degen_spinepain_jneurosurg_2005

7/28/2019 degen_spinepain_jneurosurg_2005

1/10

ADIATION therapy has become a mainstay in thetreatment of both primary36 and metastatic diseaseof the spine2 and spinal cord.5 Although radiother-

apy has proven effective in controlling tumors, side effectssuch as radiation-induced myelitis limit the safely deliver-able doses and decrease its efficacy of treating previouslyirradiated areas.46 Stereotactic radiosurgery permits high-dose treatment while minimizing exposure to surroundinghealthy tissues, increasing the safety of initial treatment,and allowing reirradiation of recurrent spinal lesions.

Stereotactic radiosurgery was first developed by Lars

Leksell in 1951 for the treatment of intracranial lesions.35

The minimal movement of intracranial contents relative tothe skull allowed treatment planning and radiation deliv-ery to be performed using a head frame as a fixed refer-ence point. Radiosurgery without the use of a head framebecame a reality with the advent of the CyberKnife (Accu-ray, Sunnyvale, CA).8

The CyberKnife is a compact 6-MV LINAC that is mount-ed on a computer-controlled robotic arm and that can deliv-er multiple, nonisocentric, noncoplanar radiation beams. Itincorporates a dynamic tracking system consisting of anorthogonal pair of diagnostic-quality x-ray imaging devicesand software that can locate fiducial markers implantednear the tumor. This provides the capability of sending up-dated position information to the robot, allowing adaptivebeam algorithms to correct for patient movement and per-mitting accurate targeting of the therapeutic beam duringtreatment. Because the radiation source can track the target,

complete target immobilization is unnecessary.1One advantage of this and similar technologies is that

radiosurgery no longer needs to be confined to the intracra-nial compartment. By using bone landmarks or implantedfiducial markers, stereotactic radiosurgery has been used totreat lesions of the spine,9,18,19,29,39,47 pancreas,29 prostate,22

and lung.51 Because radiosurgery does not require the appli-cation of a head frame, staged radiosurgery (that is, frac-tionation) is feasible.39

Many patients with spinal lesions experience pain butno neurological deficit.28 In this study we chose, therefore,to use the SF-12 Health Survey and a VAS to determinethe effects of stereotactic radiosurgery in cases of spinallesions on QOL and pain. The SF-12 reliably appraises

J Neurosurg: Spine 2:540549, 2005

540

CyberKnife stereotactic radiosurgical treatment of spinaltumors for pain control and quality of life

JEFFREY W. DEGEN, M.D., GREGORY J. GAGNON, M.D., JEAN-MARC VOYADZIS, M.D.,DONALD A. MCRAE, PH.D., MICHAEL LUNSDEN, M.S., D.A.B.R., SONJA DIETERICH, PH.D.,INGE MOLZAHN, B.S., R.R.T., AND FRASER C. HENDERSON, M.D.

Departments of Neurosurgery and Radiation Medicine, Georgetown University Hospital, Washington, DC

Object. The authors conducted a study to assess safety, pain, and quality of life (QOL) outcomes following CyberKniferadiosurgical treatment of spinal tumors.

Methods. Data obtained in all patients with spinal tumors who underwent CyberKnife radiosurgery at GeorgetownUniversity Hospital between March 2002 and March 2003 were analyzed. Patients underwent examination, visual analogscale (VAS) pain assessment, and completed the 12-item Short Form Health Survey (SF-12) before treatment and at 1, 3,6, 8, 12, 18, and 24 months following treatment.

Fifty-one patients with 72 lesions (58 metastatic and 14 primary) were treated. The mean follow-up period was 1 year.Pain was improved, with the mean VAS score decreasing significantly from 51.5 to 21.3 at 4 weeks (p 0.001). Thiseffect on pain was durable, with a mean score of 17.5 at 1 year, which was still significantly decreased (p = 0.002).

Quality of life was maintained throughout the study period. After 18 months, physical well-being was 33 (initial score32; p = 0.96) and mental well-being was 43.8 (initial score 44.2; p = 0.97). (The mean SF-12 score is 50 10 [standarddeviation].) Adverse effects included self-limited dysphagia (three cases), diarrhea (two cases), lethargy (three cases),paresthesias (one case), and wound dehiscence (one case).

Conclusions. CyberKnife radiosurgery improves pain control and maintains QOL in patients treated for spinal tumors.Early adverse events are infrequent and minor. The authors await long-term follow-up data to determine late complicationsand tumor control rates.

KEY WORDS CyberKnife neoplasm radiosurgery spine stereotaxis

R

J. Neurosurg: Spine / Volume 2 / May, 2005

Abbreviations used in this paper: BED = biologically effectivedose; EBRT = external-beam radiation therapy; HI = homogeneityindex; LINAC = linear accelerator; MCS = mental component sum-mary; NCI = nonconformality index; PCS = physical componentsummary; QOL = quality of life; SF-36 = Short Form36; VAS =visual analog scale.


2/10

physical and mental health, taking into account physicalfunctioning, role limitations due to physical or emotionalproblems, bodily pain, general health, vitality, social func-tioning, and mental health.50 It has been proven both reli-able and valid when tested in comparison with other suchsurveys.24,26,50

Clinical Material and Methods

Patient Selection and Fiducial Implantation

In March of 2002, physicians at Georgetown UniversityHospital began treating patients with the CyberKnife sys-tem. All patients treated for spinal metastatic and primarytumors between March 13, 2002, and March 12, 2003,were included in this analysis (Tables 1 and 2).

Radiographic fiducial screws were placed into osseousstructures around the vertebrae to be irradiated during tu-mor resection if radiosurgery was to be adjuvant therapy orpercutaneously as an outpatient procedure if radiosurgerywas to be the sole treatment. A detailed description of the

placement of the fiducial markers has been published pre-viously.39 For patients in whom a tissue diagnosis had notbeen established, percutaneous needle biopsy samplingwas performed at the time of fiducial placement.

Fifty-one patients (Table 1) with a total of 72 lesions (Ta-ble 2) underwent CyberKnife radiosurgery. The mean ageof the patients was 53 years. There were 58 metastatic and14 primary tumors distributed throughout the spinal col-umn. Of the 13 cases of primary spinal tumors, six (46%)were benign and seven (54%) were malignant. Twenty-sixpatients (51%) were treated for one metastatic lesion and12 (24%) for two or more. Thirty-eight lesions (53%) hadbeen previously treated with conventional EBRT, of which58% had been irradiated to maximal tolerance ( 4500

cGy in the cervical and lumbar spine and 4000 cGy inthe thoracic spine in a standard-fraction schedule). Themean follow-up duration was 350 days (range 35683days). Twenty-one patients died of systemic disease duringthe follow-up period. Six patients did not undergo analysisof pain and/or QOL assessment (in three there were no pre-treatment data for comparison and three died before follow-up data could be obtained). These patients were assessedfor complications and time to recurrence.

Patient Evaluations

Patients were evaluated by the senior author (F.C.H.)before treatment and at 1, 3, 6, 9, 12, 18, and 24 monthsafter initial treatment. Follow-up intervals were measured

from the first treatment, even if subsequent lesions weretreated in the same patient. When physical health prevent-ed the patient from coming to the clinic, data were col-lected over the telephone by a member of the neurosur-gery department. At each evaluation the SF-12 and VASwere completed. The VAS includes scores from 0 to 100for pain in each of four areas: neck, arms, trunk, and legs.Only the pain referable to the tumor was used in analysis toprevent the confounding effect of remote pain unrelatedto the treated tumor(s). Mental Component Summary andPCS scores were computed for each SF-12 evaluationaccording to standard methodology. In the general popula-tion, each value has a mean of 50 10 (standard devia-tion), with higher values signifying better health.50 Due to

missed appointments, 12% of the data (27 evaluations in-volving 19 patients) were collected retrospectively.

Imaging studies were obtained every 3 months follow-ing treatment for malignant lesions and every 6 months forbenign lesions, as well as when clinically indicated. Arecurrence was indicated if there was radiographic evi-dence of tumor growth, if the patient suffered a new neu-rological deficit, or if the patient noted increased painreferable to a treated tumor.

Treatment Planning and Radiation Delivery

In the supine position patients were treated using stan-dard radiotherapy immobilization devices and a poly-styrene bead vacuum bag. For treatment planning, a com-puterized tomography scan was obtained with 300 slices,each with 1-mm thickness and spacing, centered on thefiducial array. The attending neurosurgeon (F.C.H.) pro-vided lesion and critical structure contours. The radiationoncologist (G.J.G.) provided dose guidelines and the crit-ical structure dose limits. Inverse planning yielded treat-ment deliveries with a mean of 300 beams. Treatmentsrequired a mean of 25 minutes for setup and 75 minutesfor radiation delivery.

The NCI37 is equal to the product of the target volume

and isodose volume divided by the square of the volumeof the target that falls within the selected isodose line (tar-get isodose volume)

NCI = (TV)(IV)/TIV2

where TV is the target volume, IV is the isodose volume,and TIV the target isodose volume. This is a simple wayof mathematically defining how precisely the plannedradiation delivery scheme and the target volume overlap.The HI40 consists of the ratio of the maximum dose deliv-ered within the treatment volume to the prescribed dose:

HI = MD/PD

where MD is the maximum treatment dose and PD the


Quality of life after spinal radiosurgery

541

TABLE 1

Summary of tumor-related characteristics

Characteristic Value (%)

no. of patients 51primary tumor

benign 6 (11.8)

malignant 7 (13.7)metastatic tumor

single 26 (51.0)multiple 12 (23.5)

mean age (yrs) 52.75no. of lesions 72

metastatic 58 (80.5)primary 14 (19.4)

spinal location of lesionscervical 16 (22.2)thoracic 32 (44.4)lumbar 15 (20.8)sacral 9 (12.5)

no. of lesions previously irradiated 38 (52.8)5000 cGy 16 (22.2)5000 cGy 22 (30.5)

no. of lesions not previously irradiated 34 (47.2)


3/10

prescribed dose. For each treatment course, the NCI and HIwere computed.

Radiobiological CalculationsWith increasing conformality comes the opportunity to

deliver high-dose fraction sizes to tumors with the goal ofimproving local control. Because of concerns regardingtissue toxicities, most treatments in the modern radiothera-py era have consisted of multiple fraction treatments of sev-eral weeks duration or, at the other extreme, a single largestereotactic fraction. We believed that staged irradiation(threefive fractions) would capture much of the benefit offractionation while still providing large ablative doses. Tothat end, a series of radiobiological calculations have beenconducted to determine our current fraction schemes and tocompare them with standard fractionated treatment: model-ing was used in the manner of Dale.11,12 In this formalism,

the tumor control probability is the zero term of the Poissondistribution or

TCP = eNcS

where TCP is the tumor control probability, Nc is the esti-mated clonogen number, and S is survival fraction. Thesurvival fraction is exponentially related to dose and thisdose can be parameterized in the BED. The survival frac-tion is therefore:

S = eBED

where is a tumor or tissue-specific parameter represent-ing the single-hit radiation sensitivity of that tumor or tis-sue. A higher BED is therefore associated with better tu-

mor control probability. The BED is calculated from therelationship

BED = TD RE RF

where TD is total dose, RE is the relative effectiveness ofthat dose, and RF is the repopulation factor. Because theclonogen cell number Nc is perhaps the least determinant

parameter, we chose to compare values of BED among thetreatments and not the tumor control probability.

The BED of fractionated EBRT, assuming repopulationand repair of sublethal damage, is given by the following:

where n is fraction number, d is dose per fraction, T is treat-ment time, Tp is potential cell doubling time, is the tumoror tissue-specific parameter representing multihit radiationsensitivity of that tumor or tissue,11,12 and K = ex, where xis the time between fractions, and is the sublethal dam-age repair constant, ln(2)/repair half-life. In this formula-

tion, nd is the total dose, Tln(2)/Tp is the repopulation fac-tor RF, and the remainder of the function is the relativeeffectiveness, RE, of the fractionation scheme, includingexponential repair.

Here we calculate the BEDs obtained by fractionatedEBRT and some typical CyberKnife treatment schemes fora sample of reasonable radiobiological parameters (Table3). These include and parameters from Malaise, et al.,27

and potential doubling times, Tp, from Thames, et al.48 Avalue of = 0.346 hour1 was used in this study, repre-senting a repair half-life of 2 hours.

Statistical Analysis

Two-tailed paired t-tests were used to evaluate VASpain scores and SF-36 MCS and PCS scores, comparingeach time point with the corresponding pretreatment data.Significance was set at a probability value less than 0.05.

Regardless of the number of lesions treated, each patientwas counted only once in the statistical analysis with thedate of initial treatment used for determination of the fol-low-up duration for that individual. Subgroup analysis wasperformed using separate analyses of patients with primaryand metastatic disease. Within the group of patients withprimary spinal tumors, distinction was also made betweenthose with benign and those with malignant lesions. Thegroup involving metastatic disease was further divided toallow separate analysis of patients treated for a single lesion

and those treated for multiple spinal lesions. We did notsubdivide cases on the basis of chemotherapy because inalmost all patients initial chemotherapy had failed, and his-torically salvage response rates with second- and third-linechemotherapy have been nominal.

Results

Radiation Dosing

All patients completed their prescribed treatment cours-es. Lesions were treated with a mean of 2116 cGy and amean dose per fraction of 645 cGy. All but three lesionswere treated with hypofractionated therapy (mean 3.6 frac-tions per radiosurgery course). Most patients were treated

J. W. Degen, et al.

542 J. Neurosurg: Spine / Volume 2 / May, 2005

TABLE 2

Distribution of pathological entities

Lesion No. of Lesions No. of Patients

primary 14 13chondrosarcoma 1 1chordoma 3 3

ependymoma 2 2giant cell osteoblastoma 1 1hemangioma 1 1meningioma 1 1neuroma 3 3*paraganglioma 1 1malignant schwannoma 1 1*

metastatic lesions 58 38alveolar sarcoma 2 1bladder 1 1breast 14 12cervix 1 1colon 3 2endometrium 3 1Ewing sarcoma 1 1leiomyosarcoma 2 2

lung 2 2melanoma 6 2plasmacytoid lymphoma 2 1prostate 1 1RCC 8 6thyroid 9 3unknown adenocarcinoma 3 2

* One neuroma underwent malignant transformation and was retreated.

BEDf= nd (1d

)n (1K2) 2K (1Kn)T1n(2)

/ 1 K2 Tp


4/10


5/10

part, worsening deficits were the result of nonirradiatedtumors, with the following five exceptions: two primarylesions and three metastatic lesions that recurred immedi-ately adjacent to the irradiated field, which we will discuss.

Of patients with metastatic disease who had not previ-ously undergone irradiation, the tumor control rate was100%. Among those with metastatic disease who hadundergone prior irradiation, there were three clinical re-currences diagnosed a mean of 145 days after treatment(range 56218 days): an osteosarcoma twice previouslyirradiated, an RCC, and a melanoma metastasis, all ofwhich recurred immediately adjacent to the radiation iso-

dose line. Of cases involving primary tumors, two recur-red, one in a patient with a schwannoma that dedifferenti-ated into a malignant nerve sheath tumor, and one in apatient with a malignant osteoblastoma who had undergonethree resections and radiation to tolerance before radio-surgery. Three patients required additional surgery (oneeach with RCC, malignant nerve sheath tumor, and osteo-sarcoma) prior to retreatment with radiosurgery. One pa-tient with a clinical recurrence died of systemic diseasewithout any additional therapy being pursued.

Adverse Effects

No patients reported any complications referable toCyberKnife radiosurgery at the 3-month follow-up exam-

ination. Complications associated with radiosurgery weregenerally self-limited and mild. In one patient woundbreakdown developed at a surgical site and required de-bridement and reclosure of the wound. Other complica-tions potentially attributable to radiosurgery included in-creased nocturia (one patient), self-limited esophagitis ordysphagia (three patients), paresthesias (one patient), fa-tigue (three patients), transient diarrhea (two patients), andhoarse voice (one patient).

Discussion

The authors of previous studies have demonstrated thefeasibility and safety of delivering radiation doses to the

spine with submillimeter accuracy when using Cyber-Knife technology.9,10,19,39,42 In this study, we found that Cy-berKnife treatment lessens pain and stabilizes QOL. Onerationale for staging treatment (fractionation) is offered,and potential problems are discussed.

Pain Control and QOL

Most patients with spinal tumors present with pain ra-ther than neurological symptoms,28 and the life span islimited in most cases by systemic disease rather thanspinal metastases. We chose, therefore, to evaluate the ef-fect of CyberKnife radiosurgery on pain control and QOL.Analysis of our short-term results demonstrates a rapidand durable improvement in pain. Pain decreased within 1week of treatment in many cases and was significantlydecreased even 12 months after treatment for patients witheither metastatic disease or primary spinal tumors. Thelevel of pain increased at 18 months in the group of pa-tients with a single metastasis (Fig. 2 upper right). In largemeasure this is due to two patients in whom recurrencesdeveloped immediately adjacent to the site of irradiation;retreatment could not be attempted because conventionalirradiation of the area had been previously performed.

As demonstrated by the SF-12 data, our patients initiallyexperienced improved physical well-being (the PC scorewas significantly increased in all cases at 1 and 3 months).More importantly, the PC and MC scores remained essen-tially constant throughout the follow-up period (Table 5).We believed it an accomplishment to maintain the QOL inthese patients in whom other treatment options had beenexhausted.

Comparison of Data With Historical Controls

Comparisons with historical series are complicated bydifferent patient populations, different treatment sched-ules, and different assessment methods and end points. Inthe available prospective studies investigators did not ana-lyze bone metastases by site; thus, pain scores reflect acompilation of spinal and nonspinal sites. Because spinal

J. W. Degen, et al.


TABLE 5

CyberKnife treatmentrelated SF-36 PC and MC scores

Mean Scores at Given Follow-Up Periods (p value)

Variable* Initial 1 Mo 3 Mos 6 Mos 9 Mos 12 Mos 18 Mos

PC score

all patients 32.04 35.15 (0.01) 35.68 (0.02) 35.16 (0.27) 35.52 (0.24) 32.18 (0.56) 33.08 (0.96)metastatic lesion 32.08 34.86 (0.23) 35.10 (0.28) 33.22 (0.66) 33.39 (0.79) 31.13 (0.19) 28.58 (0.03)single 31.81 35.22 (0.28) 35.09 (0.44) 34.23 (0.93) 36.99 (0.16) 33.42 (0.85) 25.10 (0.17)multiple 32.28 34.55 (0.65) 35.74 (0.40) 33.10 (0.48) 30.20 (0.38) 26.56 (0.03) 32.05 (0.13)

all primary lesions 32.06 36.66 (0.12) 37.63 (0.02) 37.90 (0.12) 41.92 (0.03) 34.52 (0.55) 38.69 (0.29)benign 28.39 24.17 (0.94) 34.63 (0.15) 35.79 (0.34) 37.25 (0.14) 35.46 (0.21) 33.40 (0.37)malignant 34.67 43.79 (0.12) 40.13 (0.04) 40.53 (0.22) 44.26 (0.18) 33.27 (0.39) 42.22 (0.74)

MC scoreall patients 44.15 48.76 (0.64) 49.31 (0.25) 50.65 (0.11) 45.65 (0.95) 49.75 (0.96) 43.82 (0.98)

metastatic 44.06 48.99 (0.58) 48.08 (0.22) 51.37 (0.08) 44.84 (0.65) 48.57 (0.73) 42.90 (0.97)single 42.51 47.15 (0.81) 45.90 (0.29) 49.52 (0.09) 42.97 (0.24) 46.04 (0.98) 35.46 (0.82)multiple 48.06 53.58 (0.60) 53.14 (0.59) 54.99 (0.45) 46.50 (0.77) 53.63 (0.65) 50.35 (0.89)


* See Table 1 for summary of values associated with each variable.


6/10

metastases typically compose a large proportion (3050%) of the metastatic burden of bone,4,45 we believe that,despite the aforementioned limitations, there are severalexcellent prospective studies with which comparisons canbe made (Table 7).

A first consideration is the overall pain response, whichhas been reported in newer studies to be between 59 and90% when using standard radiation treatments.4,13,33,45 Thepresent series compares favorably with these pain out-comes. More importantly, the duration of pain relief needsto be considered. In nine studies investigators assessed theduration of pain relief.13 In two larger studies the authors

showed respectively the time to progression of pain to be5 to 6 months45 and a 40% rate of nonprogression of painat 12 months.4 Our results, although reflecting only thepain of the treated lesion, indicated median pain relief per-sisting beyond 12 months.

A final consideration is the retreatment rate, which is asurrogate for failure in the irradiated field. This rangedfrom 2 to 44% in seven trials;13 in two larger trials inves-tigators reported retreatment rates of 7 and 25% in one45

and 10 and 23% in the other.4 In the present series, six of72 lesions recurred in or near the treatment field for aretreatment rate of 8%, which compares favorably withthose in the literature.

Ideally, definitive statements of the efficacy of Cyber-Knife radiosurgery of spinal lesions would be based on aprospective randomized comparison of CyberKnife-relatedresults, and outcomes after conventional irradiation. Un-fortunately, in most patients treated with CyberKnife con-ventional means of irradiation have already been exhaust-ed, limiting the feasibility of such a trial.

Rationale for Staged Treatment

Investigators of clinical and in vitro studies have clearlydemonstrated efficacy when using single-fraction irradia-tion.7 Gamma Knife3,14,43,44,49 and LINAC15,31 technologieshave entailed single-dose treatments for intracranial tumorsyielding efficacy and low morbidity rates. To some extent,however, single fractionation is driven by the necessity ofplacing a rigid frame on the patient because staging (frac-tionating) radiosurgery by using a rigid frame is burden-some for both patient and physician. Because the Cyber-Knife is not constrained by the need for an external frame,staging treatment does not pose the same difficulties.

We have chosen to stage most treatment plans to mini-

mize the late effects on normal tissues and to maximizethe likelihood of tumor control. Our treatment strategy isbased on the principle that normal cells will more faith-fully undergo repair of sublethal and potentially lethalchromosomal injury than tumor cells, as manifested tosome extent in the aforedescribed linear quadratic theo-ry.11,12 Calculations suggest that staging treatments willlessen the undesired late effects in the central nervous sys-tem and increase tumoricidal effects in most malignanttumors. Other advantages include an opportunity for re-oxygenation and reassortment of cells,20,38 as well as short-er individual treatment times. Calculations based on thelinear quadratic method suggest that we were able todeliver substantially higher BEDs than would be possible

with conventional EBRT (Table 3).Our patients underwent a median of three treatments.

We devised a tumor grading scale that correlated with thenumber of staged treatments delivered. The grading scalealso correlated with the time required to develop a treat-ment plan, the treatment delivery time, and the confor-mality of treatment. Each tumor was graded using this 10-point scale that takes into account the lesions proximity tothe spinal cord, the degree to which it is wrapped aroundthe spinal cord or other critical structures, and whether theregion had been previously irradiated (Table 8). Simplelesions (12 points) were generally treated with a singledose of radiation, intermediate grade lesions (36 points)



545

TABLE 6

Summary of VAS pain scores after CyberKnife treatment

Mean Scores at Different Follow-Up Periods (p value)

Variable* Initial 1 Mo 3 Mos 6 Mos 9 Mos 12 Mos 18 Mos

all patients 51.53 21.33 (0.001) 19.00 (0.001) 23.90 (0.001) 30.00 (0.03) 17.50 (0.002) 32.73 (0.31)metastatic 52.09 16.80 (0.001) 18.96 (0.001) 25.77 (0.02) 31.33 (0.09) 16.88 (0.03) 33.33 (0.87)

single 51.70 22.22 (0.001) 19.33 (0.001) 28.33 (0.22) 39.29 (0.46) 9.00 (0.01) 56.67 (0.44)multiple 53.00 2.86 (0.03) 18.33 (0.11) 23.57 (0.07) 24.38 (0.11) 30.00 (0.52) 10.00 (0.46)


* See Table 1 for summary of values associated with each variable.

FIG. 1. Graph depicting maintenance of PCS and MCS scoresafter CyberKnife radiosurgery. Patients completed the SF-12. TheMC and PC scores were computed for each survey. The meanvalue for each score in the healthy population is 50 10. The PCscore initially improved before returning to baseline. There was nostatistically significant change in MC score. *p 0.05.


7/10

with three fractions (stages), and complex lesions (710points) with five fractions.

Radiation Delivery

The NCI defines how precisely the planned radiation

delivery volume overlaps the target. The ideal value isone, and this would be achieved if the target volume andisodose volume were the same size and shape, overlap-ping exactly. The more a radiation treatment plan deviatesfrom this ideal, the higher the NCI. We were able toachieve an NCI of 1.87. Although no other groups havepublished NCI-related data in spinal radiosurgery, thereare published data sets involving patients who underwenttreatment of intracranial lesions with the assistance ofa stereotactic head frame. Standard LINAC-based radio-surgery results in a mean NCI of approximately 2.7.32

With micromultileaf technology, this can be improved toapproximately 1.8,23 which is similar to the NCI found inour series. There is one large gamma knife series in which

the authors reported a mean NCI of 1.67,30 but cases wereexcluded from analysis if only partial coverage wasachieved due to the proximity of critical structures, heav-ily biasing the results in favor of a lower NCI.

In addition to conforming to the tumor size and shape,the CyberKnife-delivered radiation in our patients was

relatively uniform. The mean HI among our patients was1.43. Others have reported a mean HI of 1.34 for spinallesions treated with CyberKnife radiosurgery,39 and 2.0 forintracranial lesions treated with gamma knife surgery.25,41

In two studies involving intensity-modulated radiotherapyinvestigators reported HIs of 1.125 and 1.12,16 but thesestudies included only much smaller lesions (mean volume2.48 cm3). Thus, CyberKnife irradiation is relatively ho-mogeneous (with fewer so-called hot and cold spots), al-though the significance of this in spinal radiosurgery re-mains to be demonstrated.

Limitations of the CyberKnife

Although placement of fiducials is not difficult, it

J. W. Degen, et al.


FIG. 2. Graphs demonstrating alleviation of pain after CyberKnife radiosurgery. Patients undergoing radiosurgerywere asked to rate their relevant pain on a scale of 0 to 100 prior to treatment and at each follow-up visit. Pain quicklydecreased after treatment, and the effect was sustained for several months. Graphs are shown for the entire study popu-lation (upper left) as well as the subgroups analyzed: primary and metastatic tumors (upper right), single and multiplemetastases (lower left), and benign and malignant primary tumors (lower right). PCS = PCS score, MCS = MCS score.*p 0.05.


8/10

requires fluoroscopic localization and a sterile field. Anexperienced surgeon or interventionalist requires approxi-mately 5 minutes to place each fiducial. Induction of gen-eral anesthesia is necessary for placing screws in the cer-vical spine where we prefer a standard anterior (Cloward)exposure. Others prefer to use the skull for stereotacticlocalization in the cervical spine.18

Unplanned device-related down time was minimal: 3days were lost in the course of the 1st year for generalmaintenance and for repair of the x-ray imaging systems.At first, several patients with complex spinal tumors

required substantial time for treatment; however, treat-ment times have now been reduced by 50% since theintroduction of Express treatment software (Accuray),which speeds treatment delivery, and the Axum remote-controlled treatment couch (Accuray), which allows rapidpatient positioning.

Several patient characteristics militate against use of theCyberKnife. Excessive obesity causes difficulties in regis-tration of fiducials in the lower cervical, thoracic, andlumbar regions. Psychosis can potentially cause setup de-lays, missed treatments, and delays during treatment. Pa-tients with extreme pain or those who cannot lie supinemay require sedation and monitoring by an anesthesiolo-gist. Severe pulmonary disorders, especially those associ-

ated with frequent coughing, can result in excessive intra-treatment movement and frequent delays. In our 1st yearof using the CyberKnife, we excluded two cases becauseof obesity and one because of a pulmonary disorder.

Patients with sacral tumors initially presented difficulty inpositioning on the treatment table. Recognizing that somepatients breeched the theoretical safe zone (the patient wastoo close to the robot, risking a collision), we positionedthese patients feet first (toward the robot) on the treatmenttable. This entailed performing treatment planning feet first,with resulting potential for leftright confusion on the imag-ing study. We currently perform all treatments of the sacraland lumbar spine in this feet-first position, and stronglyencourage all other users to do so as well.

CyberKnife radiosurgery is a process requiring team-work. Good working relationships among the physicist,radiation oncologist, and neurosurgeon are essential toproducing good patient outcomes.

Conclusions

We have demonstrated that CyberKnife radiosurgerycan be safely performed in patients with benign and ma-lignant spinal lesions. In our study patients experiencedrelief of pain and maintenance of QOL with an acceptable

rate of adverse effects. The ability to stage (fractionate)treatment should minimize late radiation-induced effectson normal tissues, and the modality allows delivery of



547

TABLE 7

Previous studies of radiation therapy for bone metastases compared with the present study*

No. of Median Duration RetreatmentAuthors & Year Cases RT OPR (%) CPR (%) of Pain Relief Rate (%)

Niewald, et al., 1996 51 20 Gy/5 fx 77 33 35 wks 246 30 Gy/15 fx 86 31 35 wks 2

Gaze, et al., 1997 134 10 Gy/1 fx 83 39 22.5 wks 131 22.5 Gy/5 fx 89 42 24.9 wks

Jeremic, et al., 1998 109 4 Gy/1 fx 59 21 42 wks 42108 6 Gy/1 fx 73 27 50 wks 44110 8 Gy/1 fx 78 32 47 wks 38

Nielsen, et al., 1998 120 8 Gy/1 fx 62 6 mos 20119 20 Gy/4 fx 71 6 mos 11

BPTWP, 1999 383 8 Gy/1 fx 78 58 12 mos 23370 20 Gy/5 fx 78 58 12 mos 10

Bremer, et al., 1999 45 4 Gy/1 fx 55 3 mos 86 16 Gy/4/fx 80 9 mos

Steenland, et al., 1999 578 8 Gy/1 fx 72 37 20 wks 25579 24 Gy/6 fx 69 33 24 wks 7

present series 51 2124 Gy/3 fx 97 74 12 mos 8

* BPTWP = Bone Pain Trial Working Party; CPR = complete pain relief (the proportion of patients with complete relief of pain);fx = fraction; OPR = overall pain relief (the proportion of patients with any pain relief); RT = radiation therapy; not provided.

Studies with two sets of case numbers involved two different protocols.

TABLE 8

Georgetown CyberKnife grading scheme for spinal tumors*

Characteristic Point Value

proximity to critical structure (mm)10 0510 15 2

quadrants of critical structure w/in 10 mm of lesion

each quadrant 14previous irradiation

none 0tolerance 1tolerance 2

craniocaudal extent of tumor (cm)3 13 2

* Overall, 1 to 2 points indicates simple, 3 to 6 points an intermediate,and 7 to 10 points a complex lesion.

There are four quadrants around a critical structure (spinal cord). Atumor bordering the spinal cord around an arc of 180 would represent twoquadrants, and two points would be added to the tumor grade in this grad-ing scheme.

Tolerance is defined as equivalent to 4500 cGy to cervical or lumbarspinal cord or 4000 cGy to thoracic spinal cord in standard fractions.


9/10

higher BEDs than would be possible with conventionalirradiation. Tumor recurrence, with one exception, devel-oped only in patients who had previously undergone con-ventional irradiation to maximum tolerance. Our retreat-ment rate (8%) compares favorably with those reported inother large series, although longer follow up is necessaryto make definitive statements regarding efficacy.

Disclaimer

None of the authors has any financial relationship with Accuray,Inc.

References

1. Adler JR Jr, Chang SD, Murphy MJ, Doty J, Geis P, HancockSL: The Cyberknife: a frameless robotic system for radiosur-gery. Stereotact Funct Neurosurg 69:124128, 1997

2. Bilsky MH, Lis E, Raizer J, Lee H, Boland P: The diagnosis andtreatment of metastatic spinal tumor. Oncologist 4:459469,1999

3. Boethius J, Ulfarsson E, Rahn T, Lee H, Lippittz B: Gamma

knife radiosurgery for pilocytic astrocytomas. J Neurosurg 97(Suppl 5):677680, 20024. Bone Pain Trial Working Party: 8 Gy single fraction radiother-

apy for the treatment of metastatic skeletal pain: randomisedcomparison with a multifraction schedule over 12 months ofpatient follow-up. Radiother Oncol 52:111121, 1999

5. Bowers DC, Weprin BE: Intramedullary spinal cord tumors.Curr Treat Options Neurol 5:207212, 2003

6. Bremer M, Rades D, Blach M, Krenkel B, Karstens JH: Ef-fectiveness of hypofractionated radiotherapy in painful bonemetastases. Two prospective studies with 1 4 Gy and 4 4Gy. Strahlenther Onkol 175:382386, 1999

7. Brenner DJ, Martel MK, Hall EJ: Fractionated regimens for ste-reotactic radiotherapy of recurrent tumors in the brain. Int JRadiat Oncol Biol Phys 21:819824, 1991

8. Chang SD, Adler JR Jr: Current status and optimal use of radio-surgery. Oncology 15:209221, 20019. Chang SD, Main W, Martin DP, Gibbs IC, Heilbrun MP: An

analysis of the accuracy of the CyberKnife: a robotic framelessstereotactic radiosurgical system. Neurosurgery 52:140147,2003

10. Chang SD, Murphy M, Geis P, Martin DP, Hancock SL, DotyJR: Clinical experience with image-guided robotic radiosurgery(the Cyberknife) in the treatment of brain and spinal cord tu-mors. Neurol Med Chir 38:780783, 1998

11. Dale RG: The application of the linear-quadratic dose-effectequation to fractionated and protracted radiotherapy. Br J Radiol58:515528, 1985

12. Dale RG: The application of the linear-quadratic model to frac-tionated radiotherapy when there is incomplete normal tissuerecovery between fractions, and possible implications for treat-ments involving multiple fractions per day. Br J Radiol 59:919927, 1986

13. Falkmer U, Jarhult J, Wersall P, Cavallin-Stahl E: A systematicoverview of radiation therapy effects in skeletal metastases.Acta Oncol 42:620633, 2003

14. Feigl GC, Bonelli CM, Berghold A, Mokry M: Effects of gam-ma knife radiosurgery of pituitary adenomas on pituitary func-tion. J Neurosurg 97 (Suppl 5):415421, 2002

15. Friedman WA, Foote KD: Linear accelerator radiosurgery inthe management of brain tumours. Ann Med 32:6480, 2000

16. Fuss M, Salter BJ, Sadeghi A, Vollmer DG, Herezi JM, HermanTS: Fractionated stereotactic intensity-modulated radiotherapy(FS-IMRT) for small acoustic neuromas. Med Dosim 27:147154, 2002

17. Gaze MN, Kelly CG, Kerr GR, Cull A, Cowie VJ, Gregor A, et

al: Pain relief and quality of life following radiotherapy forbone metastases: a randomised trial of two fractionation sched-ules. Radiother Oncol 45:109116, 1997

18. Gerszten PC, Ozhasoglu C, Burton SA, Vogel WJ, Atkins BA,Kalnicki S, et al: CyberKnife frameless single-fraction stereo-tactic radiosurgery for benign tumors of the spine. NeurosurgFocus 14(5):E16, 2003

19. Gerszten PC, Ozhasoglu C, Burton SA, Vogel WJ, Atkins BA,Kalnicki S, et al: CyberKnife frameless stereotactic radiosur-gery for spinal lesions: clinical experience in 125 cases. Neuro-surgery 55:8999, 2004

20. Howes AE: An estimation of changes in the proportions andabsolute numbers of hypoxic cells after irradiation of trans-planted C3H mousemammary tumors. Br J Radiol 42:441447, 1969

21. Jeremic B, Shibamoto Y, Acimovic L, Milicic B, MilisavljevicS, Nikolic N, et al: A randomized trial of three single-dose radi-ation therapy regimens in the treatment of metastatic bone pain.Int J Radiat Oncol Biol Phys 42:161167, 1998

22. King CR, Lehmann J, Adler JR, Hai J: CyberKnife radiothera-py for localized prostate cancer: rationale and technical feasi-bility. Technol Cancer Res Treat 2:2530, 2003

23. Kubo HD, Wilder RB, Pappas CT: Impact of collimator leaf

width on stereotactic radiosurgery and 3D conformal radiother-apy treatment plans. Int J Radiat Oncol Biol Phys 44:937945, 1999

24. Lundberg L, Johannesson M, Isacson DG, Borgquist L: The re-lationship between health-state utilities and the SF-12 in a gen-eral population. Med Decis Making 19:128140, 1999

25. Ma L, Xia P, Verhey LJ, Boyer AL: A dosimetric comparisonof fan-beam intensity modulated radiotherapy with GammaKnife stereotactic radiosurgery for treating intermediate intra-cranial lesions. Int J Radiat Oncol Biol Phys 45:13251330,1999

26. Macran S, Weatherly H, Kind P: Measuring population health:a comparison of three generic health status measures. MedCare 41:218231, 2003

27. Malaise EP, Fertil B, Chavaudra N, Guichard M: Distribution of

radiation sensitivities for human tumor cells of specific histo-logical types: comparison of in vitro to in vivo data. Int J Ra-diat Oncol Biol Phys 12:617624, 1986

28. Markman M: Early recognition of spinal cord compression incancer patients. Cleve Clin J Med 66:629631, 1999

29. Murphy MJ, Adler JR Jr, Bodduluri M, Dooley J, Forster K, HaiJ, et al: Image-guided radiosurgery for the spine and pancreas.Comput Aided Surg 5:278288, 2000

30. Nakamura JL, Verhey LJ, Smith V, Petti PL, Lamborn KR, Lar-son DA, et al: Dose conformity of gamma knife radiosurgeryand risk factors for complications. Int J Radiat Oncol BiolPhys 51:13131319, 2001

31. Naoi Y, Cho N, Miyauchi T, Iizuka Y, Maehara T, KatayamaH: Usefulness and problems of stereotactic radiosurgery usinga linear accelerator. Radiat Med 14:215219, 1996

32. Nedzi LA, Kooy HM, Alexander E III, Svensson GK, LoefflerJS: Dynamic field shaping for stereotactic radiosurgery: a mod-eling study. Int J Radiat Oncol Biol Phys 25:859869, 1993

33. Nielsen OS, Bentzen SM, Sandberg E, Gradeberg CC, TimothyAR: Randomized trial of single dose versus fractionated pallia-tive radiotherapy of bone metastases. Radiother Oncol 47:233240, 1998

34. Niewald M, Tkocz HJ, Abel U, Scheib T, Walter K, Nieder C,et al: Rapid course radiation therapy vs. more standard treat-ment: a randomized trial for bone metastases. Int J Radiat On-col Biol Phys 36:10851089, 1996

35. Niranjan A, Lunsford LD: Radiosurgery: where we were, are,and may be in the third millennium. Neurosurgery 46:531543, 2000

36. Ozaki T, Flege S, Liljenqvist U, Hillmann A, Delling G, Salzer-Kuntschik M, et al: Osteosarcoma of the spine: experience of

J. W. Degen, et al.



10/10

the Cooperative Osteosarcoma Study Group. Cancer 94:10691077, 2002

37. Paddick I: A simple scoring ratio to index the conformity of ra-diosurgical treatment plans. Technical note. J Neurosurg 93(Suppl 3):219222, 2000

38. Palcic B, Skarsgard LD: Reduced oxygen enhancement ratio atlow doses of ionizing radiation. Radiat Res 100:328339,1984

39. Ryu SI, Chang SD, Kim DH, Murphy MJ, Le QT, Martin DP,et al: Image-guided hypo-fractionated stereotactic radiosurgeryto spinal lesions. Neurosurgery 49:838846, 2001

40. Shaw E, Kline R, Gillin M, Souhami L, Hirschfeld A, DinapoliR, et al: Radiation Therapy Oncology Group: radiosurgeryquality assurance guidelines. Int J Radiat Oncol Biol Phys 27:12311239, 1993

41. Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, etal: Single dose radiosurgical treatment of recurrent previouslyirradiated primary brain tumors and brain metastases: finalreport of RTOG protocol 9005. Int J Radiat Oncol Biol Phys47:291298, 2000

42. Shiomi H, Inoue T, Nakamura S, Inoue T: Quality assurance foran image-guided frameless radiosurgery system using radio-chromic film. Radiat Med 18:107113, 2000

43. Spiegelmann R, Lidar Z, Gofman J, Alezra D, Hadani M, PfefferR, et al: Linear accelerator radiosurgery for vestibular schwan-noma. J Neurosurg 94:713, 2001

44. Stafford SL, Pollock BE, Leavitt JA, Foote RL, Brown PD,Link MJ, et al: A study on the radiation tolerance of the opticnerves and chiasm after stereotactic radiosurgery. Int J RadiatOncol Biol Phys 55:11771181, 2003

45. Steenland E, Leer JW, van Houwelingen H, Post WJ, van den

Hout WB, Kievit J, et al: The effect of a single fraction com-pared to multiple fractions on painful bone metastases: a glob-al analysis of the Dutch Bone Metastasis Study. Radiother On-col 52:101109, 1999

46. Stewart FA: Re-treatment after full-course radiotherapy: is it aviable option? Acta Oncol 38:855862, 1999

47. Takacs I, Hamilton AJ: Extracranial stereotactic radiosurgery:applications for the spine and beyond. Neurosurg Clin N Am10:257270, 1999

48. Thames HD, Peters LJ, Withers HR, et al: Fractionation strate-gies for improving the results of radiotherapy. Cancer Bull 34:186193, 1982

49. Vesagas TS, Aguilar JA, Mercado ER, Mariano MM: Gammaknife radiosurgery and brain metastases: local control, survival,and quality of life. J Neurosurg 97 (Suppl 5):507510, 2002

50. Ware JE, Kosinski M, Keller SD: How to score the SF-12 Phy-sical & Mental Health Summary Scales. Lincoln, RI: QualityMetric Incorporated, 2002

51. Whyte RI, Crownover R, Murphy MJ, Martin DP, Rice TW,DeCamp MM Jr, et al: Stereotactic radiosurgery for lung tu-mors: preliminary report of a phase I trial. Ann Thorac Surg75:10971101, 2003

Manuscript received February 13, 2004.Accepted in final form February 9, 2005.

Address reprint requests to: Fraser C. Henderson, M.D., Depart-ment of Neurosurgery 1 PHC, Georgetown University Hospital,3800 Reservoir Road, NW, Washington, DC 20007. email: [email protected].



549

Documents

degen_spinepain_jneurosurg_2005