• Embed Doc
  • Copy Link
  • Readcast
  • Collections
  • CommentGo Back
Download
 
N
EUROSURGERY 
VOLUME 64 | NUMBER 2 | FEBRUARY 2009 SUPPLEMENT|
 A67
SPINAL RADIOSURGERY 
Iris C. Gibbs, M.D.
Department of Radiation Oncology,Stanford University Medical Center,Stanford, California
Chirag Patil, M.D.
Department of Neurosurgery,Stanford University Medical Center,Stanford, California
Peter C. Gerszten, M.D.
Departments of Neurological Surgeryand Radiation Oncology,University of Pittsburgh School of MedicineUniversity of Pittsburgh Medical Center,Pittsburgh, Pennsylvania
John R. Adler, Jr., M.D.
Department of Neurosurgery,Stanford University Medical Center,Stanford, California
Steven A. Burton, M.D.
Department of Radiation Oncology,University of Pittsburgh School of Medicine,University of Pittsburgh Medical Center,Pittsburgh, Pennsylvania
Reprint requests:
Iris C. Gibbs, M.D.,Department of Radiation Oncology,Stanford University Medical Center,875 Blake Wilbur Drive,Stanford, CA 94305-5847.Email: iris.gibbs@stanford.edu
Received,
 June 2, 2008.
Accepted,
November 6, 2008.Copyright © 2009 by theCongress of Neurological Surgeons
S
ince the classic reports of cervical spinalcord radiation myelopathy by Ahlbom (1)in 1941 and Boden (3) in 1948, radiation- induced myelopathy has been a feared com-plication of conventional radiotherapy (13).Animal studies and clinical modeling derivedfrom observational data provide the basisforour current understanding of the factorsaffecting the spinal cord’s tolerance to radia-tion. The most important of these appear to be the radiation dose schedule, which includesdose per fraction, total dose, and interfractioninterval. It is widely accepted that the thera-peutic index of radiotherapy limits the radia-tion dose near the spinal cord to such anextent that the control of many tumors is compromised. With advances in technology,especially the development of image-guidedradiosurgery, the ability to deliver a moreaggressive radiation dose adjacent to the spineis now feasible. Nevertheless, the tolerance ofthe spinal cord to such single or hypofraction-ated radiation schedules is uncertain and maywell differ markedly from conventional radia-tion therapy precedents. Combining 2 of thelargest CyberKnife (Accuray, Inc., Sunnyvale,CA) spinal radiosurgery series, we report 6
D
ELAYED
 R
ADIATION
-
INDUCED
M
YELOPATHYAFTER
S
PINAL
R
ADIOSURGERY
OBJECTIVE:
Spinal cord injury is arguably the most feared complication in radiotherapyand has historically limited the aggressiveness of spinal tumor treatment. We report acase series of 6 patients treated with radiosurgery who developed delayed myelopathy.
METHODS:
Between 1996 and 2005, 1075 patients with benign or malignant spinaltumors were treated by CyberKnife (Accuray, Inc., Sunnyvale, CA) robotic radiosurgery atStanford University Medical Center and the University of Pittsburgh Medical Center.Patients were followed prospectively with clinical and radiographic assessments at1- to 6-month intervals. A retrospective review identified patients who developed delayed radiation-induced myelopathy. Six patients (5 women, 1 man) with a mean age of 48years (range, 25–61 years) developed delayed myelopathy at a mean of 6.3 months (range,2–9 months) after spinal radiosurgery. Three tumors were metastatic; 3 were benign. Themetastases were in the upper to midthoracic spine, whereas the benign tumors were par-tially in the cervical region. Three cases involved previous radiation therapy.
RESULTS:
Dose volume histograms were generated for target and critical structures.Clinical and dosimetric factors were analyzed for factors predictive of spinal cord injury.Specific dosimetric factors contributing to this complication could not be identified,but one-half of the patients with myelopathy received spinal cord biological equiva-lent doses exceeding 8 Gy.
CONCLUSION:
Delayed myelopathy after radiosurgery is uncommon with the doseschedules used in this case series. Radiation injury to the spinal cord occurred over aspectrum of dose parameters that prevented identification of specific dosimetric fac-tors contributing to this complication. Primarily, biological equivalent dose estimateswere not usable for defining spinal cord tolerance to hypofractionated dose schedules.We recommend limiting the volume of spinal cord treated above an 8-Gy equivalentdose, because half of the complications occurred beyond this level.
KEY WORDS:
Myelopathy, Radiation complications, Radiosurgery
Neurosurgery 64:A67–A72, 2009 
DOI: 10.1227/01.NEU.0000341628.98141.B6
www.neurosurgery-online.com
ABBREVIATIONS: BED
, biological equivalent dose;
MRI
, magnetic resonance imaging
 
cases of radiation-induced myelopathy among more than 1000patients treated with the CyberKnife for benign and metasta-tic spinal tumors.
PATIENTS AND METHODS
Between 1996 and 2005, 1075 patients with benign or metastaticspinal tumors were enrolled prospectively onto Institutional Review Board-approved registry protocols and treated with CyberKnife radio-surgery at either Stanford University or the University of Pittsburgh.Patients with spinal instability, target tumors involving more than2vertebral levels, or spinal cord compression causing acute neurolog-ical deterioration were excluded from enrollment. The overall cohort of1075 patients included 156 patients with benign extramedullary tumorsand 919 patients with metastatic tumors. Of the metastatic tumors, theprimary histologies were predominantly renal cell carcinoma (142patients), breast cancer (134 patients), lung cancer (129 patients),melanoma (49 patients), gastrointestinal adenocarcinoma (57 patients),sarcoma (42 patients), and prostate (40 patients). A database of clinicaland dosimetric factors was collected and used to identify patients whodeveloped treatment-related spinal cord injury.Pretreatment setup and immobilization of cervical lesions used a sim-ple, nonrigid Aquaplast face mask (WFR/Aquaplast Corp., Wyckoff,NJ). In the lumbothoracosacral region, a vacuum foam cradle was used.Treatment plans were generated using thin-section (1.25-mm slice thick-ness) contrast computed tomographic scans through the anatomic regionof interest. As most metastatic vertebral body tumors were clearly visi- ble by computed tomography, magnetic resonance imaging (MRI) withimage fusion was used when necessary to define the extent of benignextramedullary tumors. In general, for treatment planning purposes,small metastatic tumors involving less than 25% of the vertebral bodywere contoured with an approximately 2- to 3-mm margin; all other ver-tebral body tumors encompassed the entire vertebra. Until 2004, spinaltargeting in the lower cervical, thoracic, and lumbar spine required thepreradiosurgical percutaneous insertion of either stainless steel bonescrews (Stanford University Medical Center) or subperiosteal gold seeds(University of Pittsburgh Medical Center), which served as localizingfiducials. Treatment simulation techniques were similar at the 2 institu-tions and have been described previously (4, 6). In 2004, the availabilityof fiducial-less Xsight (Accuray, Inc.) hierarchical mesh tracking obviatedthe need for metallic markers; with this system, the imaging system cor-relates images of bony anatomy rather than implanted fiducials.On the basis of institutional preference, radiosurgery was generallyadministered in a single fraction at the University of Pittsburgh,whereas patients were treated in 2 to 5 sessions at Stanford University,depending on the tumor size or proximity to the spinal cord. A detaileddescription of the treatment planning used at both institutions has pre-viously been published (6). For all patients, the total dose prescriptionvaried between 12.5 and 25 Gy (12.5–20 Gy in 1 fraction, 18 to 22 Gy in2fractions, 18 to 24 Gy in 3 fractions, 14 to 24 Gy in 4 fractions, or 25Gy in 5 fractions). The majority of patients (112 benign, 803 metastatic)were treated with a single fraction; 90 (22 benign, 68 metastatic) weretreated in 2 fractions; and 70 (22 benign, 48 metastatic) were treated in3 to 5 fractions. Care was taken to limit the maximum dose to thespinal cord or cauda equina to less than 8 to 10 Gy while aiming tooptimize the coverage of the target lesion to at least 90%. However, thisconstraint was relaxed when the volume of spinal cord receivinggreater than this dose was estimated to be only a few voxels.Dose volume histograms were generated for the target lesion andcritical structures including the spinal cord. Conformality, homogene-ity, and coverage indices were calculated and recorded using conven-
 A68
| VOLUME 64 | NUMBER 2 | FEBRUARY2009 SUPPLEMENT
www.neurosurgery-online.com
G
IBBS ET AL
.
tional measures (12, 16). Radiosurgery was performed as an outpatientprocedure lasting 30 to 60 minutes. Subsequently, patients withmetastatic lesions were followed both clinically and radiographically at1- to 3-month intervals, and patients with benign lesions were followedat 6-month intervals. Late complications were scored according to theCommon Toxicity Criteria, Versions 2.0 and 3.0 (17, 19).The dose volume effects on the spinal cord were analyzed for allpatients in this series. Because of the variety of radiation schedules,each was converted to biological equivalent doses (BED
3
) using thefollowing formula, with an assumed
α
/
β
ratio of 3 for spinal cord (7):
BED
nd
1
d
α
/
β
To determine which factors were predictive of spinal cord injury, mul-tiple clinical and dosimetric factors including sex, age, histology, present-ing symptoms, previous radiotherapy, anatomic level, prescribed dose,dose per fraction, tumor volume, and spinal cord dose were analyzed.
RESULTS
The tumor volumes in all patients ranged from 0.025 to 685mL. More than 55% of metastatic patients were previously irra-diated. Although care was taken to limit the volume of spinalcord receiving 8 Gy or higher, the maximum spinal cord doseranged from 3.6 to 29.9 Gy. In all but 2 patients treated in singlefractions, the volume of spinal cord treated to a dose of morethan 8 Gy was less than 1 mL. The average spinal cord volumetreated to 8 Gy equivalent dose or more was also less than 1 mL.Our review identified 6 patients who developed radiation- induced delayed spinal cord myelopathy. The mean time toonset was 6.3 months (range, 2–10 months). Four patients weretreated at Stanford University Medical Center and 2 at theUniversity of Pittsburgh Medical Center. The characteristics ofthe 6 patients who developed radiation-induced myelopathyafter spinal radiosurgery are shown in Tables 1 and 2. The meanage of the 5 women and 1 man was 48 years (range, 25–61years). Three of the 6 spinal tumors were metastases, and 3 were benign tumors (meningioma, schwannoma, and neurofibroma).The three metastatic tumors were located entirely in the upperto midthoracic spine, whereas all 3 benign tumors were locatedpredominantly in the cervical and cervicothoracic region. Withregard to clinical symptoms of myelopathy, 5 of the 6 patientshad motor and sensory deficits, 3 had bowel and bladder symp-toms, and 4 had significant pain (Table 3).Myelopathic symptoms were initially managed by corticos-teroids in all patients. Some patients also received a combina-tion of vitamin E and pentoxifylline (Trental; Sanofi-Aventis,Bridgewater, NJ), hyberbaric oxygen, gabapentin (Neurontin;Pfizer, New York), and/or physical therapy. After an initialworsening, 3 of the 6 patients had improvement in their myelo-pathic symptoms after treatment, 2 reached a plateau, and 1patient progressed to paraplegia. All 3 patients who showedclinical improvement had complete radiographic resolution oftheir spinal cord edema on MRI. The earliest radiographic find-ings were edema within the spinal cord, signified by hyperin-tensity on the T2-weighted MRI sequence above and below theregion of injury. In most patients, clinical symptoms shortly
 
N
EUROSURGERY 
VOLUME 64 | NUMBER 2 | FEBRUARY 2009 SUPPLEMENT|
 A69
S
PINAL
M
YELOPATHY AFTER
R
ADIOSURGERY
preceded or were coincident with these findings. Within 6 to 8months, an area of contrast enhancement at the level of injurydeveloped. By 12 to 18 months from onset, a region of myelo-malacia was seen within the spinal cord at the area of injury.Figure 1shows these characteristic radiographic findings of radiation-induced myelopathy. Two of the patients hadreceived irradiation before radiosurgery at doses of 50.4 and39.6 Gy in 1.8-Gy fractions, at 70 and 81 months, respectively.The estimated maximum spinal cord doses in these previouscourses were 25.2 and 40 Gy, respectively. Also of interest is that2 of the 6 patients who developed these injuries received anantiangiogenic or epidermal growth factor inhibitor-targetedtherapy within 2 months of developing clinical myelopathy.One patient died of systemic disease progression 17 monthsafter treatment and 7 months after the onset of myelopathy. Ofnote, the treated lesion in this patient appeared radiographi-cally stable at the last follow-up examination.An analysis of the cohort of myelopathic patients shows thatthe mean prescribed dose to the tumor margin was 21.6 Gy.The mean treatment volume was 9.4 mL. The mean maximumBED
3
for the spinal cord dose was 78.7 Gy (range, 32.6–140.3 Gy;standard deviation, 46.4 Gy) (Table 2). Three of the 6 patientsreceived a maximum spinal cord dose of less than the averagedose received among all patients. One patient received a dose
a
BED
3
(8 Gy)
biological equivalent dose of 8 Gy in 1 fraction
29 Gy.
FIGURE 1.
Characteristic radio- graphic findings of delayed radia-tion-induced myelopathy.
A
,spinal cord edema at the onset of symptoms (brackets) and vertebrallevel of the T1 spinous process ini-tially treated (arrow).
B
, spinalcord enhancement (brackets) at 6to 8 months after onset.
C
, myelo-malacia (brackets) 12 to 18 monthsafter onset of symptoms.
BAC
TABLE 1.Patient characteristics and tumor locationOnset ofPrevious radiationPatientAge (y)/Tumor typeLocationmyelopathyand intervalPrevious Previousno.sex(mo)between courses (mo)surgerychemotherapy
159/FRenal cell carcinomaT69NoNoYes255/FBreast cancerT56Yes (80.8)NoYes359/FBreast cancerT14Yes (70.0)YesYes429/FMeningiomaC7–T29NoYesNo561/FSchwannomaC62NoYesNo625/MNeurofibromaC–C75NoNoNo
TABLE 2.Tumor and radiosurgery treatment plan characteristics
a 
TreatmentTotalMaximumMaximumVolume of cordPatientTumor typeLocationvolumeprescriptionNo. of dose to spinalspinal cordreceiving
BED
3
no.(cm
3
)dose (Gy)treatmentscord (Gy)BED
3
(Gy)(8 Gy) (cm
3
)
1Renal cell T6 vertebral body 13.725226.2140.63.4carcinomaand spinal canal2Breast cancerT5 vertebral body10.520219.280.62.63Breast cancerT1 posterior elements18.921213.946.10.34MeningiomaRight C7–T2 spinal 7.624329.9129.24.0canal5SchwannomaRight C6 spinal canal4.52018.532.60.16NeurofibromaRight C7 spinal canal1.22011043.30.2
of 00

Leave a Comment

You must be to leave a comment.
Submit
Characters: ...
You must be to leave a comment.
Submit
Characters: ...