Technical and clinical experience with spine radiosurgery: a new technology for management of localized spine metastases

This study is to demonstrate the technical and clinical experience of applying image guided spinal radiosurgery for treatment of localized spinal metastasis. A dedicated shaped beam radiosurgery unit with intensity modulated radiotherapy (IMRT) and x-ray based image-guided radiotherapy (IGRT) were used for the radiosurgery procedure. A total of 196 patients with 270 lesions of spinal metastases were treated with this procedure from May 2001 to October 2005. All patients received single dose radiosurgery to the involved spine only. The radiosurgery dose was escalated from 10 to 18 Gy in 2 Gy increments. The technical experience using IMRT planning and IGRT implementation has been summarized. Clinical results reporting pain relief responses have been analyzed for the first 49 patients treated with this procedure. For IMRT treatment planning, seven posterior/oblique fields were generally used for spinal radiosurgery as the optimal setup to balance conformality versus complexity. A criterion of 10 Gy to 10% of the adjacent spinal cord volume has been met with satisfactory target dose coverage for most of the cases. When the spinal cord dose exceeded this constraint, the tumor coverage was somewhat compromised. Accurate target localization has been achieved for all patients using the x-ray image-guided system. The preliminary clinical results have demonstrated that pain response was achieved in 85% of patients, with neurological improvement in patients with spinal cord compression. Patients tolerated the treatment well without major acute toxicities. Image guided spinal radiosurgery can be successfully applied to treat patients with focal spine metastases.     

Partial volume tolerance of the spinal cord and complications of single-dose radiosurgery

BACKGROUND: Spine radiosurgery causes a rapid dose fall-off within the spinal cord. The tolerance of partial volume of the spinal cord may determine the extent of clinical application. The study analyzed the partial volume tolerance of the human spinal cord to single fraction radiosurgery. METHODS: A total of 230 lesions with spine metastases in 177 patients were treated with radiosurgery with single fraction of 8 to 18 Gy, prescribed to the 90% isodose line that encompassed the target volume. Spinal cord volume was defined as 6 mm above and below the radiosurgery target volume. Spinal cord dose was calculated from the radiation dose/spinal cord volume histogram and correlated with clinical/neurological status and radiographic studies. Median follow-up was 6.4 months (range, 0.5-49 months). The 1-year survival rate was 49%. RESULTS: The average spinal cord volume defined at the treated spinal segment was 5.9 +/- 2.2 mL. The average dose to the 10% spinal cord volume was 9.8 +/- 1.5 Gy, calculated from the dose-volume histogram in the group of 18 Gy prescribed dose. The spinal cord volume that received higher than 80% of the prescribed dose was 0.07 +/- 0.10 mL, which represented 1.3 +/- 1.8% of the cord volume. Among the 86 patients who survived longer than 1 year there was 1 case of radiation-induced cord injury after 13 months of radiosurgery. There were no other cases of spinal cord sequelae. CONCLUSIONS: Whereas the maximum spinal cord tolerance to single-dose radiation is not known, partial volume tolerance of the human spinal cord is at least 10 Gy to 10% of the spinal cord volume defined as 6 mm above and below the radiosurgery target.     

Central nervous system complications by malignant lymphomas: radiation schedule and treatment results

During a 10-year period, 28 patients with spinal cord compression due to epidural malignant lymphoma and 47 patients with cerebral involvement of lymphoma were treated with radiation at our institution. Fifty-four percent of the patients with spinal cord compression had this complication at the time of initial presentation of the disease, whereas only 4% with cerebral involvement presented with CNS symptoms. Only one patient had primary lymphoma solely located in the brain. Characteristically, a majority of the patients with spinal cord compression complained of back pain several months before developing neurological symptoms. Because only one-third of the patients had positive spine roentgenograms at the time of spinal cord compression, a CT scan is suggested in patients with malignant lymphoma suffering from back pain in order to verify a paraspinal lymphoma. Thus spinal cord compression may be avoided by early diagnosis and treatment. Among the patients with spinal cord compression, Hodgkin's and non-Hodgkin's histology were equally represented, whereas only 6% had Hodgkin's lymphoma among the patients with cerebral involvement of lymphoma. The response to treatment defined as improvement in neurological deficit in the patients with spinal cord compression was approximately 90% in both the Hodgkin's and the non-Hodgkin's group. No difference in response was found among patients who had laminectomy compared to patients who did not. Patients receiving high dose, short-term treatment (5 Gy X 5-6) responded equally to patients receiving low dose, long-term treatment (2 Gy X 18-20). The median survival from initiation of radiation therapy in patients developing spinal cord compression or cerebral involvement during relapse was 30 months. In patients with spinal cord compression at initial presentation of the disease, median survival had not been reached after 5 years. Among patients with cerebral involvement 50% had improvement of neurological symptoms with no difference between patients receiving high dose, short-term and patients receiving low dose, long-term treatment. It is concluded that high dose, short-term irradiation is as effective as low dose treatment. Especially in patients with neurological complications at relapse, this treatment schedule is preferred because of the extremely short survival of these patients.     

Radiosurgery for the treatment of spinal lung metastases

BACKGROUND: Spinal metastases are a common source of pain as well as neurologic deficit in patients with lung cancer. Metastases from lung cancer traditionally have been believed to be relatively responsive to radiation therapy. However, conventional external beam radiotherapy lacks the precision to allow delivery of large single-fraction doses of radiation and simultaneously limit the dose to radiosensitive structures such as the spinal cord. The current study evaluated the efficacy of single-fraction radiosurgery for the treatment of spinal lung cancer metastases. METHODS: In the current prospective cohort evaluation, 87 lung cancer metastases to the spine in 77 patients were treated with a single-fraction radiosurgery technique with a follow-up period of 6 to 40 months (median, 12 months). The indication for radiosurgery treatment was pain in 73 cases, as a primary treatment modality in 7 cases, for radiographic tumor progression in 4 cases, and for progressive neurologic deficit in 3 cases. RESULTS: Tumor volume ranged from 0.2 to 264 cm(3) (mean, 25.7 cm(3)). The maximum tumor dose was maintained at 15 to 25 grays (Gy) (mean, 20 Gy; median, 20 Gy). No radiation-induced toxicity occurred during the follow-up period. Long-term axial and radicular pain improvement occurred in 65 of 73 patients (89%) who were treated primarily for pain. Long-term radiographic tumor control was observed in all patients who underwent radiosurgery as their primary treatment modality or for radiographic tumor progression. CONCLUSIONS: Spinal radiosurgery was found to be feasible, safe, and clinically effective for the treatment of spinal metastases from lung cancer. The results of the current study indicate the potential of radiosurgery in the treatment of patients with spinal lung metastases, especially those with solitary sites of spine involvement, to improve long-term palliation.     

Stereotactic radiosurgery and hypofractionated stereotactic radiotherapy: normal tissue dose constraints of the central nervous system

Single-fraction stereotactic radiosurgery (SRS) and hypofractionated stereotactic radiotherapy (SRT) are radiation planning and delivery techniques used for the treatment of intracranial and spine/spinal cord tumors and targets. For cranial SRS and SRT, critical normal tissues/structures include the brainstem, cranial nerves, cochlea and normal brain parenchyma. For spine SRS/SRT, critical normal tissues/structures include the spinal cord, cauda equina as well as neighboring organs. This paper reviews clinical studies investigating central nervous system dose tolerances after cranial or spinal SRS/SRT. The impact of dose, volume, fractionation, and other relevant clinic-pathologic variables are discussed, as are limitations of the published data.     

Repositioning accuracy of a commercially available thermoplastic mask system.

BACKGROUND AND PURPOSE: To evaluate the repositioning accuracy of a commercially available thermoplastic mask system for single dose radiosurgery treatments and fractionated treatment courses. PATIENTS AND METHODS: The repositioning accuracy of the Raycast-HP mask system (Orfit Industries, Wijnegem, Belgium) was analyzed. Twenty-two patients that were treated by intensity-modulated radiation therapy (IMRT) or intensity modulated radiosurgery (IMRS) for 43 intracranial lesions, underwent repeated CT imaging during their course of treatment, or as a positional control immediately before radiosurgery. We evaluated multiple anatomical landmark coordinates and their respective shifts in consecutive repeated CT-controls. An iterative optimization algorithm allowed for the calculation of the x, y and z-components of translation of the target isocenter(s) for each repeated CT, as well as rotation in the respective CT data sets. In addition to absolute target isocenter translation, the total magnitude vector (i.e. sum-vector) of isocenter motion was calculated along with patient rotations about the three principle axes. RESULTS: Fifty-five control CT datasets were analyzed for the target isocenter's respective position relative to the original treatment planning CT simulation. Mean target isocenter translation was 0.74+/-0.53, 0.75+/-0.60 and 0.93+/-0.78 mm in x, y and z-directions, respectively. Mean rotation about the x, y and z-axes was 0.67+/-0.66, 0.61+/-0.63 and 0.67+/-0.61 degrees, respectively. The respective median and mean magnitude vectors of isocenter translation were 1.28 and 1.59+/-0.84 mm. Analysis of the accuracy of the first setup control, representative of setup accuracy for radiosurgery treatments, compared with setup accuracy throughout a fractionated radiation treatment course were statistically equivalent (P= 0.15) thus indicating no measurable deterioration of setup accuracy throughout the treatment course. CONCLUSIONS: The analyzed Orfit thermoplastic mask system performed favorably compared with other mask immobilization systems for which peer-reviewed repositioning data exist. While the performance of the system for fractionated treatment courses was considered to be excellent, use of this mask system for radiosurgery immobilization in our clinic is subject to additional quality assurance measures to prohibit the delivery of treatments with target dislocations larger than 2 mm. The measured data in the present study should enable the users of this system to assign appropriate margins for the generation of planning target volumes.     

Radiosurgical decompression of metastatic epidural compression

BACKGROUND:

Surgical decompression of metastatic epidural compression (MEC) improved ambulatory function. Spine radiosurgery can accurately target the epidural tumor and deliver high radiation doses for tumor control. Therefore, a clinical trial was performed to quantitatively determine the degree of epidural decompression by radiosurgery of metastatic epidural compression.

METHODS:

Sixty‐two patients with a total of 85 lesions of metastatic epidural compression were treated. Epidural compression was diagnosed by magnetic resonance imaging (MRI) scans. Main criteria of inclusion were neurological status with muscle power 4 of 5 or better. Radiosurgery was performed to the involved spine segment, including the epidural mass with median dose of 16 Gy (range 12‐20 Gy) in a single session. All patients had prospective clinical follow‐up, ranging from 1‐48 months (median 11.5 months), and 36 patients had pretreatment and post‐treatment imaging, ranging from 2‐33 months (median 9.3 months). Primary endpoints were epidural tumor control and thecal sac decompression.

RESULTS:

The mean epidural tumor volume reduction was 65 ± 14% at 2 months after radiosurgery. The epidural tumor area at the level of the most severe spinal cord compression was 0.82 ± 0.08 cm² before radiosurgery and 0.41 ± 0.06 cm² after radiosurgery (P < .001). Thecal sac patency improved from 55 ± 4% to 76 ± 3% (P < .001). Overall, neurological function improved in 81%.

CONCLUSIONS:

This study demonstrated a radiosurgical decompression of epidural tumor. Although neurosurgical decompression and radiotherapy is the standard treatment in patients with good performance, radiosurgical decompression can be a viable noninvasive treatment option for malignant epidural compression. Cancer 2010. © 2010 American Cancer Society.     

IMRT dose escalation for positive para-aortic lymph nodes in patients with locally advanced cervical cancer while reducing dose to bone marrow and other organs at risk

PURPOSE: To evaluate the feasibility of dose-escalated para-aortic lymph node (PALN) intensity modulated radiation therapy (IMRT) in reducing the dose to bone marrow, bowel, spinal cord, and kidneys, compared with conventional radiation techniques of PALNs in patients with locally advanced cervical cancer and PALN metastases. METHODS AND MATERIALS: Computed tomography scans and MRI studies of 5 cervical cancer patients with PALN involvement were transferred to an IMRT treatment planning workstation (Eclipse/Helios) for image fusion and definition of target volumes and critical structures. The positive PALNs identified on fused computed tomography-magnetic resonance images were defined as the gross target volume (GTV), and the PALN region was defined as the planning target volume. There were 2 distinct treatment regions: the PALN region superior to the isocenter and the whole-pelvis region inferior to the isocenter. Three treatment planning techniques were compared: AP/PA (both regions), 4-field box (both regions), and PALN-IMRT with 4-field box to the whole-pelvis field. With IMRT, the radiation dose to the GTV was escalated from the conventional 45 Gy to 60 Gy (2.4 Gy/fraction), whereas the planning target volume and whole-pelvis region received 45 Gy. The treatment planning isocenter was placed at the L4-L5 vertebral body interspace, and this allowed the two treatment regions to be abutted using independent jaws. RESULTS: This study has demonstrated the feasibility of escalating the dose delivered to grossly positive PALNs to 60 Gy (2.4 Gy/fraction) with a 95.6% median GTV coverage, concomitantly with conventional treatment of the whole-pelvis region. PALN-IMRT significantly reduced V(40) bone marrow compared to the AP/PA and 4-field box techniques with a median of 21.3%, 98%, and 49.7%, respectively. The PALN-IMRT and 4-field box techniques showed a reduction in V(45) bowel over the AP/PA technique, but a level of statistical significance was not reached. The spinal cord received a significantly higher maximum dose when PALNs were treated with AP/PA fields. Alternatively, the use of the 4-field box technique yielded a significant increase in V(22) kidney on both sides. The placement of the treatment planning isocenter at the L4-L5 interspace allowed the PALN-IMRT and whole-pelvis regions to be treated with a relatively uniform dose at the abutment region. CONCLUSION: In this dosimetric analysis, we demonstrated that dose-escalated PALN-IMRT with conventional whole-pelvis radiotherapy is feasible with significant sparing of critical normal structures compared to PALN conventional radiation techniques.     

Clinical use of dual image-guided localization system for spine radiosurgery

The recently released Novalis TX linac platform provides various image guided localization methods including a stereoscopic X-ray imaging technique (ExacTrac) and a volumetric cone beam computed tomography (CBCT) imaging technique. The ExacTrac combined with the robotic six dimensional (6D) couch provides fast and accurate patient setup based on bony structures and offers "snap shot" imaging at any point during the treatment to detect patient motion. The CBCT offers a three dimensional (3D), volumetric image of the patient's setup with visualization of anatomic structures. However, each imaging system has a separate isocenter, which may not coincide with each other or with the linac isocenter. The aim of this paper was to compare the localization accuracy between Exactrac and CBCT for single fraction spine radiosurgery treatments. The study was performed for both phantom and patients (96 clinical treatments of 57 patients). The discrepancies between the isocenter between the ExacTrac and CBCT in four dimensions (three translations and one rotation) were recorded and statistically analyzed using two-tailed t-test.     

Phase I clinical evaluation of near-simultaneous computed tomographic image-guided stereotactic body radiotherapy for spinal metastases

PURPOSE: To evaluate in a Phase I study the safety, feasibility, and patient-positioning accuracy of treating patients with intensity-modulated, near-simultaneous, computed tomographic (CT) image-guided stereotactic body radiotherapy (SBRT). PATIENTS AND METHODS: Fifteen consecutive patients with metastatic spinal disease who met protocol eligibility criteria were entered into a Phase I clinical trial. Each patient received five treatments of intensity-modulated, near-simultaneous CT image-guided SBRT, for a total of 75 treatments with 90 isocenter setups during the course of the study. Patients uniformly received 30 Gy (if possible) of radiotherapy in 5 fractions to the clinical target volume. The total dose was constrained by limiting the spinal cord to a maximum dose of 10 Gy. To verify correct daily patient positioning before each treatment and to determine the daily treatment setup error after radiation delivery, axial CT scans were taken before and immediately after each treatment without moving the patient from the treatment position, for comparison with the planning CT scan. Toxicity was measured using the Common Toxicity Criteria, the Late Effects of Normal Tissue scoring system and a neurologic function scale. Follow-up was conducted 4 weeks after completion of SBRT, and then 2, 3, 6, 9, 12, and every 6 months thereafter. RESULTS: The procedure was technically feasible to perform in all patients. No neurologic toxicity was observed in any patient. The median follow-up time was 9 months (range 6-16). The Clopper-Pearson upper bound on the probability of paralysis with 95% confidence is no greater than 0.181. The positional setup error was determined to be within 1 mm of planning isocenter. CONCLUSIONS: This Phase I study shows that intensity-modulated, near simultaneous, CT image-guided SBRT is a feasible, and highly precise technique for the noninvasive treatment of spinal metastases. Although no paralysis has developed in the 15 patients treated, continued monitoring for spinal cord toxicity is warranted, as larger numbers of patients will be needed to more precisely define the upper bound on the probability of spinal cord myelopathy.     

Where to locate the isocenter? The treatment strategy for repeat trigeminal neuralgia radiosurgery

PURPOSE: The purpose of this study is to investigate how the spatial relationship between the isocenters of the first and second radiosurgeries affects the overall outcome. METHODS AND MATERIALS: We performed a retrospective study on 40 patients who had repeat gamma knife radiosurgery for trigeminal neuralgia. Only one 4-mm isocenter was applied in both first and second radiosurgeries, with a maximum radiation dose of 75 Gy and 40 Gy, respectively. The MR scan of the first radiosurgery was registered to that of the second radiosurgery by a landmark-based registration algorithm. The spatial relationship between the isocenter of the first and the second radiosurgeries was thus determined. The investigating parameters were the distance between the isocenters of the two separate radiosurgeries and isocenter proximity to the brainstem. The outcome end points were pain relief and dysesthesias. The median follow-up for the repeat radiosurgery was 28 months (range, 6-51 months). RESULTS: Pain relief was complete in 11 patients, nearly complete (> or =90%) in 7 patients, partial (> or =50%) in 8 patients, and minimal (<50%) or none in another 14 patients. The mean distance between the two isocenters was 2.86 mm in the complete or nearly complete pain relief group vs. 1.93 mm in the others. Farther distance between isocenters was associated with a trend toward better pain relief (p = 0.057). The proximity of the second isocenter to the brainstem did not affect pain relief, and neither did placing the second isocenter proximal or distal to the brainstem compared with the first one. Three patients developed moderate dysesthesias (score of 4 on a 0-10 scale), and 2 other patients developed more significant dysesthesias (score of 7) after the second radiosurgery. Dysesthesias related neither to distance between isocenters nor to which isocenter was closer to the brainstem. CONCLUSIONS: Image registration between MR scans of the first and second radiosurgeries helps target delineation and radiosurgery treatment planning. Increasing the isocenter distance between the two radiosurgeries treated a longer segment of the trigeminal neuralgia nerve and was associated with a trend toward improved pain relief.     

Isocenter accuracy in frameless stereotactic radiotherapy using implanted fiducials

PURPOSE: The stereotactic radiotherapy (SRT) system verifies isocenter accuracy in patient space. In this study, we evaluate isocenter accuracy in frameless SRT using implanted cranial gold markers. MATERIALS AND METHODS: We performed frameless SRT on 43 intracranial tumor patients between August 1997 and December 2000. The treatment technique was determined by the tumor shape and volume, and by the location of critical organs. The coordinates of anterior-posterior and lateral port film were inputted to ISOLOC software, which calculated (1) the couch moves translation distance required to bring the target point to the isocenter, and (2) the intermarker distance comparisons between the CT study and the treatment machine films. We evaluated the isocenter deviation based on the error between orthogonal film target coordinates and isocenter coordinates. RESULTS: The mean treatment isocenter deviations (x, y, z) were -0.03, 0.14, and -0.04 mm, respectively. The systematic component isocenter standard deviations were 0.28, 0.31, and 0.35 mm (1 SD), respectively, and the random component isocenter standard deviations were 0.53, 0.52, and 0.50 mm (1 SD), respectively. CONCLUSIONS: The isocenter accuracy in the frameless SRT-implanted fiducial system is highly reliable and is comparable to that of other stereotactic radiosurgery systems.     

Metastatic spinal cord compression

Approximately 70% of cancer patients have metastatic disease at death. The spine is involved in up to 40% of those patients. Spinal cord compression may develop in 5% to 10% of cancer patients and up to 40% of patients with preexisting nonspinal bone metastasis (>25,000 cases/y). Given the increasing survival times of patients with cancer, greater numbers of patients are likely to develop this complication. The role of surgery in the management of metastatic spinal cord compression is expanding. The management of metastatic spine disease can consist of a combination of surgery, radiation treatment, and chemotherapy. Treatment modalities are not mutually exclusive and must be individualized for patients evaluated in a multidisciplinary setting.     

Surgical treatment of spinal cord compression in kidney cancer

Forty-three patients with renal-cell carcinoma underwent treatment for spinal cord compression over a 7-year period. Of these, 32 patients underwent surgery, while 11 patients underwent radiation alone. Before operation, 25 patients had relapsed following prior radiation, while seven others received postoperative radiation. A more aggressive surgical approach, tailored to the site of compression within the spinal canal, was used with the majority undergoing gross total tumor resection by an anterior approach. Immediate stability of the spine was achieved with methyl-methacrylate reconstruction of the resected segments. Preoperative spinal angiography with embolization of hypervascular tumors was carried out in eight patients. Patient parameters in the surgical and irradiated groups were comparable, except that a greater proportion of the radiation alone group had more than one organ system involved (64% v 44%). The median survival of the surgically treated patients was 13 months, compared with 3 months for those treated by radiation alone. In addition, a greater proportion of the surgically treated patients were benefitted neurologically (70%) compared with those treated by radiation (45%). With the development of effective surgical treatment for spinal metastases, early consideration for surgical treatment (before radiation) should be considered in selected patients. Preoperative spinal angiography and embolization are recommended whenever feasible to minimize intraoperative blood loss.     

Spinal cord compression in small cell lung cancer: a retrospective study of 610 patients

The records of 610 consecutive patients with small cell lung cancer, treated on a common protocol in a multicentre trial, were reviewed and 24 (4%) cases of spinal cord compression identified. Five hundred patients had isotope bone scans performed at presentation, and in 131 (26%) there was abnormal isotope uptake in the spinal column; only 7% of these patients developed spinal cord compression. However, of the 24 patients who presented with back pain and had a positive bone scan affecting the spine, 36% developed cord compression. Cerebral metastases occurred at some stage in 19.5% of all patients and in 45% of patients with cord compression. The combination of cerebral metastases and a positive bone scan gave a 25% chance of developing spinal cord compression. There were two distinct forms of clinical presentation. Six patients (group A) presented with cord compression: All had back pain and positive bone scans, five out of six had sphincter disturbance, and median survival from cord compression was 30 weeks. Eighteen patients (group B) developed cord compression while on treatment: 28% had positive initial bone scans, 44% back pain and 61% sphincter disturbance, and median survival from cord compression was 4 weeks. Spinal cord compression is an important cause of morbidity and mortality in small cell lung cancer. We suggest that it may be possible to select patients who should receive radiotherapy to the spine to try to prevent the development of this complication.     

Detection of unsuspected spinal cord compression in melanoma patients by 18F-fluorodeoxyglucose-positron emission tomography.

AIM: Positron emission tomography (PET) using (18)F-fluorodeoxyglucose can detect early or small metastatic deposits of melanoma and guide subsequent correlative anatomical imaging and treatment. The aim of this study was to assess the value of PET in demonstrating spinal cord compression by otherwise unsuspected metastatic disease. METHODS: Reports of 1365 PET studies performed on patients with melanoma were reviewed. Fifty patients considered to be at risk of spinal cord compression on the basis of PET were identified and 35 patients were analysed. Magnetic resonance imaging and computed tomography were used to confirm or refute the diagnosis. The symptoms and signs at the time of PET and follow-up status were compared between patients with and without confirmed spinal cord compression. RESULTS: In nine patients (26%) compression of the spinal cord or adjacent neurological structures was confirmed and eight of these patients had immediate treatment. Survival was poor in both patient groups, but three patients with confirmed compression maintained good neurological functional status following treatment. CONCLUSION: PET can detect imminent, unsuspected spinal cord compression in patients with metastatic melanoma. Immediate anatomical imaging of the spine is recommended in patients who have evidence of spinal cord compression on PET.     

The effect of palliative radiation therapy on epidural compression due to metastatic malignant melanoma.

The efficacy of palliative radiation therapy in the treatment of spinal cord and cauda equina compression due to metastatic malignant melanoma was evaluated in 38 sites in 35 patients treated between 1970 and 1990. All patients had radiographic documentation of epidural compression. The median dose of radiation therapy was 2850 cGy (range, 500 to 4000 cGy), with daily fractions ranging from 200 to 800 cGy. Twenty-eight sites in 26 patients were evaluable 1 month after completion of radiation therapy, and symptoms responded completely in 11 of 28 (39%) sites. Fourteen sites (46%) showed a partial response of symptoms. Response lasting until death was documented in 21 of 26 patients (81%). Patients receiving a total dose of 3000 cGy or greater were more likely to achieve a complete response than those receiving less than 3000 cGy (62% versus 20%) by univariate (P = 0.025) and multivariate (P = 0.048) analyses. A treatment program of radiation therapy and corticosteroids is effective in palliating the symptoms of epidural compression due to metastatic malignant melanoma. It is recommended to deliver an accelerated course of radiation therapy to a dose of 3000 cGy or greater without exceeding spinal cord tolerance (e.g., 3000 cGy in ten fractions at 300 cGy per fraction).