Int. J. Radiation Oncology Biol. Phys., Vol. 79, No. 4, pp. 1151–1157, 2011Copyright � 2011 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/$–see front matter

jrobp.2009.12.038

doi:10.1016/j.i

CLINICAL INVESTIGATION Metastasis

PREDICTORS OF LOCAL CONTROL AFTER SINGLE-DOSE STEREOTACTICIMAGE-GUIDED INTENSITY-MODULATED RADIOTHERAPY FOR

EXTRACRANIAL METASTASES

CARLO GRECO, M.D.,* MICHAEL J. ZELEFSKY, M.D.,* MICHAEL LOVELOCK, PH.D.,y ZVI FUKS, M.D.,*

MARGIE HUNT, M.S.,y KENNETH ROSENZWEIG, M.D.,* JOAN ZATCKY, B.S., N.P.,* BALEM KIM, B.A.,*

AND YOSHIYA YAMADA, M.D.*

Departments of *Radiation Oncology and yMedical Physics, Memorial Sloan-Kettering Cancer Center, New York, NY

ReprinRadiation1275 Yor639-6802

Purpose: To report tumor local control after treatment with single-dose image-guided intensity-modulated radio-therapy (SD-IGRT) to extracranial metastatic sites.Methods and Materials: A total of 126 metastases in 103 patients were treated with SD-IGRT to prescription dosesof 18–24 Gy (median, 24 Gy) between 2004 and 2007.Results: The overall actuarial local relapse–free survival (LRFS) rate was 64% at a median follow-up of 18 months(range, 2–45 months). The median time to failure was 9.6 months (range, 1–23 months). On univariate analysis,LRFS was significantly correlated with prescription dose (p = 0.029). Stratification by dose into high (23 to 24Gy), intermediate (21 to 22 Gy), and low (18 to 20 Gy) dose levels revealed highly significant differences inLRFS between high (82%) and low doses (25%) (p < 0.0001). Overall, histology had no significant effect onLRFS (p = 0.16). Renal cell histology displayed a profound dose–response effect, with 80% LRFS at the highdose level (23 to 24 Gy) vs. 37% with low doses (#22 Gy) (p = 0.04). However, for patients who received thehigh dose level, histology was not a statistically significant predictor of LRFS (p = 0.90). Target organ (bone vs.lymph node vs. soft tissues) (p = 0.5) and planning target volume size (p = 0.55) were not found to be associatedwith long-term LRFS probability. Multivariate Cox regression analysis confirmed prescription dose to be a signif-icant predictor of LRFS (p = 0.003).Conclusion: High-dose SD-IGRT is a noninvasive procedure resulting in high probability of local tumor control.Single-dose IGRT may be effectively used to locally control metastatic deposits regardless of histology and targetorgan, provided sufficiently high doses (> 22 Gy) of radiation are delivered. � 2011 Elsevier Inc.

Image-guided radiotherapy, IMRT, Single fraction, Metastases.

INTRODUCTION

One of the basic tenets of classic clinical radiobiology is that

fractionation of the radiation optimizes outcomes because it

affords protection to normal tissues relative to the targeted tu-

mor. For radiotherapy to extracranial disease sites, the pres-

ence of dosimetric and treatment delivery uncertainties,

largely due to organ motion, have traditionally necessitated

the use of considerable safety margins around the target vol-

ume, leading to greater exposure of normal tissues to higher

radiation doses. Until recently, in the extracranial setting,

treatment with ultra-hypofractionated or single-dose regi-

mens has often been regarded as unfeasible owing to the

potential increased risk for radiation-induced complications.

During the last 5–10 years, advances in radiotherapy plan-

ning and delivery have led to a major paradigm shift in radio-

therapy practice. Treatment plans featuring steep dose

t requests to: Michael J. Zelefsky, M.D., Department ofOncology, Memorial Sloan-Kettering Cancer Center,

k Avenue, Box 22, New York, NY 10065. Tel: (212); Fax: (212) 639-8876; E-mail: zelefskm@mskcc.org

1151

gradients at the planning target volume (PTV) edge can be

achieved with intensity modulation of the radiation beams,

thereby maximally reducing exposure to adjacent nontarget

tissues. More recently, the advent of image-guided radiother-

apy (IGRT) has made it possible to achieve near-real-time

quality assurance of the treatment plan, enabling safe deliv-

ery of unprecedented dose exposures (>10 Gy) with single-

dose or ultra-hypofractionated regimens (1). Evidence of

the feasibility and effectiveness of this approach is rapidly

emerging in the scientific literature, mostly in the field of

hypofractionated stereotactic body radiotherapy (SBRT).

Reports on SBRT for lung, liver, and bone metastases have

consistently shown improved tumor-control outcomes with

regimens using higher hypofractionated biologically equiva-

lent doses compared with lower dose levels (2–7). Hereto-

fore, published clinical experience using single-dose

Conflict of interest: none.Received Sept 30, 2009, and in revised form Dec 4, 2009.

Accepted for publication Dec 4, 2009.

Table 1. Patient characteristics

All lesions (n = 124) n

GenderMale 71Female 32

Age (y), median (range) 64 (33–91)Treatment site

Bone 94Lymph node 14Lung 8Liver 6Other soft tissues 2

Histologic typeProstate 42Renal cell 35Colorectal 15Sarcoma 5Non–small cell lung cancer 4Cholangiocarcinoma 4Breast 3Melanoma 3Bladder 2Leydig cell 2Small-cell lung cancer 2Thyroid cancer 2Esophageal 1Germ cell 1Chordoma 1Ovarian 1Pancreatic 1

1152 I. J. Radiation Oncology d Biology d Physics Volume 79, Number 4, 2011

image-guided intensity-modulated radiotherapy (SD-IGRT)

has been far more limited, and the optimal dose required

for local tumor control is not yet defined (8, 9). Stereotactic

radiosurgery of intracranial tumors using single doses of

18–24 Gy is an established mode of treatment and results

in remarkable outcomes. A dose–response relationship has

been reported for single-dose stereotactic radiosurgery of

brain metastases, with approximately 50% freedom from lo-

cal relapse at 1 to 2 years after 15–18 Gy compared with

$80% after 22–24 Gy, regardless of the histologic phenotype

of the primary tumor (10, 11). Similar dose–response data for

extracranial sites are not available. Recent studies using sin-

gle doses of 24 to 25 Gy have shown $80% local control

rates for inoperable primary lung cancer (12–14) and oligo-

metastatic disease (4, 15, 16). Results from our institution re-

ported by Yamada et al. (16) have shown improved outcomes

for metastatic spinal lesions using SD-IGRT of $22 Gy

compared with lower doses. The present study reports on

our experience using SD-IGRT for nonspinal lesions; initial

single-dose levels used ranged from 18 to 20 Gy, and subse-

quently, as part of a Phase I dose-escalation study, doses of

22 and 24 Gy were used. Our findings indicate that higher

doses were associated with improved and durable tumor-con-

trol outcomes irrespective of histology and lesion size.

METHODS AND MATERIALS

Between February 2004 and December 2007, 103 consecutive pa-

tients with a total of 124 lesions underwent SD-IGRT directed to fo-

cal metastatic disease within bone, lymph nodes, or soft tissue using

prescription doses ranging between 18 Gy and 24 Gy. Patients were

initially treated to dose levels of 18–20 Gy and, beginning in 2006,

a Phase I dose-escalation study was activated at 22 Gy. The intent of

this dose-escalation study was to determine the maximally tolerated

dose of single-fraction high-dose IGRT to soft tissues, lymph node,

and bone lesions. Once 20 patients were treated at this dose level

with a minimum follow-up of 6 months, subsequent eligible patients

were accrued to the next dose level of 24 Gy. Although the Phase I

trial is currently accruing patients at the 26-Gy level, the present re-

port includes patients treated with dose ranges of 18–24 Gy until De-

cember 2007. The study was internally approved by the Memorial

Sloan-Kettering Cancer Center Research Board, and all patients

signed informed consent forms. No patient had undergone surgical

resection of the lesion of interest or prior radiotherapy to this region,

and adequate cross-sectional imaging was acquired before treat-

ment.

The clinical characteristics of these patients are summarized in

Table 1. The most frequently treated metastatic lesion was of pros-

tate origin (n = 42), followed by kidney (n = 35) and colorectal (n =

15). The median follow-up was 18 months (range, 2–45 months).

No patient has been lost to follow-up.

Treatment planningThe techniques used for treatment planning and delivery of IGRT

at our institution have been previously described (16, 17). All pa-

tients were immobilized in a customized cradle developed at our in-

stitution for IGRT to prevent any inadvertent patient motion. Before

simulation, if deemed necessary, the patient underwent implantation

of radio-opaque fiducial markers in the vicinity of the target lesion to

ensure target localization before treatment delivery. Patients were

simulated in the cradle using 2-mm slice thickness CT images. Dur-

ing treatment, patient movement was monitored with infrared ste-

reoscopic cameras, and if any motion was noted the treatment was

temporarily stopped and only continued when movement was no

longer observed.

Prescription doseTreatment planning was generated on in-house software using an

inverse treatment algorithm. The dose was prescribed to the 100%

maximum isodose line, which completely encompassed the PTV.

The prescribed doses in these patients ranged between 18 and

24 Gy. The planning target volume was created with a 2-mm expan-

sion around the clinical target volume. Dose constraints of <12 Gy

were set as the maximum allowable dose to the spinal cord contour

and of #16 Gy to bowel, rectum, bladder, and other identified rele-

vant normal tissue structures. This was determined using a standard-

ized best-fit inverse optimization process that takes into account

normal tissue constraints, clinical target volume, and PTV coverage.

Typically, seven to nine coplanar fields were set to a single isocenter

using dynamic multileaf collimation. Six- and/or 15-MV photons

were used to provide optimal coverage and normal tissue sparing.

IGRT techniqueImage verification was performed by creating digitally recon-

structed radiographs from the simulation studies for each field’s

beam’s eye view, which were used as the gold standard of compar-

ison. Linear accelerator on-board cone-beam CT was used for veri-

fication. These images were digitally overlaid with the reference

images to calculate isocenter corrections. The calculated corrections

were then verified for each field before actual treatment.

Table 2. Distribution by prescription dose (all lesions,n = 124)

Planning target volume dose (Gy) n

18 1020 221 322 3823 124 70

Fig. 1. Actuarial local control (Kaplan-Meir method) by dose level.Y axis represents local relapse-free survival (%).

Fig. 2. Actuarial local control (Kaplan-Meir method) by histology.Y axis represents local relapse-free survival (%).

Predictors of local control after SD-IGRT for extracranial metastases d C. GRECO et al. 1153

Intrafraction motion was monitored with infrared stereoscopic cam-

eras. Four spherical infrared reflectors were placed on the patient

above and below the treated region to track intrafractional motion.

If motion >2 mm was detected, treatment was temporarily stopped

and the position was verified again with both two-dimensional kilo-

voltage and three-dimensional beam CT imaging for accurate repo-

sitioning. A cone-beam CT scan was also acquired to verify

positioning after treatment was completed.

EndpointsIn general, patients were examined 8 weeks after treatment.

Repeat diagnostic CT, magnetic resonance imaging, and positron

emission tomography/CT scans were acquired at approximately 3

to 4-month intervals to assess local control for the first 2 years. Local

response to treatment was scored as complete response, stable

disease, or failure on the basis of a thorough assessment of all

cross-sectional imaging/metabolic studies available. Local failure

was scored as an event if a lesion increased in size by $20% accord-

ing to the Response Evaluation Criteria in Solid Tumors and when-

ever persistence of disease was confirmed pathologically (18).

Estimates of local relapse-free survival (LRFS) were calculated us-

ing the Kaplan-Meier method (19).

RESULTS

The overall 2-year actuarial local control rate for all treated

lesions was 64%. The median time to local failure was 9.6

months (range, 0.8–23.5 months). Overall, 76% of the treated

lesions (95 of 124) showed durable local control. Complete

responses occurred in 22% of treatments (21 of 95) (1 lung

nodule, 9 lymph nodes, 11 skeletal lesions). No complete re-

sponses (0 of 12) were observed at lower dose levels (18–20

Gy), 25% (1 of 4) occurred at 21 Gy, 21% (8 of 38) at 22 Gy,

and 17% (12 of 71) at 24 Gy (p = 0.15). Of the 29 local fail-

ures, distribution by histology was as follows: 5 prostate, 8

renal cell, 8 colorectal, 4 sarcomas, 1 non–small-cell lung

cancer, 1 small-cell lung cancer, 1 breast, 1 ovarian. Colorec-

tal lesions fared significantly worse (p < 0.001) than prostate

vs. renal cell vs. all other histologies grouped together. In par-

ticular, four of five colorectal liver metastases recurred

despite receiving a prescription dose of 24 Gy.

Effect of doseThe distribution of numbers of cases at each prescription

dose is summarized in Table 2. Analysis of the effect of

dose on local response was carried out on a total of 118 le-

sions (with the exclusion of liver lesions). On univariate anal-

ysis, prescription dose significantly correlated with LRFS

(p = 0.029). For this analysis patients were stratified into three

dose groups: high (23 to 24 Gy), intermediate (21 to 22 Gy),

and low (18–20 Gy). Figure 1 shows actuarial LRFS curves

as a function of dose. The difference between high dose

(82%) vs. low dose (25%) was highly significant (p <

0.0001), indicating a steep dose–response curve. The local

control rates for lesions treated with intermediate (69%) vs.

low doses (25%) was also significantly different (p = 0.04).

HistologyFigure 2 demonstrates the Kaplan-Meier overall probabil-

ities of LRFS by histology (p = 0.16).

Because of the relatively small number of cases treated at

doses between 18 and 20 Gy, subset analysis of the effect of

dose on different histologic phenotypes was carried out di-

chotomically for high (23 to 24 Gy) vs. low (#22 Gy) dose

levels. No correlation of dose and LRFS probability was ob-

served for prostate cancer (not shown; p = 0.8), whereas renal

cell histology (Fig. 3) displayed significantly improved local

control at higher doses (80%) vs. lower doses (37%) (p =

0.04). For the highest dose level only, analysis of LRFS prob-

abilities showed no statistical significance as a function of

histology (p = 0.90) (Fig. 4), with local control rates of

Fig. 3. Actuarial local control (Kaplan-Meir method) showing theeffect of high dose vs. lower doses in renal cell histology. Y axis rep-resents local relapse-free survival (%).

Fig. 5. Actuarial local control (Kaplan-Meir method) as a functionof lesion location. Y axis represents local relapse-free survival (%;p = 0.50).

1154 I. J. Radiation Oncology d Biology d Physics Volume 79, Number 4, 2011

85%, 80%, 75%, and 72% for prostate, renal cell, colorectal,

and other histologies, respectively.

Target organsNinety-four lesions were osseous, 14 were located in

lymph nodes, 8 in the lungs, 6 in the liver, 1 in the adrenal

gland, and 1 was subcutaneous. Four of the liver lesions re-

curred early on, and no patients with hepatic metastases

were alive beyond the median survival of 18 months owing

to systemic progression. Fig. 5 shows the Kaplan-Meier

LRFS probabilities for bone, lymph node, and lung lesions

(p = 0.5).

The relatively large number of bone lesions allowed fur-

ther analysis of the effect of dose and histology in this site.

A positive association (Fig. 6) between dose and actuarial

LRFS was observed in bone metastases (p = 0.019). More-

over, the difference between high dose (83%) vs. intermedi-

ate dose (60%) was also significant (p = 0.04). At the high

dose level, 2-year actuarial LRFS probabilities of osseous

metastases were 86% for prostate, 80% for renal cell, 75%

for colorectal, and 83% for all other histologies combined

(p = 0.89).

Fig. 4. Actuarial local control (Kaplan-Meir method) at the highdose level for all histologies. Y axis represents local relapse-free sur-vival (%; p = 0.90).

Effect of lesion sizeMean PTV volume was 104.8 cm3 (median, 54.9 cm3;

range, 8.5–1150 cm3). No correlation between PTV and

LRFS as a function of prescribed dose was found (p = 0.55).

Multivariate analysisMultivariate analysis was performed using the Cox propor-

tional hazards multiple regression model with prescription

dose, PTV size, histology and target organ as covariates. Pre-

scription dose retained significance as a predictor of long-

term LRFS (p = 0.003).

ToxicityTreatment was generally well tolerated. Gastrointestinal

Grade 3 acute toxicity was observed in 2 cases (1 patient

treated to a dose of 24 Gy for a liver metastasis, and 1 treated

to 22 Gy for a T6 lesion). Grade $2 late toxicity was ob-

served in 11 of 103 patients. Grade 2 dermatitis with telan-

gectasias was observed in 4 cases. One case of Grade 3

gastrointestinal toxicity (esophageal stricture) occurred after

severe acute toxicity after 22 Gy. Grade 2 radiculopathy was

Fig. 6. Actuarial local control (Kaplan-Meir method) of bone le-sions as a function of dose. Y axis represents local relapse-free sur-vival (%).

Predictors of local control after SD-IGRT for extracranial metastases d C. GRECO et al. 1155

observed in 3 cases. Three patients treated in the pelvic re-

gion developed a radiation-induced Grade 3 peripheral neu-

ritis syndrome (1 at 22 Gy and 2 at 24 Gy). Two of three

cases gradually improved with time and aggressive physical

therapy interventions. The overall incidence of Grade 3 late

toxicity was <4%.

DISCUSSION

Similar to other recent reports, the results of the present

analysis confirm that SD-IGRT is a powerful tool for achiev-

ing long-term local control of metastatic lesions. One of the

most important findings in the present series was the recogni-

tion of a dose–response relationship, with a highly statisti-

cally significant difference in tumor control between high

vs. low single doses (p < 0.0001). A significant difference be-

tween intermediate and low dose (p = 0.04) was also ob-

served, indicating the presence of a steep dose–response

curve. These findings are consistent with results from our in-

stitution published by Yamada et al. (16) which showed

higher doses (24 Gy vs. #23 Gy) to be a predictor of

improved local control (p = 0.03) for metastatic spinal

lesions.

It is noteworthy that an overall actuarial LRFS of 82% was

observed at the high dose level regardless of tumor histology,

lesion size, and target tissue. Analysis of the effect of histol-

ogy on the likelihood of local control revealed a pronounced

dose effect for renal cell histologies, a known radioresistant

histology according to classic radiobiologic ranking. A sim-

ilar effect was not observed for any other histology, likely

owing to the relatively small number of lesions treated at

lower doses in this series. This finding, however, is of great

interest in light of the fact that inherent radioresistance of re-

nal cell cancer seemed to be overcome at a dose of 24 Gy, for

which LRFS rates similar to those of other histologies were

achieved.

The finding of poor local control rates for colorectal liver

metastases after high-dose single-fraction irradiation is not

unexpected. Herfarth et al. (15) carried out a prospective

Phase I/II trial for limited hepatic metastases with dose esca-

lation from 14 Gy to 26 Gy using a single-fraction regimen.

In the initial report actuarial freedom from local failure for the

entire group was 67% at 18 months, but in a subsequent up-

date of the study higher failure rates were observed, suggest-

ing that the single-fraction dose level should be further

escalated (3). Wulf et al. (4), however, reported an 82% actu-

arial local control at 24 months after a single fraction of 26 Gy

in 9 patients. This finding may be due to the steep dose–

response curve of single-dose irradiation, whereby relatively

moderate increases may lead to significant improvements in

outcome. The disappointing local control for colorectal liver

metastases observed in our limited series warrants further in-

vestigation of its underlying causes. Aside from liver loca-

tions, in our cohort, colorectal metastases seemed to

respond relatively well, suggesting that colorectal histology

per se is likely not responsible for the poor outcome.

The findings of the present report and of similar studies re-

porting on high-dose-per-fraction irradiation defy the princi-

ples of classic radiobiology. There are abundant experimental

and clinical data to show that the standard linear-quadratic

(LQ) model performs well for comparison of isoeffective

doses in fractionated radiotherapy up to 8 Gy per fraction.

However, the biologically equivalent doses calculated by

LQ modeling are inconsistent with the clinical outcomes of

1–5 fractions exceeding 12 Gy each (20). It has been disputed

that the LQ model overestimates cell killing at high single

doses (21). Moreover, if the effect of hypoxia is taken into ac-

count, a single dose of 20 Gy is expected to result in a surviv-

ing fraction of 10�3 with 20% hypoxic tumor cells vs.

a survival of 10�6 after 30 fractions of 2 Gy, assuming full

reoxygenation of the tumor (22). On the basis of these as-

sumptions, 40 Gy in a single fraction would be isoeffective

to 60 Gy in 2-Gy fractions (23). Experimental models and

emerging clinical data, however, consistently show that sig-

nificantly lower single exposures are able to achieve high lo-

cal control rates, leading to the hypothesis that the underlying

mechanisms of tumor-cell killing may be different from frac-

tionated radiotherapy.

The biologic foundation of radiation-induced cell death af-

ter high-dose radiation exposure (>10 Gy per fraction) is be-

ginning to be unraveled. Recent studies have challenged the

prevailing hypothesis that at high-dose exposures tissue stem

cell clonogens represent exclusive and independent targets

for radiation-induced tissue damage through reproductive

cell death. There is emerging evidence that tumor endothe-

lium may be impacted to a greater degree with high radiation

doses, resulting in apoptosis, microvasculature dysfunction,

and cell death through parallel and independent pathways

(24–27).

The efficacy of SD-IGRT seems to be durable, suggesting

that the patients most likely to benefit are those with limited

metastatic disease. The beneficial impact of achieving local

control in patients with limited metastases is corroborated

by extensive surgical series of complete excision of brain,

pulmonary, and liver metastases with long-term survival

(28–33). The term oligometastases was coined to indicate

a state of limited metastatic progression in which overt wide-

spread dissemination of the disease has not yet occurred (34).

According to this hypothesis, a disease state, comprised be-

tween locoregionally confined and widely disseminated,

may exist early on in the evolution of metastatic progression,

in which only a limited number of lesions and/or destination

organs are present (35–37). A window of therapeutic oppor-

tunity, therefore, may be exploited in carefully selected pa-

tients who may still benefit from a local form of treatment.

This therapeutic approach may achieve long-term survival

of patients who have been historically considered incurable.

Indeed, if the detected lesions represent the only sites of dis-

tant disease, effective local therapy may result in cure (38,

39) and potentially prevent further disease dissemination

(40).

In our experience, SD-IGRT is extremely well tolerated

and results in minimal radiation-induced morbidity, provided

1156 I. J. Radiation Oncology d Biology d Physics Volume 79, Number 4, 2011

strict dose/volume constraints for the critical structures are

complied with. Overall, Grade 3 complications were <4%

in this series. The instance of a Grade 3 esophageal stricture

occurred in early 2005, before adequate dose/volume con-

straints had been established. A thorough analysis of thoracic

toxicity in paraspinal SD-IGRT recently carried out at our in-

stitution showed a low rate of treatment-related toxicity and

overall safety of this therapeutic approach (41).

Because treatment is administered in 1 day, SD-IGRT also

has practical advantages and seems particularly appealing in

the context of metastatic disease (42, 43). Furthermore, SD-

IGRT is also less likely to conflict with systemic therapy

schedules. As treatment procedures become more stream-

lined, single-fraction radiotherapy will allow improved pa-

tient throughput per treatment machine.

CONCLUSIONS

The present series suggests that SD-IGRT is a highly effec-

tive intervention regardless of target tissue, lesion size, and

histology, provided high enough doses of radiation are deliv-

ered. Further investigation on the clinical role of SD-IGRT is

warranted to ascertain whether dose escalation beyond 24 Gy

improves outcome in specific clinical settings.

REFERENCES

1. Jaffray DA. Image-guided radiation therapy: From concept topractice. Semin Radiat Oncol 2007;17:243–244.

2. Wulf J, Baier K, Mueller G, et al. Dose-response in stereotacticirradiation of lung tumors. Radiother Oncol 2005;77:83–87.

3. Herfarth KK, Debus J, Wannenmacher M. Stereotactic radiationtherapy of liver metastases: Update of the initial phase-I/II trial.Front Radiat Ther Oncol 2004;38:100–105.

4. Wulf J, Guckenberger M, Haedinger U, et al. Stereotactic radio-therapy of primary liver cancer and hepatic metastases. ActaOncol 2006;45:838–847.

5. Casamassima F, Masi L, Bonucci I, et al. Relevance of biolog-ically equivalent dose values in outcome evaluation of stereo-tactic radiotherapy for lung nodules. Int J Radiat Oncol BiolPhys 2008;71:145–151.

6. Guckenberger M, Wulf J, Mueller G, et al. Dose-response rela-tionship for image-guided stereotactic body radiotherapy of pul-monary tumors: Relevance of 4d dose calculation. Int J RadiatOncol Biol Phys 2009;74:47–54.

7. McCammon R, Schefter TE, Gaspar LE, et al. Observation ofa dose-control relationship for lung and liver tumors after stereo-tactic body radiation therapy. Int J Radiat Oncol Biol Phys2009;73:112–118.

8. Park C, Papiez L, Zhang S, et al. Universal survival curve andsingle fraction equivalent dose: Useful tools in understandingpotency of ablative radiotherapy. Int J Radiat Oncol Biol Phys2008;70:847–852.

9. Tome WA. Universal survival curve and single fraction equiv-alent dose: Useful tools in understanding potency of ablative ra-diotherapy: In regard to Parks et al. (Int J Radiat Oncol BiolPhys 2008;72:1620-1621). Int J Radiat Oncol Biol Phys 2009;73:1286.

10. Shiau CY, Sneed PK, Shu HK, et al. Radiosurgery for brain me-tastases: Relationship of dose and pattern of enhancement to localcontrol. Int J Radiat Oncol Biol Phys 1997;37:375–383.

11. Vogelbaum MA, Angelov L, Lee SY, et al. Local control ofbrain metastases by stereotactic radiosurgery in relation todose to the tumor margin. J Neurosurg 2006;104:907–912.

12. Fritz P, Kraus HJ, Blaschke T, et al. Stereotactic, high single-dose irradiation of stage I non-small cell lung cancer (NSCLC)using four-dimensional CT scans for treatment planning. LungCancer 2008;60:193–199.

13. Hara R, Itami J, Kondo T, et al. Clinical outcomes of single-fraction stereotactic radiation therapy of lung tumors. Cancer2006;106:1347–1352.

14. Le QT, Loo BW, Ho A, et al. Results of a phase I dose-escala-tion study using single-fraction stereotactic radiotherapy forlung tumors. J Thorac Oncol 2006;1:802–809.

15. Herfarth KK, Debus J, Lohr F, et al. Stereotactic single-dose ra-diation therapy of liver tumors: Results of a phase I/II trial. JClin Oncol 2001;19:164–170.

16. Yamada Y, Bilsky MH, Lovelock DM, et al. High-dose, single-fraction image-guided intensity-modulated radiotherapy formetastatic spinal lesions. Int J Radiat Oncol Biol Phys 2008;71:484–490.

17. Lovelock DM, Hua C, Wang P, et al. Accurate setup of paraspi-nal patients using a noninvasive patient immobilization cradleand portal imaging. Med Phys 2005;32:2606–2614.

18. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelinesto evaluate the response to treatment in solid tumors. EuropeanOrganization for Research and Treatment of Cancer, NationalCancer Institute of the United States, National Cancer Instituteof Canada. J Natl Cancer Inst 2000;92:205–216.

19. Kaplan EL, Meier P. Nonparametric estimations from incom-plete observations. Am J Stat Assos 1958;53:457–481.

20. Fowler JF, Tome WA, Fenwick JD, et al. A challenge to tradi-tional radiation oncology. Int J Radiat Oncol Biol Phys 2004;60:1241–1256.

21. Guerrero M, Li XA. Extending the linear-quadratic model forlarge fraction doses pertinent to stereotactic radiotherapy.Phys Med Biol 2004;49:4825–4835.

22. Wouters BG, Brown JM. Cells at intermediate oxygen levelscan be more important than the ‘‘hypoxic fraction’’ in determin-ing tumor response to fractionated radiotherapy. Radiat Res1997;147:541–550.

23. Brown JM, Koong AC. High-dose single-fraction radiotherapy:Exploiting a new biology? Int J Radiat Oncol Biol Phys 2008;71:324–325.

24. Fuks Z, Kolesnick R. Engaging the vascular component of thetumor response. Cancer Cell 2005;8:89–91.

25. Chmura SJ, Nodzenski E, Kharbanda S, et al. Down-regulationof ceramide production abrogates ionizing radiation-inducedcytochrome c release and apoptosis. Mol Pharmacol 2000;57:792–796.

26. Deng X, Yin X, Allan R, et al. Ceramide biogenesis is requiredfor radiation-induced apoptosis in the germ line of. C. elegans.Science 2008;322:110–115.

27. Garcia-Barros M, Paris F, Cordon-Cardo C, et al. Tumor re-sponse to radiotherapy regulated by endothelial cell apoptosis.Science 2003;300:1155–1159.

28. Long-term results of lung metastasectomy. Prognostic analysesbased on 5206 cases. The International Registry of Lung Metas-tases. J Thorac Cardiovasc Surg 1997;113:37–49.

29. Headrick JR, Miller DL, Nagorney DM, et al. Surgical treat-ment of hepatic and pulmonary metastases from colon cancer.Ann Thorac Surg 2001;71:975–979.

30. McCormack PM, Burt ME, Bains MS, et al. Lung resection forcolorectal metastases. 10-year results. Arch Surg 1992;127:1403–1406.

31. Regine WF, Rogozinska A, Kryscio RJ, et al. Recursive parti-tioning analysis classifications I and II: Applicability evaluated

Predictors of local control after SD-IGRT for extracranial metastases d C. GRECO et al. 1157

in a randomized trial for resected single brain metastases. Am JClin Oncol 2004;27:505–509.

32. Selzner M, Morse MA, Vredenburgh JJ, et al. Liver metastasesfrom breast cancer: Long-term survival after curative resection.Surgery 2000;127:383–389.

33. Simmonds PC, Primrose JN, Colquitt JL, et al. Surgical re-section of hepatic metastases from colorectal cancer: A sys-tematic review of published studies. Br J Cancer 2006;94:982–999.

34. Hellman S, Weichselbaum RR. Oligometastases. J Clin Oncol1995;13:8–10.

35. Hellman S, Weichselbaum RR. Importance of local control inan era of systemic therapy. Nat Clin Pract Oncol 2005;2:60–61.

36. Rubin P, Brasacchio R, Katz A. Solitary metastases: Illusionversus reality. Semin Radiat Oncol 2006;16:120–130.

37. Tait CR, Waterworth A, Loncaster J, et al. The oligometa-static state in breast cancer: Hypothesis or reality. Breast2005;14:87–93.

38. Ben-Josef E, Lawrence TS. Using a bigger hammer: The role ofstereotactic body radiotherapy in the management of oligometa-stases. J Clin Oncol 2009;27:1537–1539.

39. Macdermed DM, Weichselbaum RR, Salama JK. A rationalefor the targeted treatment of oligometastases with radiotherapy.J Surg Oncol 2008;98:202–206.

40. Withers HR, Lee SP. Modeling growth kinetics and statisticaldistribution of oligometastases. Semin Radiat Oncol 2006;16:111–119.

41. Gomez DR, Hunt MA, Jackson A, et al. Low rate of thoracictoxicity in palliative paraspinal single-fraction stereotacticbodyradiation therapy. Radiat Oncol 2009;93:414–418.

42. Carey SM, Katz A, Constine LS. Stereotactic body radiationtherapy for extracranial oligometastases: Does the sword havea double edge? Semin Radiat Oncol 2006;16:67–76.

43. Kavanagh BD, McGarry RC, Timmerman RD. Extracranial ra-diosurgery (stereotactic body radiation therapy) for oligometa-stases. Semin Radiat Oncol 2006;16:77–84.