STATE OF THE ART: CONCISE REVIEW

Practical Considerations Arising from the Implementationof Lung Stereotactic Body Radiation Therapy (SBRT) at a

Comprehensive Cancer Center

Max Dahele, MBChB,*† Shannon Pearson, MRT(T),* Tom Purdie, PhD,*†Jean-Pierre Bissonnette, PhD,*† Kevin Franks, MBChB,‡ Anthony Brade, MDCM,*†

John Cho, MD,*† Alex Sun, MD,*† Andrew Hope, MD,*† Andrea Marshall, MRT(T),*Jane Higgins, MRT(T),*† and Andrea Bezjak, MDCM*†

Introduction: With the anticipation of improved outcomes, espe-cially for patients with early-stage non-small cell lung cancer,stereotactic body radiation therapy (SBRT) has been rapidly intro-duced into the thoracic radiation oncology community. Although atfirst glance lung SBRT might seem methodologically similar toconventional radiotherapy, there are important differences in itsexecution that require particular consideration. The objective of thispaper is to highlight these and other issues to contribute to the safeand effective diffusion of lung SBRT. We discuss practical chal-lenges that have been encountered in the implementation of lungSBRT at a single, large institution and emphasize the importance ofa systematic approach to the design of lung SBRT services.Methods: Specific technical and clinical components that wereidentified as being important during the development of lung SBRTat Princess Margaret Hospital are described. The clinical system thatevolved from these is outlined.Results: Using this clinical framework the practical topics ad-dressed include: patient assessment, simulation and treatment plan-ning, tumor and organ at risk delineation, trial set up before treat-ment, on-line image-guidance, and patient follow-up.Conclusions: The potential gain in therapeutic ratio that is theoret-ically possible with lung SBRT can only be realized if the tumor isadequately irradiated and normal tissue spared. A discussion of thecomponent parts of lung SBRT is presented. It is a complex processand specific challenges need to be overcome to effect the satisfac-tory transition of lung SBRT into routine practice.

Key Words: Stereotactic body radiation therapy, Lung cancer,Clinical process, Quality assurance.

(J Thorac Oncol. 2008;3: 1332–1341)

There is increasing interest among the oncology commu-nity in the potential advantages of hypo-fractionated,

high-precision radiation therapy, which are embodied by‘Stereotactic Body Radiation Therapy’ (SBRT). Lung SBRThas been introduced for patients with early-stage non-smallcell lung cancer (NSCLC) and more recently, pulmonarymetastases from other primary tumors. It represents an evo-lution of rigid frame intracranial radio-surgery techniquesthat were initially developed at the Karolinska Institute inSweden in the 1950s.1 From the mid 1990s onwards, reportsof extracranial SBRT appeared from groups in Europe, Japan,and North America.2–4 Most of the data to support lung SBRThas been accumulated for patients with medically inoperableNSCLC and in contrast to what has been achieved with conven-tionally fractionated radiotherapy (RT),5 initial local control andsurvival results have been very encouraging.6,7 Indeed lungSBRT is already being discussed as a potential option forpatients with medically operable NSCLC.6 Many dose-frac-tionation regimens for NSCLC SBRT are in use, some of themost common are 48 Gy in 4 fractions (Japan), 54 or 60 Gyin 3 fractions (North America), and 60 Gy in 5 or 8 fractions(Europe).4,8,9 Biologically, these represent very high, ablativedoses of radiation that exceed the conventional tolerance ofcertain normal tissues. By way of comparison (and acknowl-edging the uncertainty in applying such formulae to ex-tremely hypo-fractionated RT), 70 Gy fractionated in 35fractions, represents an estimated biologically effective doseto the tumor of 84 Gy (�/� � 10), compared with 105 to 132Gy for 48 Gy in 4 fractions or 60 Gy in 5 or 8 fractions and180 Gy for 60 Gy in 3 fractions.

The acute toxicity associated with lung SBRT is typi-cally mild and short-lived. Our own experience at PrincessMargaret Hospital (PMH) is that grade 1–2 fatigue, cough,chest wall pain, and skin erythema are the most common sideeffects in the first 3 months after treatment. Although the

*Radiation Medicine Program, Princess Margaret Hospital, UniversityHealth Network, Toronto, Canada; †Department of Radiation Oncology,University of Toronto, Toronto, Ontario, Canada; and ‡St. James Insti-tute of Oncology, The Leeds Teaching Hospitals NHS Trust, Leeds,England.

Disclosure: Partial funding supplied by Elekta Synergy Research Group andAddie MacNaughton Chair in Lung Cancer Radiotherapy.

Address for correspondence: Dr. Andrea Bezjak, Addie MacNaughton Chairin Lung Cancer Radiotherapy, Princess Margaret Hospital, 610 Univer-sity Avenue, Toronto, ON, M5G 2M9, Canada. E-mail: andrea.bezjak@rmp.uhn.on.ca

Copyright © 2008 by the International Association for the Study of LungCancerISSN: 1556-0864/08/0311-1332

Journal of Thoracic Oncology • Volume 3, Number 11, November 20081332

radiobiology of SBRT is incompletely understood,10 there arenow several reports containing somewhat longer follow-updata for lung SBRT6,11,12 that permit some general observa-tions about late toxicity. For very peripheral tumors, incor-porating chest wall structures in the high-dose volume canlead to side effects that include rib fracture and chest wallpain.12 Serious toxicity has been attributed to radiation dosesof the order of 60 Gy in 3 fractions to the central airways,11

and there are individual reports of other toxicities includingfatal esophageal ulceration13 and pneumonitis.14 Sometimes itcan be difficult to ascertain the relative contributions of totaldose, fractionation, and treatment technique (including treat-ment planning margins and image-guidance) to toxicity.

Many treatment centers now possess the technical ca-pability to deliver lung SBRT. Nevertheless variation in itsdelivery15 argues for initiatives that aim to promote effectivelearning and dialogue between institutions and enhance theclinical diffusion of lung SBRT. The aim of this report is todescribe the experience in one institution that has imple-mented lung SBRT, and in so doing, try to aid in a practicalway, the fast-occurring transition of lung SBRT from clinicaltrials to routine clinical treatment. Specific aspects of design-ing a clinical system to support the safe and effective deliveryof lung SBRT are discussed. Historical aspects, detailedtechnical issues, dose selection, or outcomes relating toSBRT are not comprehensively reviewed. For such topics, thereader is referred to several other publications e.g.,6,12,16–19 orto individual SBRT protocols. This manuscript does notconstitute a ‘recipe’ for thoracic SBRT and technical detailsare presented to illustrate specific scenarios encountered inthe implementation of SBRT.

PATIENTS AND METHODSThe lung SBRT program at our institution, a large

academic comprehensive cancer center in Canada, com-menced in fall 2004, and has evolved over the last 3 years.Important technical and clinical issues that had to be ad-dressed were identified early on and the treatment processdeveloped around them. Many of the issues are generic andcan be grouped under the following headings:

Technical RequirementsFor SBRT to be successful, it requires accuracy and

precision to consistently irradiate the tumor while simulta-neously sparing organs at risk (OAR). In recognition of this,treatment planning and delivery needs to incorporate thefollowing features: (i) reliable and consistent patient immo-bilization, for example with a rigid stereotactic body frame or‘frameless’ semi-rigid evacuated bag, that is comfortableenough to accommodate extended treatment sessions, (ii)high-precision imaging for radiation treatment planning thatfacilitates tumor delineation and allows assessment of tumormotion to inform margin selection,20 (iii) the potential tomanage tumor motion in a reproducible fashion, e.g., usingabdominal compression,2 (iv) precise beam shaping and theuse of multiple nonopposing, coplanar and noncoplanartreatment beams to deliver a highly conformal, robust dosedistribution with rapid high-dose fall-off, (v) accurate andprecise dose-calculation algorithms, and (vi) on-line image-

guidance to accurately verify tumor and OAR location priorto and during treatment.21

Clinical RequirementsWe considered that an SBRT program required (i) a

clinical forum for patient evaluation and discussion, (ii) arobust quality assurance program, (iii) protocols for treatmentplanning and delivery, (iv) an integrated clinical team withdesignated roles, and (v) consideration of whether to developthe SBRT program within the context of a research ethicsboard-approved multicenter, or institutional protocol, and ifnot, to then put in place adequate independent mechanismsfor patient follow up that is required to ascertain tumorcontrol and toxicity and validate specific techniques.22

To meet these requirements there are several key stepsin the lung SBRT process. These include patient selection/assessment, treatment simulation, tumor and OAR delinea-tion, treatment planning, quality assurance, treatment deliv-ery, and assessment of outcomes. Figure 1 presents anoverview of the clinical system that was implemented atPMH to support the delivery of lung SBRT. Each of thecomponent parts is considered in more detail in the Results

FIGURE 1. An overview of the Lung Stereotactic BodyRadiation Therapy (SBRT) process at Princess MargaretHospital (PMH).

Journal of Thoracic Oncology • Volume 3, Number 11, November 2008 Practical Considerations from the Implementation of Lung SBRT

Copyright © 2008 by the International Association for the Study of Lung Cancer 1333

section, with specific examples of issues that we have en-countered.

RESULTS

Patient Assessment for Lung SBRTPatients’ individual circumstances will vary and multi-

disciplinary discussion is suggested, with the aim of ensuringthat patients receive the treatment with the best risk/benefitratio. There is some retrospective data from Japan reporting thatmedically operable patients with stage I NSCLC who receivedlung SBRT to a biologically effective dose of �100 Gy had animpressive 5-year overall survival rate of 71%.6 But atpresent, most patients with early-stage NSCLC are treatedwith surgical resection and SBRT is typically reserved forpatients who are medically inoperable. It is worth noting thatalthough there is general agreement on what factors contrib-ute to medical inoperability, the criteria are neither absolutenor firmly established and their application is a matter ofclinical judgment that may well differ between individualsand centers. If SBRT is being considered for medicallyoperable patients then it needs to be discussed in the contextof other treatments that include “standard procedure” surgery(e.g., lobectomy), lung-sparing surgery (e.g., segmentectomy,the role of which is being examined by randomized trials) andconventional RT. In common with many surgical series, we donot require histologic confirmation in all cases before SBRT. Insome patients the appearance of a new or growing lesion withincreased uptake on positron emission tomography (PET) im-aging will be deemed sufficient for decision-making. We recog-nize that individual centers will vary in this regard.

The critical appraisal of a tumor for possible SBRT ismore stringent than for conventional RT and we have foundit useful to consider several specific issues when assessingtreatment options, including (i) tumor location: for SBRT,ideal lesions are away from central thoracic structures, in-cluding proximal bronchial tree, esophagus and spinal cordand other radio-sensitive structures such as stomach, bowel,heart, brachial plexus, liver, and spleen, (ii) tumor size:although SBRT protocols have generally allowed for tumorsup to 5 cm maximum dimension, most results are based onsmaller lesions. The median size of the tumors that we havetreated is 2 to 3 cm, (iii) ease of tumor identification onimaging: ideally the tumor edges should be well visualized oncomputed tomography (CT) and not obscured by atelectasis,although if this is the case, fusion of PET images for treat-ment planning may be useful, (iv) histology: there is someconcern that bronchoalveolar cancers have a pattern of in-traalveolar microscopic spread that may not be ideally suitedto the tight treatment planning margins used in lung SBRT,and (v) number of lesions: we do consider treating multiplelesions, either in patients who have more than one primarylung cancer, or oligometastatic disease (see below).

Important patient factors to assess include the ability tolie in the treatment position for an appropriate length of time.Patients need sufficient flexibility to maintain the arms in anelevated position, but where this is not possible, it may befeasible to treat in an alternative position, such as one arm upand one arm down. Depending on its location, prior radiation

treatment need not preclude SBRT. Assessing cumulativedosing does, however, pose a challenge and each case mustbe carefully considered by the clinical team-there are no hardand fast rules.

Distant failure remains a significant problem in medi-cally inoperable patients treated with SBRT. In terms ofstaging patients our present practice is therefore to use mag-netic resonance imaging (or where contraindicated, CT) ofthe brain, CT of the thorax and abdomen, and Fluorodeoxy-glucose (FDG) PET/CT scan; an isotope bone scan is ob-tained if PET/CT is not performed. Every patient has pulmo-nary function testing, although we do not specify lower limitsthat would preclude SBRT. In practice, treatment fields areoften small, minimizing the amount of lung damage from RTand so even patients with extremely limited lung function,including those on home oxygen, may be candidates forSBRT, particularly if they have a peripheral lung lesion.

Treatment SimulationTwo issues of critical importance at the time of simu-

lation for lung SBRT are patient immobilization, and assess-ment of tumor motion.

Patient ImmobilizationReproducible and stable patient positioning is essential

to facilitate accurate treatment and to permit the small mar-gins typical of SBRT treatment planning. Many differentimmobilization devices exist: most common among these arethose incorporating a stereotactic frame23 and those that relyon evacuated bags. We use the latter (Vac-Lok MEDTEC,Orange City, IA).

Cranial, as well as extracranial SBRT was initiallydeveloped using a stereotactic frame to localize the target andguide the beams in three dimensions. This rigid devicesurrounds the patient and is often combined with an evacu-ated bag to immobilize the patient and abdominal compres-sion to reduce tumor motion. An important function is toenable the spatial location of the patient and tumor to bedescribed by an external coordinate system to improve theaccuracy and precision of patient positioning and contributeto reducing planning margins. Nevertheless by themselvessuch frames have been shown to be insufficient for extracra-nial tumor localization, which still requires the use of someform of image-guidance. With the frameless image-guidedsystem that we use, the tumor itself becomes the fiducial.Initial setup is achieved by the use of traditional external skinmarks and cone-beam CT is then used to correctly repositionthe tumor to the planning location. All three cardinal planes areassessed and on-line image registration is carried out before andduring treatment. Even though a frame is not used the paradigmof accuracy, precision and three-dimensional tumor localiza-tion, hallmarks of stereotactic treatment, prevails.

It is apparent that the ability of simulator units toaccommodate immobilization devices and specific patientpositions will be influenced by the dimensions of the machinebore and patient factors such as size, anatomy (e.g., kyphosis)and upper limb flexibility. Compared with conventional RT,each SBRT treatment can take a long time (e.g., 30–60minutes) and more effort is needed to ensure that the patient

Dahele et al. Journal of Thoracic Oncology • Volume 3, Number 11, November 2008

Copyright © 2008 by the International Association for the Study of Lung Cancer1334

is as comfortable as possible to maximize their stability.Careful positioning in the immobilization device, supportingthe hands and shoulders, and in some patients, premedicationwith analgesia (e.g., to prevent shoulder pain) or an anxiolyticmay need to be considered.

Target MotionMaximizing the therapeutic ratio of SBRT is in part

predicated on maintaining small treatment volumes. Thismakes limiting breathing-induced tumor motion important.24

There are several approaches to accounting for motion whenplanning lung SBRT. These range from using fixed cranio-caudal and lateral margins, to characterizing individual tumormotion and incorporating this into the definition of volumes.This can be achieved with methods such as fluoroscopy,‘slow-CT’ or respiratory-sorted (‘4D’) CT. With the lattertechnique the combination of maximum-exhale and inhalefree-breathing images (as at PMH), or a maximum intensityprojection can be used to denote the extent of tumor motion.

Assessing tumor motion in real-time, means that mo-tion management can be instituted at the simulator if it isexcessive (our action level is typically �1 cm in any direc-tion). Again, there is a choice when it comes to reducingmotion. Our preference is abdominal compression, a simpleand relatively well-tolerated dampening technique that is alsoused with frame-based SBRT. Motion is reassessed at thesame simulator session (aiming for �0.5–1 cm). Other op-tions include the Active Breathing Coordinator that effec-tively suspends motion and respiratory gating which limitsthe ‘beam-on’ time to certain phases of the breathing cycle.Although mobile, there are data to suggest that tumor move-ment is often fairly limited and relatively stable. For example,Franks et al. from PMH evaluated tumor motion using 4DCTand respiratory-sorted (4D) cone beam computed tomography(CBCT) and found that in 18 patients 87% of the motion in anydirection on 4D CBCT was �4.5 mm. Eighty-nine percent ofthe differences between the 4D treatment planning CT andCBCT motion were within �2 mm.25 Fewer than 25% of thepatients treated at PMH with lung SBRT have required abdom-inal compression.

Tumor and OAR DelineationSBRT is a high-precision technique requiring precise

delineation of both tumor(s) and normal structures. Oneapproach is described below.

Gross Tumor Volume (GTV)Using 4DCT imaging, the GTV is delineated on the

exhale 4DCT dataset using lung windows and then again onthe inhale dataset. Any solid portion of the tumor is includedin the GTV contour and surrounding areas of opacification orhaziness are assessed individually. In selected patients intra-venous CT contrast may help to identify the GTV. Smallvolumes and tight margins are desirable in SBRT to reducethe amount of normal tissue that is irradiated. When PETimaging is available (either in the diagnostic or preferably,the treatment position) it is fused to the exhale CT and maybe used to inform the contouring process, especially ininstances where there is a neighboring region of atelectasis.

Potential inaccuracy in the fusion process needs to be recog-nized and at this point in time the optimal integration of FDGPET into NSCLC delineation remains unresolved. The ex-hale-inhale GTV contours are fused in preparation for delin-eating the clinical target volume.

If respiratory-sorted CT images are unavailable, or ifthey are not usable, for example, if there is excessive artifact(e.g., ‘stepping’ in the images, particularly in the region of thetumor), or if the patient’s breathing pattern is unstable, thenfree-breathing helical images can be used for treatment plan-ning. If helical images are used, then additional informationabout tumor motion (e.g., from fluoroscopy) can still be usedto generate individualized planning margins, or alternativelya standard margin recipe (e.g., GTV- planning target volume[PTV] expansion of 10 mm cranio-caudal and 5 mm else-where) may be used. Although we typically obtain FDG PETimages at treatment planning and fuse these with the planningCT, the PET is not respiratory-correlated or ‘deblurred’ and itis at present unclear how to view or segment free-breathingPET images to best reflect true motion.

Clinical Target VolumeHaving characterized the GTV contours a decision

needs to be made as to how much of a margin to add forsuspected microscopic disease.26 In a departure from conven-tional treatment planning, and in common with others, we donot at present add a margin for this (and so the clinical targetvolume � internal target volume, [ITV]). This is predicatedon the assumption that the margin added to generate the PTV,the very high doses being used in SBRT and the relativelyisotropic dose fall-off, in combination with image-guidanceshould result in any microscopic disease being adequatelyirradiated. Early data from our own and other centers that usethis approach show very high local control rates and supportthis practice, however, longer-term data are awaited.

Planning Target Volume (PTV)For the remaining uncertainty a setup margin is re-

quired. A uniform expansion of 5 mm is typically applied tothe 4DCT-based ITV to generate the PTV. In certain circum-stances, for example OAR proximity, this may be individu-alized. The margin takes into account such factors as thetolerance in the image-guidance process and the stability ofpatient setup, which should be verified by individual institutions.

OAR DelineationWith the availability of 4DCT imaging we delineate the

normal structures on the exhale dataset for two reasons.Firstly, the tumor tends to spend proportionately more timelocated near this phase of breathing and secondly, using thedose-volume histogram for lung at full-exhale is generallymore conservative in terms of estimating normal lung spar-ing. By more accurately identifying actual tumor motion andimproving the conformity of treatment, 4DCT may reduce theamount of normal tissue that is irradiated.27

The identification of critical structures is another areawhere there are important differences with conventional RT,both in the structures delineated and the accuracy required.

Journal of Thoracic Oncology • Volume 3, Number 11, November 2008 Practical Considerations from the Implementation of Lung SBRT

Copyright © 2008 by the International Association for the Study of Lung Cancer 1335

There is the potential for greater risks to normal tissues withSBRT. This requires the delineation of central structures(trachea, proximal bronchial tree, esophagus, heart, and spi-nal canal), the brachial plexus and other OAR (such asstomach, bowel, liver, and spleen). This means that a greaterappreciation of imaging anatomy is required. At PMH, thisentailed a learning curve and prompted peer-review of con-tours. A second Radiation Oncologist from the lung SBRTteam looks at them before treatment planning commences,and if necessary, they are revised or discussed further withinthe SBRT team. They are subsequently reviewed in multidis-ciplinary SBRT rounds and where necessary we will alsoconsult with radiologists. Structures far away from the pri-mary tumor that would not usually be considered an OAR inconventional treatment may be identified as important, espe-cially when noncoplanar beams are used. Hypo-fractionationmay also render tissues such as great vessels, trachea, prox-imal airways, and skin, important dose-limiting structures.

Radiation Treatment PlanningDose Prescription

Dose is prescribed to the prescription isodose which ischosen to ensure adequate PTV coverage (typically �95% ofthe PTV should receive at least the prescription dose and�99% of the PTV should be covered by at least 90% of theprescribed dose). The prescription isodose should be between60 and 90%, where the center of mass of the PTV isnormalized to 100%. Doses greater than 105% of the pre-scribed dose should be located inside the PTV where sub-stantial heterogeneity is allowed (for example doses up to 100Gy would be permitted if 60 Gy was prescribed to the 60%isodose). In some situations, such as when the tumor is nearthe chest wall, it is desirable to try and avoid ‘hot spots’ overcertain normal tissues, in this case the rib and intercostaltissues, which may be located inside the PTV. In terms ofdose prescriptions, we have two main schedules, each withtheir own normal tissue constraints:

a. Fifty-four or 60 Gy in 3 fractions for peripheral tumorsaway from critical central structures (outside the ‘zoneof the proximal bronchial tree’ as defined by the Radi-ation Therapy Oncology Group, [RTOG]11). For theseprescriptions our normal tissue constraints and indicesof plan quality are based on those used in prospectiveRTOG studies.19 These criteria represent ‘expert opin-ion’ that takes into account international experiencewith SBRT and are the best we have for such potentregimens. One area where a practice adjustment hasbeen made, however, concerns tumors against the chestwall. In this location rib fracture and chest wall pain arepotential side-effects of SBRT. At PMH T1 tumors inthis location are treated with 48 Gy in 4 fractions.Prospective evaluation will provide data on tumor con-trol and the relative incidence of toxicity with thisapproach as compared with historical experience.

b. Forty-eight Gy in four fractions for T1 tumors (�3 cm)close to OAR (e.g., against the chest wall). This regi-men, and the normal tissue tolerances are taken from theJapanese Clinical Oncology Group 0403 study.15 Once

again these constraints represent ‘expert opinion’ andthey are also being tested prospectively.

Such normal tissue criteria as referred to above repre-sent present day knowledge and by adopting these constraintsit will be possible to benchmark the outcome data from ourinstitution against prospective studies.

While selected tumors located within the zone of theproximal bronchial tree may be considered for treatment, thisis an evolving area and as noted, caution is required. Atpresent, at PMH for lesions within 2 cm of the proximalairways, consideration is given to treating selected patientswith extended fractionation � reduced total dose (e.g., 60 Gyin 8 fractions28 or 50 Gy in 10 fractions). Not all manuscriptsprovide details of, or define tumor location and normal tissueconstraints may not be provided. As so often, the ‘devil is inthe detail’ and even among central tumors, some will be moreintimately in contact with critical structures than others.Sometimes by attending meetings and engaging with expertpractitioners additional practical information can be gleaned.There are, however, some data on the treatment of centraltumors and the reader is directed to the following referencesfor example (although many of these describe small oruncertain numbers of patients and a wide range of doseschedules).6,9,11,28–32 The largest reported experience may befrom Lagerwaard et al.28 who used a regimen of 60 Gy in 8fractions for tumors ‘adjacent to the heart, hilus or mediasti-num.’ In their series of 206 patients, they report using thisschedule in 12%, apparently without significant toxicity (butfollow-up is currently short at a median of 12 months for thewhole group).

Location of Tumor and OARBecause SBRT plans typically use coplanar and non-

coplanar beams to permit conformal dose distributions andrapid dose fall-off, organs such as the stomach, bowel, liveror brachial plexus may be in the irradiated volume, even ifthey are far away from the primary tumor (Figure 2) and theirdose-volume histogram must be assessed. There is someguidance on the tolerance of these organs in the body SBRTliterature.33,34 The composite dose to the skin requires con-sideration and in situations such as a peripheral tumor, wherebeams are close together, or where there is a skin fold, thedose may be higher than expected, necessitating changes tothe beam angle or energy and tighter constraints.

Treatment BeamsIn designing an SBRT treatment the plan should have

high conformity, rapid isotropic dose fall-off and a compactintermediate to low dose volume. A high degree of confor-mity serves to reduce high-dose spillage beyond the PTV andcan be described by the ratio of the volume of the prescriptionisodose to the PTV. The compactness and shape of the dosefall off beyond this can be described by the ratio of thevolume of the isodose representing 50% of the prescribeddose (R50%) to the PTV and the maximum dose at a distanceof 2 cm from the PTV (D2cm). Published tables providedesirable objectives for these criteria.19 To achieve these, atotal of nine nonopposing treatment beams (e.g., seven co-

Dahele et al. Journal of Thoracic Oncology • Volume 3, Number 11, November 2008

Copyright © 2008 by the International Association for the Study of Lung Cancer1336

planar and two noncoplanar) are commonly used at PMH.Initial beam parameters have been scripted into the planningsystem but they can be modified. If the tumor is close to thespinal canal for example, an OAR expansion, representing theplanning OAR volume is used to account for variation inpatient positioning and motion, and anisotropic fall off can beengineered to reduce dose to the OAR. Where additionaltreatments may be considered in the future (e.g., for coexis-tent or future lesions) the potential for cumulative dosing toan OAR needs to be considered and beam arrangementschosen to minimize this (e.g., to keep the total spinal corddose as low as possible).

At our institution, open conformal beams are standardand intensity modulated radiation therapy (IMRT) is onlyused in individual cases (e.g., where OAR structures are veryclose to, or within the PTV). In general we feel that the useof open conformal fields to treat a mobile tumor in only a fewfractions provides a more robust treatment than one that isdelivered using IMRT. Therefore, we only use IMRT whenclinically mandated (e.g., to create subvolumes with differentdose levels inside the PTV). And then only a small number ofsegments are usually permitted to reduce the potential influ-ence of motion on delivered dose.

When evaluating the treatment plan, unexpectedly highindividual beam weightings should prompt further assess-ment of the. If a minimum field size is used to permit reliabledosimetry (we use 3 � 3 cm in conjunction with multileafcollimator leaves �5 mm) then when tumors are small, theeffective treatment aperture may be larger than the PTV. Asa result conformity/dose spillage constraints may not be met.

Multiple TargetsMultiple tumors may be treated concurrently or sequen-

tially. So far we have treated two lesions in six patients, threeconcurrently and three sequentially (the lesions having gen-erally been considered synchronous or metachronous primarytumors). Where a decision is taken to treat more than onelesion simultaneously and they are close together, we havetypically used a single isocenter. In this situation, consider-

ations include whether or not more forgiving ITV-PTV margins(i.e., �5 mm) will be required, because it may be difficult tomatch two lesions simultaneously to their respective locationsat the time of simulation. Although a single isocenter maysometimes be associated with reduced conformity and in-creased dose spillage, it does serve to limit time on thetreatment couch, which should improve patient stability.

Heterogenity Correction (HC)Although the density along a path through the chest

wall and into the lung to reach the tumor may vary apprecia-bly, for a tumor in the lung parenchyma, it might be supposedthat the loss of electronic equilibrium may be compensatedfor by the greater energy retained by photons passing throughlung tissue. Nevertheless as demonstrated by differencesbetween plans corrected, or not, for heterogeneity, this is notalways be the case. Nonetheless, the decision to use hetero-genity correction (HC) in SBRT plans may vary betweencenters. Although it is inherently attractive to account fordifferences in tissue density along the beam path and betterappreciate the delivered dose, users should be aware thatdifferent planning systems may handle heterogeneity differ-ently15 and at present there are no clinical data to suggest thatusing HC improves outcomes (indeed recent studies, includ-ing the RTOG 0236 were performed without HC). If HC isused, then it may affect the total (corrected) dose that isdeliverable and the utility of non-HC corrected, SBRT qualitymetrics, especially those relating to intermediate dose volumeand dose fall-off.35 Individual institutions may need to con-sider modifications to existing quality metrics. Our standardpractice at this point in time is to use HC and this isperformed with the Collapsed Cone Convolution Superposi-tion algorithm.

With HC, our experience has been that particularly forsmall, mobile tumors surrounded by lung parenchyma (suchthat the volume of the PTV is considerably greater than thatof the GTV, and the overall PTV density is low), it has oftennot been possible to deliver 60 Gy in 3 fractions and simul-

FIGURE 2. Multiple coplanar and noncoplanar treatment beams (A) are typically used in thoracic stereotactic body radiation ther-apy (SBRT) to achieve dose conformity and rapid high-dose fall-off (B, Green isodose � 60 Gy, Light Blue � 30 Gy, Orange � 18Gy). The use of noncoplanar beams means that additional organs at risk (OAR) structures such as brachial plexus, proximal airways,and liver must be contoured so that the radiation dose they will receive can be assessed (C).

Journal of Thoracic Oncology • Volume 3, Number 11, November 2008 Practical Considerations from the Implementation of Lung SBRT

Copyright © 2008 by the International Association for the Study of Lung Cancer 1337

taneously maintain acceptable dose conformity and spillage.In such cases though, 54 Gy in 3 fractions is often achievable.

Quality Assurance (QA) Including SBRT RoundsRegular SBRT rounds are a focal point for the multi-

professional, multidisciplinary team that delivers lung SBRT(e.g., radiation oncologists, physicists, planners, clinical re-search associates, and treatment therapists). They are seen asparticularly important given the evolving nature of lungSBRT and the relative immaturity of this technique. Theyprovide an opportunity to discuss patients, to review treat-ment plans, serve as a forum for wider technical consider-ations, education and research, to highlight specific clinicalscenarios and to improve clinical processes. It is useful tohave facilities available for the projection of treatment plans,patient records, diagnostic images, radiation treatmentrecords, and verification images from the treatment unit.

Because of its increased requirement for accuracy andprecision, and the lack of an opportunity to make up forinadequacies in the delivery of any given fraction, robustquality assurance (QA) is very important in SBRT.20,36 Thisincludes independent QA of patient immobilization, the4DCT simulator, linear accelerators, and image-guidancesystems,37,38 as well as tumor/OAR contour and treatmentplan review. If intensity modulated radiotherapy (IMRT) isbeing used then additional QA is required on the treatmentunit. A mini multileaf collimator improves treatment confor-mity, and a linear accelerator with high output will contributeto reducing overall treatment time. Although larger centersmay have several linear accelerators that are SBRT capable,machine down time may still pose challenges. We do notdeliver SBRT on units without volumetric cone-beam imag-ing, nor do we substitute alternative fractionations. On occa-sion we have had to modify the overall treatment time orscheduling because of machine maintenance. On this note, itis worth highlighting that there are weekly SBRT schedules.39

If a machine is down, most problems will be fixed rapidly, butif not, then adaptation of the treatment schedule should befeasible. Characterizing the stability of patient immobiliza-tion systems during SBRT treatment can assist in determiningthe adequacy of institutional planning margins.40 We havedemonstrated that tumor drift from its initial location in-creases after about half an hour from initial localization21 andtherefore use intrafraction CBCT to verify patient stability.

Treatment DeliveryTrial Set Up

A trial setup session at which no treatment is deliveredhas been introduced for all patients to provide reliable feed-back on patient immobilization, positioning and comfort,stability of the lesion (e.g., same size, no interval develop-ment of atelectasis or effusion, reproducible motion), con-touring and margin selection, and to ensure that the treatmentis deliverable as planned (e.g., no treatment unit and patientor equipment collisions). CBCT imaging allows the matchbetween CBCT and simulation CT to be verified and tumormotion reevaluated (see below). It is also a further opportu-nity to review the treatment plan and potential OARs and achance to discuss patient setup and image matching with the

treating therapists. The ability to review these aspects oftreatment off-line is useful as there is less pressure on theteam to make immediate decisions.

This dry run typically takes place 48 to 72 hours beforethe scheduled start of treatment to allow enough time for recon-touring and replanning in the event that problems are identified.In recognition of the variability in breathing motion, respiratory-sorted (4D) CBCT images are also reconstructed using in-housesoftware to further confirm the adequacy of the ITV and PTV,and the stability of tumor motion.41 Other techniques can also beused to reevaluate motion. In a review of the outcome of trialsetup for lung SBRT at our institution, we observed a require-ment for treatment adjustments in 20%, ITV/PTV adjustment in11% and in 4% of cases, not all of the gantry angles wereachievable, confirming the usefulness of this procedure.

Image-GuidanceCBCT image-guidance to the tumor is a more accurate

means of localization than using bony landmarks.21 Purdie etal. demonstrated that the mean difference between lesionlocalization derived from soft-tissue (tumor) and bony match-ing was 6.8 mm, and exceeded 13.9 mm 10% of the time.Therefore, relying on bony anatomy as a surrogate for thetumor can result in suboptimal localization and verification.

In response to data like these, our present practice is tolocalize the tumor using a two-step on-line matching process:firstly, automatic registration to bony anatomy (spine) whichpermits an assessment of rotation and identifies major inac-curacies in patient setup. Because we cannot at presentcorrect for rotation (some couch-systems will correct forthis), it is ‘zeroed-out’ prior to step 2: a soft-tissue match inwhich the average tumor image on CBCT is colocalized withthe ITV/PTV treatment planning contours. Adequate OARsparing also needs to be verified at this point. If rotation is �3to 5 degrees (or less if there is perceived to be too great animpact on tumor or OAR location) then the patient is re-moved from the treatment couch and setup again. Patientposition and tumor stability is verified during treatment (typ-ically before the couch-kick to deliver the final two, nonco-planar beams) to minimize intrafraction drift in PTV cover-age or OAR avoidance. Repeat CBCT imaging at the end oftreatment can provide further information on patient stabilityand margin adequacy.

When using volumetric image-guidance, the original le-sion and normal tissue contours, as well as contours derivedfrom specific isodoses, especially those representing the pre-scription dose and critical structure constraints, are exported tothe treatment unit so that they can be used during on-line imagematching to verify both tumor and OAR location (Figure 3).

Such frequent use of on-line CBCT during treatmenthas facilitated role extension in the SBRT team and radiationtherapists’ autonomously perform image-guidance at PMH.Deviation beyond specific thresholds leads to a radiationoncologist being contacted. In certain situations, for examplewhen the tumor is less visible on CBCT or close to criticalstructures, the radiation oncologist may be present for someor all of the treatment matching, although we recognize thatin some jurisdictions their presence may be mandatory at all

Dahele et al. Journal of Thoracic Oncology • Volume 3, Number 11, November 2008

Copyright © 2008 by the International Association for the Study of Lung Cancer1338

fractions. All couch shifts made during treatment are re-corded prospectively for future analysis and all projectionsand reconstructed CBCT images are stored. The standardprocess that we have instituted for on-line image-guidancepromotes uniformity in practice and improves workflow onthe treatment unit.

On-Treatment Patient ReviewUnexpected toxicity needs to be identified to avoid

compromising treatment in which up to one third of the totaldose may be delivered in a single fraction. In most availableseries’ low-level acute morbidity predominates and this alsoreflects our experience to date.6,42,43

Follow-Up and Outcomes AssessmentFollow-up is essential to fully characterize toxicity and

efficacy. While opportunities for biologic and anatomic im-aging are now available in many centers, the optimal imagingparadigm for post-SBRT response assessment has not beendefined. At present our protocol mandates FDG PET/CTbefore and 3 months after completing treatment. This helpsstaging prior to SBRT and allows early assessment of tumorand normal tissue metabolism in response to SBRT. Becauseimaging appearances post-SBRT may vary from conventionalRT, dialogue between radiologists and oncologists may berequired to best characterize posttreatment findings. As anexample, Takeda et al. recently reported the development offibrosis near to the primary tumor a year or more post-SBRTthat can be difficult to distinguish from tumor recurrence44

(Figure 4). The interpretation of post-SBRT imaging oftenbecomes clearer with sequential scans but in selected patientshistologic evaluation may need to be considered.

Potential late toxicities of SBRT that are not typicallyseen (or seen less frequently) in conventional RT include: (a)rib fracture which may be symptomatic or asymptomatic andonly identified on close scrutiny of follow-up imaging, (b)partial lung collapse due to central bronchial toxicity, whichshould generally be further assessed with bronchoscopy, (c)chronic chest wall pain which may be neuropathic, promptingconsideration of adjuvant analgesics and if necessary, theinvolvement of pain specialists, and (d) soft tissue chest wallmasses that on histologic examination are consistent withpost-SBRT fibrosis. Findings like these will depend in largepart where the tumor and PTV are located. Education andsupport are important in informing patients and colleagues,who may be unfamiliar with SBRT, of its potential sideeffects, especially since the apparent incidence of these mayevolve over time as experience with the technique increases,more patients are treated and follow up lengthens. In practice,we subject all significant treatment toxicity and suspicion oftreatment failure to multidisciplinary and peer review by the

FIGURE 3. Cone beam CT (CBCT) is used to register thetarget and organs at risk to the simulation position, aided byexported planning contours. Respiratory-sorted CBCT withinternal target volume (ITV) (red) and planning target vol-ume (PTV) (green) is pictured.

FIGURE 4. Normal tissue changes postlung stereotacticbody radiation therapy (SBRT) demonstrating left rib fractureand chest wall fibrosis-which was confirmed histologically (A)and parenchymal lung fibrosis (B).

Journal of Thoracic Oncology • Volume 3, Number 11, November 2008 Practical Considerations from the Implementation of Lung SBRT

Copyright © 2008 by the International Association for the Study of Lung Cancer 1339

lung SBRT team in case there is an indication to change someaspect of the clinical process. We have established a lungSBRT clinic to facilitate patient assessment and follow-up.

DISCUSSIONOutside of certain institutions, lung SBRT is a rela-

tively recent incarnation. Furthermore, different treatmentschedules, planning systems and delivery techniques exist.This argues for a coordinated approach to the deployment oflung SBRT and highlights the importance of gathering long-term outcome information and imparting the uncertainty in itstherapeutic ratio to potential patients. Although perhaps out-wardly similar to conventional lung RT, important differ-ences have been highlighted. Foremost among these is thatSBRT is inherently less forgiving of inaccuracy. Thereforeparticular attention must be paid to patient comfort andstability, tumor motion, tumor/OAR delineation during plan-ning, and localization during treatment. Additional checks inthe clinical process are also warranted, including peer reviewof contours, trial setup before treatment and SBRT rounds.The tumor and normal tissue response differs between SBRTand conventional RT. Coupled with the relatively limitedfollow-up data and the potential for long-term survival inearly-stage lung cancer, this means that unexpected toxicity ispossible and protracted follow up is advocated.

Based around the component parts of lung SBRT andillustrated with specific clinical scenarios, this paper hasdescribed certain features of current lung SBRT practice atour institution. It highlights the multistep nature of thistreatment technique and emphasizes the importance of asystematic approach to the design of clinical SBRT services.At PMH, key features of the SBRT process include multidis-ciplinary patient selection, peer-review of tumor and OARcontours, a trial set up/dry run before treatment, SBRTrounds, and a lung SBRT clinic. Each of these individualmeasures, as well as the overall clinical system is subject tocontinuous improvement. Managing the transition from clin-ical trials to a routine treatment technique is an important partof allowing patients to fully benefit from lung SBRT. Tech-nical evolution by itself is insufficient and robust clinicalsystems are essential, especially for complex treatment tech-niques. These systems are intended to standardize and en-hance patient care, maximizing therapeutic gains as well astreatment efficacy, efficiency, and safety. They are beingaugmented by informal SBRT networks that are developingamong interested specialists and by collaboration with indus-try partners. Benchmarking and the reporting and sharing ofpractical experiences and solutions are beneficial to success-fully achieving change and should be encouraged.45

ACKNOWLEDGMENTSSupported by Elekta Synergy Research Group, Elekta

Ltd, Crawley, United Kingdom Addie MacNaughton Chair inLung Cancer Radiotherapy, and Princess Margaret Hospital.

REFERENCES1. Leksell L. The stereotaxic method and radiosurgery of the brain. Acta

Chir Scand 1951;102:316–319.2. Blomgren H, Lax I, Naslund I, Svanstrom R. Stereotactic high dose

fraction radiation therapy of extracranial tumors using an accelerator.Clinical experience of the first thirty-one patients. Acta Oncol 1995;34:861–870.

3. Uematsu M, Shioda A, Tahara K, et al. Focal, high dose, and fraction-ated modified stereotactic radiation therapy for lung carcinoma patients:a preliminary experience. Cancer 1998;82:1062–1070.

4. Timmerman R, Papiez L, McGarry R, et al. Extracranial stereotacticradioablation: results of a phase I study in medically inoperable stage Inon-small cell lung cancer. Chest 2003;124:1946–1955.

5. Sibley GS, Jamieson TA, Marks LB, Anscher MS, Prosnitz LR. Radio-therapy alone for medically inoperable stage I non-small-cell lungcancer: the Duke experience. Int J Radiat Oncol Biol Phys 1998;40:149–154.

6. Onishi H, Shirato H, Nagata Y, et al. Hypofractionated stereotacticradiotherapy (HypoFXSRT) for stage I non-small cell lung cancer:updated results of 257 patients in a Japanese multi-institutional study.J Thorac Oncol 2007;2:S94–S100.

7. Hoyer M, Roed H, Traberg Hansen A, et al. Phase II study on stereo-tactic body radiotherapy of colorectal metastases. Acta Oncol 2006;45:823–830.

8. Hiraoka M, Matsuo Y, Nagata Y. Stereotactic body radiation therapy(SBRT) for early-stage lung cancer. Cancer Radiother 2007;11:32–35.

9. Senan S, Haasbeek NJ, Smit EF, Lagerwaard FJ. Stereotactic radiother-apy for centrally located early-stage lung tumors. J Clin Oncol 2007;25:464.

10. Fowler JF, Tome WA, Fenwick JD, Mehta MP. A challenge to tradi-tional radiation oncology. Int J Radiat Oncol Biol Phys 2004;60:1241–1256. Erratum in: Int J Radiat Oncol Biol Phys 2005;61:962.

11. Timmerman R, McGarry R, Yiannoutsos C, et al. Excessive toxicitywhen treating central tumors in a phase II study of stereotactic bodyradiation therapy for medically inoperable early-stage lung cancer.J Clin Oncol 2006;24:4833–4839.

12. Baumann P, Nyman J, Lax I, et al. Factors important for efficacy ofstereotactic body radiotherapy of medically inoperable stage I lungcancer. A retrospective analysis of patients treated in the Nordic coun-tries. Acta Oncol 2006;45:787–795.

13. Onimaru R, Shirato H, Shimizu S, et al. Tolerance of organs at riskin small-volume, hypofractionated, image-guided radiotherapy forprimary and metastatic lung cancers. Int J Radiat Oncol Biol Phys2003;56:126 –135.

14. Yamashita H, Nakagawa K, Nakamura N, et al. Exceptionally highincidence of symptomatic grade 2–5 radiation pneumonitis after stereo-tactic radiation therapy for lung tumors. Radiat Oncol 2007;2:21.

15. Matsuo Y, Takayama K, Nagata Y, et al. Interinstitutional variations inplanning for stereotactic body radiation therapy for lung cancer. Int JRadiat Oncol Biol Phys 2007;68:416–425.

16. Bradley J. Radiographic response and clinical toxicity following SBRTfor stage I lung cancer. J Thorac Oncol 2007;2:S118—S124.

17. Timmerman RD, Park C, Kavanagh BD. The North American experi-ence with stereotactic body radiation therapy in non-small cell lungcancer. J Thorac Oncol 2007;2:S101—S112.

18. Chang JY, Roth JA. Stereotactic body radiation therapy for stage Inon-small cell lung cancer. Thorac Surg Clin 2007;17:251–259.

19. Timmerman R, Galvin J, Michalski J, et al. Accreditation and qualityassurance for radiation therapy oncology group: multicenter clinicaltrials using stereotactic body radiation therapy in lung cancer. ActaOncol 2006;45:779–786.

20. Underberg RW, Lagerwaard FJ, Cuijpers JP, et al. Four-dimensional CTscans for treatment planning in stereotactic radiotherapy for stage I lungcancer. Int J Radiat Oncol Biol Phys 2004;60:1283–1290.

21. Purdie TG, Bissonnette JP, Franks K, et al. Cone-beam computedtomography for on-line image guidance of lung stereotactic radiother-apy: localization, verification, and intrafraction tumor position. Int JRadiat Oncol Biol Phys 2007;68:243–252.

22. Ball D, Withers HR. Stereotactic radiotherapy for stage I non-small-celllung cancer–the triumph of technology over biology? Nat Clin PractOncol 2007;4:614–615.

23. Murray B, Forster K, Timmerman R. Frame-based immobilization andtargeting for stereotactic body radiation therapy. Med Dosim 2007;32:86–91.

24. Purdie TG, Moseley DJ, Bissonnette JP, et al. Respiration correlated

Dahele et al. Journal of Thoracic Oncology • Volume 3, Number 11, November 2008

Copyright © 2008 by the International Association for the Study of Lung Cancer1340

cone-beam computed tomography and 4DCT for evaluating target mo-tion in stereotactic lung radiation therapy. Acta Oncol 2006;45:915–922.

25. Franks KN, Purdie TG, Bezjak A, et al. Quantifying Inter and Intra-Fraction Tumor Motion using Respiration-Correlated Cone Beam CT inLung Stereotactic Body Radiotherapy (SBRT). Int J Radiat Oncol BiolPhys 2007;69:S489–S490.

26. Grills IS, Fitch DL, Goldstein NS, et al. Clinicopathologic analysis ofmicroscopic extension in lung adenocarcinoma: defining clinical targetvolume for radiotherapy. Int J Radiat Oncol Biol Phys 2007;69:334–341.

27. Underberg RW, Lagerwaard FJ, Slotman BJ, et al. Benefit of respiration-gated stereotactic radiotherapy for stage I lung cancer: an analysis of4DCT datasets. Int J Radiat Oncol Biol Phys 2005;62:554–560.

28. Lagerwaard FJ, Haasbeek CJ, Smit EF, et al. Outcomes of risk-adaptedfractionated stereotactic radiotherapy for stage I non-small-cell lungcancer. Int J Radiat Oncol Biol Phys 2008;70:685–692.

29. Wulf J, Haedinger U, Oppitz U, et al. Stereotactic radiotherapy forprimary lung cancer and pulmonary metastases: a noninvasive treatmentapproach in medically inoperable patients. Int J Radiat Oncol Biol Phys2004;60:186–196.

30. Joyner M, Salter BJ, Papanikolaou N, et al. Stereotactic body radiationtherapy for centrally located lung lesions. Acta Oncol 2006;45:802–807.

31. Song DY, Benedict SH, Cardinale RM, et al. Stereotactic body radiationtherapy of lung tumors: preliminary experience using normal tissuecomplication probability-based dose limits. Am J Clin Oncol 2005;28:591–596.

32. Milano MT, Katz AW, Schell MC, Philip A, Okunieff P. DescriptiveAnalysis of Oligometastatic Lesions Treated with Curative-Intent Ste-reotactic Body Radiotherapy. Int J Radiat Oncol Biol Phys 2008; (Epubahead of print).

33. Dawson LA, Eccles C, Craig T. Individualized image guided iso-NTCPbased liver cancer SBRT. Acta Oncol 2006;45:856–864.

34. Kavanagh BD, Schefter TE, Cardenes HR, et al. Interim analysis of aprospective phase I/II trial of SBRT for liver metastases. Acta Oncol2006;45:848–855.

35. Chang D, Liu C, Dempsey JF, et al. Predicting changes in dose

distribution to tumor and normal tissue when correcting for heterogene-ity in radiotherapy for lung cancer. Am J Clin Oncol 2007;30:57–62.

36. Potters L, Steinberg M, Rose C, et al. American Society for TherapeuticRadiology and Oncology; American College of Radiology. AmericanSociety for Therapeutic Radiology and Oncology and American Collegeof Radiology practice guideline for the performance of stereotactic bodyradiation therapy. Int J Radiat Oncol Biol Phys 2004;60:1026–1032.

37. Bissonnette JP. Quality assurance of image-guidance technologies. Se-min Radiat Oncol 2007;17:278–286.

38. Sharpe MB, Moseley DJ, Purdie TG, et al. The stability of mechanicalcalibration for a kV cone beam computed tomography system integratedwith linear accelerator. Med Phys 2006;33:136–144.

39. Salazar OM, Sandhu TS, Lattin PB, et al. Once-weekly, high-dosestereotactic body radiotherapy for lung cancer: 6-Year analysis of 60early-stage, 42 locally advanced, and 7 metastatic lung cancers. Int JRadiat Oncol Biol Phys 2008; (Epub ahead of print).

40. Jin JY, Ajlouni M, Ryu S, et al. A technique of quantitatively monitoringboth respiratory and nonrespiratory motion in patients using externalbody markers. Med Phys 2007;34:2875–2881.

41. Hugo G, Vargas C, Liang J, et al. Changes in the respiratory patternduring radiotherapy for cancer in the lung. Radiother Oncol 2006;78:326–331.

42. Timmerman R, Abdulrahman R, Kavanagh BD, Meyer JL. Lung cancer:a model for implementing stereotactic body radiation therapy intopractice. Front Radiat Ther Oncol 2007;40:368–385.

43. Zimmermann FB, Geinitz H, Schill S, et al. Stereotactic hypofraction-ated radiotherapy in stage I (T1–2 N0 M0) non-small-cell lung cancer(NSCLC). Acta Oncol 2006;45:796–801.

44. Takeda A, Kunieda E, Takeda T, et al. Possible misinterpretation ofdemarcated solid patterns of radiation fibrosis on CT scans as tumorrecurrence in patients receiving hypofractionated stereotactic radiother-apy for lung cancer. Int J Radiat Oncol Biol Phys 2008;70:1057–1065.

45. Massoud MR, Nielsen GA, Nolan K, et al. (Eds.) A Framework forSpread: From Local Improvements to System-Wide Change. IHI Inno-vation Series white paper. Cambridge, MA: Institute for HealthcareImprovement; 2006.

Journal of Thoracic Oncology • Volume 3, Number 11, November 2008 Practical Considerations from the Implementation of Lung SBRT

Copyright © 2008 by the International Association for the Study of Lung Cancer 1341