is an effective way of expanding the spinal canal. Its associated low incidence of postoperative instability and kyphotic deformity should make this procedure a desirable substitute for cervical laminectomy under many circumstances.

Early MRI findings in high grade glioma

Magnetic resonance imaging (MRI) is more sensitive than computerized tomography in the detection of many intracerebral lesions; however, the significance of some MRI findings may be unclear. Over four years, nine patients, aged 40–79 years, have been encountered whose initial MRI scans were negative or had minimal abnormalities and soon thereafter had high grade glioma. Initial MRI was performed in eight patients for new-onset seizures and one patient for a focal deficit. MRI was negative in four of the patients and mildly abnormal in five of the patients (small areas of increased T2 and/or minimal enhancement). The initial diagnoses usually included inconclusive differentials of stroke and infection with neoplasm less frequently considered. Radiographic progression leading to the diagnosis of high grade glioma became evident on repeat MRI in 1–8 months with six patients showing progression within three months. All patients underwent surgery and had histologic diagnosis of glioma. Although MRI is quite sensitive, four of the initial scans were negative with reasonable quality studies. Conversely, in five of the initial scans, the tumors were detected when so small that the radiographic findings were not typically diagnostic. Glioma must be considered as a possible cause of initial seizures or new neurologic deficits in adults with normal or minimally abnormal MRI. In this group, seizures were the overwhelming hallmark of presentation. In such a clinical situation, close follow-up with short interval repeat MRI should be performed.

Spinal metastases of malignant intracranial meningioma

BACKGROUND Spinal metastasis of intracranial meningiomas has rarely been reported. Three out of ten previously reported cases of malignant meningioma metastasizing to the spine had undergone surgical debulking with no neurological improvement. The authors retrospectively reviewed the treatment course of three patients with malignant meningioma metastasizing to the spine who underwent early magnetic resonance imaging (MRI) and radiotherapy without surgical debulking. CASE DESCRIPTION Three patients with intracranial malignant meningiomas underwent multiple resections of intracranial lesions, and developed spinal intradural metastases an average of 64 months (range, 27-102 months) from their initial presentation. All three patients had at least two operations for recurrent intracranial tumors. All had localized back pain with motor weakness, and MRI scans demonstrated spinal involvement. No surgical exploration was performed for the spinal lesions; rather, all patients received steroids and radiotherapy for the spinal lesions. All three patients improved neurologically after the steroid and radiation treatments, and went on to survive from 3 to 18 months. CONCLUSION Early MRI should be performed in patients with spinal symptoms and signs after the treatment of intracranial meningiomas. Radiotherapy is an effective palliative treatment for spinal metastases.

Third ventricular glioblastoma multiforme: Case report

Intraventricular neoplasms of the third ventricle represent a small minority of intracranial tumors. Furthermore, few cases of solitary, third ventricular glioblastoma multiforme were reported in the literature. The authors report a case of a fifty-nine year olf man who presented with progressive gait disturbance, disorientation, diffuse headache, and vomiting. Computed tomography and magnetic resonance imaging (MRI) confirmed a solitary third ventricular ringenhancing lesion and obstructive hydrocephalus. The final pathology of the lesion upon permanent section was glioblastoma multiforme (GBM). The clinical course, radiologic findings, the possible origin of this lesion, as well as possible treatment regimen are discussed.