Spine and Spinal Cord Surgery

Spring 2001
Volume 12, Number 1

 Next Article  

Originally used to produce tiny lesions in the brain, stereotactic radiosurgery (SRS) now describes the precise localization and delivery of a high dose of radiation to any defined anatomic target. SRS is rapidly becoming a useful treatment modality for spinal neoplasms and vascular malformations of the spinal cord. At CNI, neurosurgeons and radiation oncologists are increasingly optimistic about the future of stereotactic radiosurgery as an alternative treatment for tumors and AVMs of the spine. This article outlines the historical development and present status of this exciting new treatment modality.

Introduction. Although developed originally for use in the brain or other intracranial targets, stereotactic radiosurgery is rapidly becoming a useful treatment modality for tumors and vascular malformations of the spine and spinal cord. Radiosurgery is a term attributed to Lars Leksell, a Swedish neurosurgeon who was instrumental in the development of precise localization and delivery of radiation to an intracranial target. Originally, it was used by neurosurgeons to produce tiny, circumscribed, non-invasive lesions in the brain for movement disorders and pain. Stereotaxis refers to the precise localization of an anatomic target in 3-dimensional space, and radiosurgery is a term that now describes any precise, tightly circumscribed delivery of radiation, usually as a high dose in a single fraction, to a defined anatomic target. Multiple fractions may be used, a procedure termed stereotactic radiotherapy. Together, they are an important means of treating both benign and malignant tumors as well as vascular malformations located near critical radiosensitive structures, such as optic nerve or spinal cord. Utilization of this treatment modality outside the head has been quite limited so far, but many of the technological hurdles are being overcome, and the future of stereotactic spinal radiosurgery is rapidly approaching.

Most spinal stereotactic radiosurgery utilizes a linear accelerator (LINAC) as the radiation source. Used stereotactically, this device employs multiple intersecting radiation arcs that produce a roughly spherical dose distribution, or isocenter. Multiple isocenters can be used together to shape a dose field to roughly match the shape of a target. Stereotactic radiosurgical planning software topographically contours the prescription (or treatment) isodose level conformally to the targeted tissue while allowing the neurosurgeon and radiation oncologist to exclude nearby structures, such as spinal cord from high doses of radiation. Intensity modulated radiotherapy (IMRT) uses precision multileaf collimation to modulate or shape the beam to conform to a target profile with even greater accuracy The radiation can be delivered in a single large dose, or divided into multiple fractions. Single fraction therapy may have some practical and radiobiological advantages, but probably has higher neurotoxicity, whereas fractionated therapy has the disadvantage of requiring reproducible target fixation and localization over multiple treatments. Nevertheless it has its own set of biological advantages, such as improving tissue tolerance of nearby critical radiosensitive structures and allowing higher overall doses.

History. As early as 1969, Hitchcock described a device that allowed precise localization of spinal lesions based on anatomical landmarks.1 The device was difficult to use, variably accurate and never widely utilized. The subsequent explosion of frameless stereotaxis and image-based surgical guidance systems for the brain led to attempts to use these systems in the spine. The functionality of these devices proved to be limited because of difficulty rigidly fixating the flexible spine and precise and sustainable anatomic localization. Patel, et al, evaluated the potential role of an articulated stereotactic guidance system designed for intracranial use in the localization and minimally invasive excision of a sacral osteoblastoma.2 Their assessment of the limitations encountered in extracranial use of such localization systems was informative, and provided a glimpse into the technical difficulty encountered in stereotactic delivery of radiation to the spine.

A prototype spinal fixation frame for accurate stereotactic localization and linear accelerator (LINAC) based radiosurgical treatment of spinal targets was developed by Hamilton, et al, in 1995.3 The system was designed to employ spinal or skeletal fixation to immobilize the area of interest and then encircle the targeted region with a traditional orthogonal, 3-axis coordinate system. During the initial assessment of the device, radiolucent calibration targets and CT scan imaging demonstrated a mean localization error of less than one millimeter. Using the LINAC to irradiate these same targets delivered an overall radiation treatment accuracy ranging from 1.4 to 2.0 mm. The authors concluded that extracranial stereotactic radiosurgery was technically feasible and that the accuracy of treatment utilizing osseous fixation would be acceptable for clinical treatment. The authors subsequently delivered stereotactic radiosurgery using a modified linear accelerator to metastatic neoplasms in the cervical, thoracic, and lumbar regions in 5 patients.4 In all patients, the neoplasms had failed to respond to spinal cord tolerance doses delivered by standard external fractionated radiation therapy to a median dose of 45 Gy (range, 33-65 Gy/11-30 fractions). The tumors were treated with single-fraction stereotactic radiosurgery using the spinal stereotactic frame for immobilization, localization, and treatment. The median number of isocenters was one (range, 1 to 5) with a median single fraction dose of 10 Gy (range, 8-10 Gy) with median normalization to 80% isodose contour (range, 80%-160%). There was a single complication of esophagitis from radiosurgery of a tumor involving the C6 to T1 segments, which resolved with medical therapy. Median follow-up in this group of patients was 6 months (range, 1 to 12 mo), with no radiographic or clinical progression of the treated tumor in any patient. Two patients died from systemic metastatic disease. In the 3 surviving patients, there was computed tomographic- or magnetic resonance-documented regression of the treated tumor with a decrease of thecal sac compression with a median follow-up of 6 months (range, 3 to 14 mo). These 5 patients represented the first clinical application of stereotactic radiosurgery in the spine. The results suggested that extracranial radiosurgery could be used for the treatment of paraspinal neoplasms, even after external fractionated radiation therapy has failed.

The authors reported a second series on the use of this externally fixed spinal stereotactic radiosurgery frame employed for the treatment of 9 patients who presented with recurrent spinal neoplasms.5 All patients had failed standard therapy consisting of surgery, external fractionated radiation therapy, and/or chemotherapy. Eight of the lesions represented metastatic tumors in the vertebral column, one of the lesions was a primary osteosarcoma involving multiple vertebral bodies. The lesions were found at multiple levels, from the cervical through the sacral region. Six out of the 9 patients presented with epidural compression: 4 of the 9 patients with evidence of myelopathy: 2 of the 9 patients with radicular symptoms secondary to compression from the tumor, and 1 patient was free of any compressive symptoms. All patients had pain requiring narcotics. Patients were treated with a median radiosurgical dose of 800 cGy (range 800-1.000) with a median of 1 isocenter (range 1-7 isocenters) and median normalization of 80% to the isodose contour (range 80-160). Median dose delivered to the already prior irradiated spinal cord was 179 cGy (range 52-320 cGy) with a median spinal cord dose of 34 (range 4-68). No major complications were reported. Five of the 9 patients died during the follow-up period, all from causes unrelated to the spinal radiosurgery. Three out of the 9 patients have been followed for more than 1 year. In all 3, there was radiographic regression of the tumor and of epidural compression. In 2 patients, there was histologic confirmation of absence of tumor in the treated site. One patient who died of unrelated causes was found to be tumor free 12 months after treatment. Long-term outcomes of frame based stereotactic spinal radiosurgery from the first 19 patients treated compare favorably to conventional radiotherapy.6

Devices. The utilization of a rigid osseous fixation system described above represents only one method of solving the twin problems of precision localization and immobilization. These problems are compounded if fractionation is being employed. Externalized spine fixation used in combination with a linear accelerator has the advantage of superb accuracy and demonstrated reproducibility. Respiration induced target drift in spinal stereotactic radiosurgery using skeletal fixation was evaluated in a porcine model, and the results suggest this type of system may indeed be the most reproducibly accurate.14 However, it has the distinct disadvantage of requiring an invasively fixed frame for immobilizing the region of interest. Accordingly, LINAC-based stereotactic radiosurgery has been used with other non-invasive localization and immobilization schemes. These techniques vary significantly in the degree to which reproducibility and accuracy have been confirmed. It is very important that more than simple mechanical accuracy is examined, and that actual dose distributions are measured and compared to the theoretical and anticipated distributions based on the radiosurgical plan. Not all systems commercially available have applied these standards.

The majority of systems being developed for fractionated stereotactic radiotherapy rely on molded couches and orthogonal x-ray films to reproduce and confirm positioning. However, the problem of guaranteeing reproducible accuracy over multiple treatments remains. One possible solution involves the use of implantable fiducials which allow topographic localization using orthogonal plane x-ray films. These fiducials can be surgically or percutaneously implanted in nearby stable vertebra, or even injected directly into tumor at the time of needle biopsy. Multiple fiducials are required to confirm positioning in 3 dimensions. The use of ultrasound for “real-time” imaging of bony structures is being explored with the potential advantage of imaging target shifts induced by respiratory movement.

Lohr, et al,13 evaluated the accuracy that can be achieved with noninvasive patient fixation based on a body cast attached to an external stereotactic body frame during fractionated extracranial stereotactic radiotherapy. They evaluated setup accuracy in 31 CT studies (20 or more slices 3 mm thick) from 5 patients immobilized in a body cast attached to a stereotactic body frame for treatment of paramedullary tumors of the thoracic or lumbar spine. The immobilization device consisted of a customized wrap-around body cast that extended from the neck to the thighs and a separate head mask. Each CT study was performed immediately before or after every second or third actual treatment fraction without repositioning the patient between CT and treatment. The stereotactic localization system was mounted and the isocenter as initially located stereotactically was marked with fiducials for each CT study. Deviation of the treated isocenter as compared to the planned position was measured in all 3 dimensions. Mean patient movements of 1.6 mm+/-1.2 mm (laterolateral [LL]), 1.4 mm+/-1.0 mm (anterior-posterior [AP]), 2.3 mm+/-1.3 mm (transversal vectorial error [VE]) and less than the 3 mm CT slice thickness (craniocaudal [CC]) were recorded for the targets in the thoracic spine and 1.4 mm+/- 1.0 mm (LL), 1.2 mm+/-0.7 mm (AP), 1.8 mm+/-1.2 mm (VE), and < 3 mm (CC) for the lumbar spine. The worst case deviation was 3.9 mm for the first patient with the target in the thoracic spine (in the LL direction). Combining those numbers (mean transversal VE for both locations and maximum CC error of 3 mm), the mean 3-dimensional vectorial patient movement (and thus the mean overall accuracy) could be estimated to be less than 3.6 mm. The authors concluded that the combination of a body cast and head mask system in a rigid stereotactic body frame ensured reliable noninvasive patient fixation for fractionated extracranial stereotactic radiotherapy. Although the methods and conclusions can be challenged, the authors do make the important point that adequate accuracy and reproducibility would enable dose escalation for less radioresponsive tumors near spinal cord or other critical locations while minimizing the risk of radiation induced complications.

Rather than using rigid frame immobilization, a unique instrument for performing frameless stereotactic radiosurgery uses a robotic image-guided radiosurgical system that relies on a real time radiographic image-to-image correlation algorithm for target localization.7, 8 The system utilizes a novel, lightweight, high-energy radiation source mounted on a precision robotic arm. Since multiple individually targeted “pencil-beams” are used instead of arcs, the treatment isodose contour takes shape without using individual isocenters, and could theoretically be planned to exclude critical structures entirely. This instrument has several distinct advantages over frame-based systems, including improved patient comfort, increased treatment degrees of freedom, and the potential to more easily target extracranial lesions. The system has been used to treat an arteriovenous malformation in the cervical spine, a recurrent schwannoma of the thoracic spine, a metastatic adenocarcinoma of the lumbar spine, and several pancreatic cancers.

The device has been modified to treat extracranial sites by using implanted radio-opaque fiducials and vertebral landmarks to locate treatment targets. During each treatment, the image guidance system monitors the position of fiducials and target site and relays target coordinates to the beam-pointing system at discrete time intervals. The pointing system then dynamically aligns the beam with the lesion, automatically compensating for shifts in target position. Breathing- related motion is managed if necessary by coordinating beam gating with breath-holding by the patient. The system can maintain radiographic alignment with spine lesions to within +/- 0.2 mm on average9 If independent dosimetry measurements confirm that treatment delivery is reliable and that maximum variance is acceptable, the device may represent a revolutionary method of radiation delivery to spinal neoplasms.

The role of stereotactic radiosurgery as in the management of spinal neoplasms and vascular malformations is examined in a review article by Hamilton, et al.12 The authors outline specific problems in the application of stereotactic radiosurgery to the spine, and discuss the 3 techniques for spinal stereotaxis outlined above: bone screw fixation, contour mold fixation, and frameless stereotaxis.

Indications. Any spinal neoplasm or vascular malformation that has traditionally been treated with conventional radiotherapy can potentially be treated with stereotactic radiosurgery. The leeway to use higher doses near critical structures opens the possibility of treating more radio-resistant neoplasms or supplying boost doses to previously irradiated fields. Small focal spinal neoplasms are ideal targets for these techniques. Stereotactic radiosurgery is being used successfully to treat surgically inaccessible or multiple hemangioblastomas, such as those found in von Hippel-Lindau disease. Chang, et al, retrospectively reviewed their experience of 29 hemangioblastomas in 13 patients with von Hippel-Lindau disease treated from 1989 to 1996 with linear accelerator-based radiosurgery.11 The mean patient age was 40 years (range, 31 to 57 yr). The radiation dose to the tumor periphery averaged 23.2 Gy (range, 18-40 Gy). The mean tumor volume was 1.6 cm3 (range, 0.07-65.4 cm3). Tumor response was evaluated in serial, contrast-enhanced, computed tomographic and magnetic resonance imaging scans. The mean follow-up period was 43 months (range, 11-84 mo). Only one (3%) of the treated hemangioblastomas progressed. Five tumors (17%) disappeared, 16 (55%) regressed, and 7 (24%) remained unchanged in size. Five of 9 patients with symptoms referable to treated hemangioblastomas experienced symptomatic improvement. During the follow-up period, one patient died as a result of progression of untreated hemangioblastomas in the cervical spine. Three patients developed radiation necrosis at treated sites, 2 of whom were symptomatic. The report concluded that stereotactic radiosurgery provides a high likelihood of local control of these tumors and is an attractive alternative to multiple surgical procedures for patients with von Hippel-Lindau disease.

Conclusion. Conventional radiation therapy may not always be the best alternative for spinal neoplasm or vascular malformation. In 1999, Schuller, et al, reviewed their 35 year experience in treating ependymomas with surgical resection followed by conventional radiotherapy.10 They questioned the value of large-field techniques of whole brain and spinal irradiation, and concluded that stereotactic radiotherapy might be able to deliver a higher tumor dose without increasing toxicity to the spinal cord. They postulated this may be particularly true if IMRT is used to exclude the spinal cord and fractionation is used to reduce neurotoxicity. Their experience with traditional radiotherapy serves as a call to improve on a past marked by frustrating failure and unacceptable complications. At the CNI, neurosurgeons and radiation oncologists are increasingly optimistic about the future of stereotactic radiosurgery as an alternative treatment of spinal neoplasms and vascular malformations.

Address comments and questions to:

John H. McVicker, MD
701 E. Hampden Avenue
Suite 510
Englewood, CO 80113

 Next Article