Original Study

Incidental Mediastinal Dose Does Not ExplainLow Mediastinal Node Recurrence Rates in

Patients With Early-Stage NSCLC Treated WithStereotactic Body Radiotherapy

Jean-Claude M. Rwigema,1 Allen M. Chen,1 Pin-Chieh Wang,1 Jay M. Lee,2

Edward Garon,3 Percy Lee1

AbstractThis study assessed the impact of incidental mediastinal radiation dose in 46 patients with early stage NSCLCtreated with stereotactic body radiation therapy (SBRT). The majority of the patients received < 5 Gy to allmediastinal nodal stations. Thus, incidental mediastinal dose from lung SBRT does not directly explain the lowmediastinal recurrences in stage I NSCLC.Background: Patients with stage I nonesmall-cell lung cancer (NSCLC) treated with stereotactic body radiotherapy(SBRT) do not undergo a staging mediastinoscopy, yet reported mediastinal recurrence rates appear lower than inpatients undergoing surgical resection. We determined incidental SBRT doses to assess whether this could accountfor the low rates of recurrence. Patients and Methods: Between March 2009 and September 2012, we reviewedcases of patients with inoperable lung tumors (n ¼ 136) treated with SBRT at our institution. The SBRT regimen was54 Gy in 3 fractions with positron emission tomography/computed tomography (PET/CT) staging. Incidental doses tothe mediastinal lymph node stations (MLNSs), primary tumor control, locoregional (LR), distant control (DC), andoverall survival (OS) rates were determined. Results: Forty-six patients with stage I NSCLC met the inclusion criteria.The calculated median incidental SBRT dose to all MLNSs was < 5 Gy for the majority of patients (75%). At a medianfollow-up of 16.8 months (0.6-38.9 months), the 1- and 2-year primary tumor control, LR, OS, and DC rates were 100%and 95.5%, 97.4% and 81.7%, 88.1% and 81%, and 96.9% and 86.9%, respectively. Only 2 patients (4.9%) hadmediastinal recurrence, with incidental SBRT doses to MLNSs that were similar to the rest of patients (P > .05).Conclusion: Low mediastinal recurrence rates in stage I NSCLC treated with SBRT validates the omission of stagingmediastinoscopy. The low incidental dose to MLNSs does not seem to explain the low mediastinal recurrence in themajority of patients. Our findings also confirm that prophylactic radiation to the mediastinum is not necessary andsupport the hypothesis that local ablation of the primary lesion could indirectly affect subclinical nodal disease throughunknown mechanisms.

Clinical Lung Cancer, Vol. -, No. -, --- ª 2014 Elsevier Inc. All rights reserved.Keywords: Mediastinal recurrence, Mediastinoscopy, Nonesmall-cell lung cancer, PET/CT staging,

Stereotactic body radiotherapy

IntroductionSurgical resection is the standard therapy for stage I nonesmall-

lung cancer (NSCLC), with excellent local control at 90% and5-year overall survival (OS) in the range of 50% to 70%.1,2

This work was presented as an oral presentation at the 55th Annual Meeting of theAmerican Society for Radiation Oncology; September 22-25, 2013; Atlanta, GA

1Department of Radiation Oncology2Division of Thoracic Surgery3Department of Medical OncologyRonald Reagan UCLA Medical Center, David Geffen School of Medicine at UCLA,Los Angeles, CA

1525-7304/$ - see frontmatter ª 2014 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.cllc.2014.01.004

However, a substantial number of patients have compromisedpulmonary function or medical comorbidities, or both, that placethem at unacceptably high perioperative risk. For these patientswith medically inoperable disease, conventional fractionated

Submitted: Nov 30, 2013; Revised: Jan 19, 2014; Accepted: Jan 21, 2014

Address for correspondence: Percy Lee, MD, Department of Radiation Oncology,David Geffen School of Medicine at UCLA, 200 UCLA Medical Plaza, B265, LosAngeles, CA 90095Fax: 310-794-9795; e-mail contact: percylee@mednet.ucla.edu

Clinical Lung Cancer Month 2014 - 1

Impact of SBRT on Mediastinal Recurrence

2 - Cli

radiotherapy using standard external-beam techniques has histori-cally been offered as an alternative to surgery. However, conven-tional fractionated radiotherapy achieves inferior disease control andOS when compared with surgery, likely because of technical limi-tations in attaining adequate radiation doses to the primary tumorwithout causing significant toxicity.3,4

Stereotactic body radiotherapy (SBRT) has become an increas-ingly accepted strategy to improve primary tumor control withablative doses of radiation while minimizing radiation to sur-rounding normal organs with a highly precise delivery of radiation.5

A number of studies, including 2 multicenter phase II trials, re-ported similar local control and OS to those observed in surgicalseries for early-stage NSCLC.6,7 Additionally, the convenienceafforded by a short treatment course of SBRT in a few fractions withexceptional tolerability as well as lower overall cost has contributedto the increasing use of SBRT worldwide in patients with inoperablestage I NSCLC.8

For those fit enough to undergo surgery, staging mediastinoscopyidentifies subclinical nodal disease in > 20% of patients with stage INSCLC, leading to mediastinal node failure rates of up to 17%after resection.9-12 Conversely, the incidence of mediastinal noderecurrence among patients treated with SBRT, most of whom areclinically staged without mediastinoscopy, appears lower than thatreported after surgical resection of stage I NSCLC.13 In the currentstudy, we sought to determine whether incidental SBRT doses tothe mediastinum may account for the low rates of recurrence in acohort of patients with stage I NSCLC treated with SBRT.

Patients and MethodsPatient Selection and Treatment Characteristics

Between March 2009 and September 2012, all consecutive casesof patients with medically inoperable lung tumors (n ¼ 136) treatedwith SBRT at our institution were retrospectively reviewed. Thestudy was approved by our institutional review board, and retro-spectively acquired data were deidentified according to the HealthInsurance Portability and Accountability Act guidelines. Inclusioncriteria were patients with stage I NSCLC who were � 18 years ofage, had Karnofsky performance score � 70, and had peripheraltumors and small central tumors (� 3 cm). We defined centraltumors as those in contact with or in close proximity to mediastinalor hilar structures, whereas tumors that were completely surroundedby lung parenchyma at least 2 cm away from the proximal bronchusor mediastinal and hilar structures, or both, were characterized asperipheral tumors. Pathology reports were available and reviewed toconfirm NSCLC, and those with carcinoma in situ (n ¼ 29), lungmetastases (n ¼ 10), or carcinoma of other histologic type (n ¼ 12)were excluded. Patients with large central tumors requiring alter-native total dose and fractionation regimens (ie, other than 54 Gy in3 fractions) were excluded (n ¼ 39). All patients underwent posi-tron emission tomography/computed tomography (PET/CT) atinitial diagnosis. The maximum standardized uptake value rangedfrom 2.3 to 16.5 (median, 6.2) for the primary tumor, and thebaseline maximum standardized uptake value at the mediastinalregion ranged from 1.3 to 2.4 (median, 1.7).

All patients underwent a 4-dimensional CT (4DCT) simulationscan with intravenous contrast medium and a slice thickness of 1.5 to3 mm using a VacLoc immobilization device (Med-Tech, Orange

nical Lung Cancer Month 2014

City, IA) in the supine position. The 4DCT simulation scan wasfused with the original staging PET/CT using iPlan TreatmentPlanning System (BrainLAB, Munich Germany). An internal targetvolume was outlined on the maximum-intensity projection imagegenerated from the 4DCT scan with 8 phases of respiration. Aplanning target volume (PTV) was defined as the internal targetvolume with a margin of 3 mm in the transverse direction anda 6-mm margin in the longitudinal direction. We outlined criticalstructures, including the ipsilateral and contralateral lungs, trachea,and proximal bronchi, spinal cord, esophagus, heart, and, wheneverindicated, the ipsilateral brachial plexus and the rib in prox-imity to the primary tumor. Maximum point dose constraints fornormal structures according to the Radiation Therapy OncologyGroup (RTOG) 0618 study were spinal cord < 6 Gy per fraction,heart < 8 Gy per fraction, brachial plexus < 8 Gy per fraction, totallungs < 8 Gy per fraction, esophagus < 9 Gy per fraction, andbronchi/trachea < 8 Gy per fraction. The maximum point doseconstraint used for ribs derived from our clinical experience was set atV30 < 30 cc.

Patients were treated with 6-mV photons with a dose of 54 Gyin 3 fractions, which was prescribed to the edge of the PTV.Target coverage at 95% of the PTV was normalized to prescriptiondose, with heterogeneity within the tumor up to 15% to 40%.Treatment was delivered every other day, with 8 to 12 coplanar ornoncoplanar beams as necessary to meet dose constraints. Imageguidance was used daily to confirm the position of the target.Intensity-modulated radiotherapy was not routinely used unlessnormal tissue dose constraints were exceeded with conformal beams.A Monte Carlo treatment planning algorithm was used.14

Follow-Up and Statistical AnalysisPatients were followed with clinical examination and PET/CT 2

months after SBRT or CT every 3 months during the first 2 yearsand then every 6 months thereafter to assess treatment response(Response Evaluation in Solid Tumors15) and toxicity (NationalCancer Institute Common Toxicity Criteria, version 3.0). Duringfollow-up, patients whose computed tomographic scans presentedequivocal results or suspected tumor relapse required PET/CT toascertain disease status.

The primary objective of the study was to perform dosimetricevaluation of incidental SBRT doses to mediastinal lymph nodestations (MLNSs). First, MLNSs that are typically sampled in apreoperative staging mediastinoscopy were contoured on the 4DCTsimulation scan using a CT-based atlas proposed by the InternationalAssociation for the Study of Lung Cancer.16 MLNSs were defined asupper paratracheal (level 2 right [R], 2 left [L]), lower paratracheal(level 4R, 4L), subaortic (level 5), subcarinal (level 7), and hilarnodes (level 10R, 10L). Figure 1 illustrates an example of a patientwith contoured MLNSs. Incidental mean doses to the differentLNSs were then obtained using the Monte Carlo algorithm.

The secondary objective was to report clinical outcomes, includingmediastinal recurrence, primary tumor control, locoregional (LR)control, distant control (DC) and OS rates, which were estimatedusing Kaplan-Meier methods with SAS statistical software, version9.4 (SAS Institute Inc, Cary, NC), as well as treatment-relatedtoxicity. Recurrence in the mediastinum was defined as recur-rence in both ipsilateral and contralateral hilar, mediastinal, and

Figure 1 Contoured Mediastinal Lymph Node Stations (MLNSs)in a Patient With Stage IA NSCLC. (A) UpperParatracheal (Level 2 Right, Orange; Level 2 Left,Yellow). (B) Lower Paratracheal (Level 4 Right, Green;Level 4 Left, Blue), Subaortic (Level 5, Orange), andthe Peripheral Tumor (Internal Target Volume, Red;PTV, Purple). (C) Subcarinal (Level 7, Yellow) andHilar (Level 10 Left, Red; Level 10 Right, Light Green)

Table 1 Patient and Treatment Characteristics

Characteristic Patients, No. (%)Age

Median, years 80

Range 42-95

Sex

Male 19 (41.3)

Female 27 (58.7)

Stage

IA 40 (86.9)

IB 6 (13.1)

Histologic Type

Adenocarcinoma 35 (76.1)

Squamous cell carcinoma 4 (8.7)

Adenosquamous 2 (4.3)

Nonesmall-cell lung cancer, NOS 5 (10.9)

Tumor Location

Central 5 (10.9)

LUL 2 (4.3)

RUL 3 (6.6)

Peripheral 41 (89.1)

LLL 4 (8.7)

LUL 9 (19.6)

RML 3 (6.5)

RLL 11 (23.9)

RUL 14 (30.4)

Karnofsky Performance Score

Median 80

Range 70-100

Radiation Dose, Gy

Total 54

Per fraction 18

BED10 151.2

PFTs

Available 38 (82.6)

Median FEV1 64 (39-116)

Median DLCO 51 (29-113)

Not performeda 8 (17.4)

Abbreviations: BED10 ¼ biological effective dose when a/b ¼ 10; DLCO ¼ diffusing capacityfor carbon monoxide; FEV1 ¼ forced expiratory volume in 1 second; LLL ¼ left lower lobe;LUL ¼ left upper lobe; NOS ¼ not otherwise specified; PFTs ¼ pulmonary function tests;RLL ¼ right lower lobe; RML ¼ right middle lobe; RUL ¼ right upper lobe.aPatients without PFTs who were deemed inoperable because of other comorbidities.

Jean-Claude M. Rwigema et al

supraclavicular LNSs. Primary tumor failure was defined as an in-crease in the primary tumor diameter� 20%, in addition to the levelof suspicion for recurrence by the diagnostic radiologist, in whichpattern of growth and overall appearance are considered,17 and theclinical assessment by the treating radiation oncologist. LR recur-rence was defined as failure in the primary tumor, recurrence in thesame lobe, or ipsilateral hilum, lung, or mediastinum. Distant failurewas defined as recurrence in the contralateral mediastinum, lung, orother distant sites. A univariate Cox regression model was used to

correlate doses to MLNSs with nodal recurrence in which P < .05was considered statistically significant.

ResultsA total of 46 patients with stage I lung NSCLC met the inclusion

criteria. Detailed patient and treatment characteristics are shown inTable 1. Overall median incidental SBRT doses to MLNSs andMLNS volumes are summarized in Table 2. We showed a distri-bution of incidental SBRT doses to MLNSs in 5-Gy incrementaldose groups (Fig. 2) and observed that across all MLNSs, the

Clinical Lung Cancer Month 2014 - 3

Table2

SBRT

Dosesto

Mediastinal

Lymph

Node

Stations

(MLN

Ss)andCo

rrelationBe

tweenDo

sesandtheRisk

ofRe

currence

MLN

Ss

SBRT

Dose,M

edian(Range),Gy

MLN

SsVolume,

Mean±SD

(mL)

Univariate

CoxRe

gression

AllT

umors

(n[

46)

Perip

heralT

umors

(n[

41)

CentralT

umors

(n[

5)AllT

umors

(n[

46)

Perip

heralT

umors

(n[

41)

CentralT

umors

(n[

5)HR

95%

HRCI

PValue

2L0.6(0.2-29.3)

0.5(0.2-14)

8.1(5.1-29.3)

5.4�

4.2

5.1�

3.8

6.1�

4.6

1.10

0.79-1.52

.57

2R0.5(0.2-21.6)

0.5(0.2-15)

10.5

(5.1-21.6)

5.3�

4.0

4.9�

3.3

6.8�

7.5

1.08

0.92-1.28

.36

4L1.9(0.2-15.8)

1(0.2-15.8)

5.5(0.7-15.1)

6.5�

2.8

6.4�

2.8

7.2�

3.2

0.99

0.67-1.45

.94

4R1.9(0.2-15.4)

1.2(0.2-15.4)

8.4(4.4-12.1)

8.5�

4.1

8.5�

4.2

8.4�

3.3

1.04

0.75-1.44

.80

51.1(0.2-14.8)

0.8(0.2-14.8)

2.6(0.7-15.1)

5.7�

2.1

5.6�

2.1

5.8�

2.3

0.89

0.58-1.38

.61

70.7(0.2-13.7)

0.7(0.2-13.7)

0.7(0.5-1.2)

3.2�

1.4

3.2�

1.5

2.4�

0.6

0.74

0.26-2.1

.57

10L

0.8(0.3-21.2)

0.9(0.3-21.2)

0.5(0.4-1.6)

22.4

�7.9

22.9

�8.1

19.0

�5.7

1.01

0.75-1.36

.94

10R

1.2(0.2-26)

1.4(0.2-26)

0.7(0.6-3.9)

18.7

�6.1

18.9

�6.1

17.6

�6.5

0.68

0.19-2.43

.55

Abbreviations:CI

¼confidenceinterval;HR

¼hazard

ratio;L¼

left;

R¼

right;SBRT

¼stereotacticbody

radiotherapy;SD

¼standard

deviation.

Impact of SBRT on Mediastinal Recurrence

4 - Clinical Lung Cancer Month 2014

majority of patients (75%) received < 5 Gy. As expected, patientswith tumors that were close to the mediastinum had higher dosesto the mediastinum, as in 1 patient with a central left upper lobetumor who had a mean 29.3 Gy to the left upper paratrachealLNS (ie, 2L).

Only 2 patients (4.9%) had mediastinal node recurrences.Table 3 shows detailed characteristics of the 2 patients. One patienthad mediastinal recurrence in the ipsilateral lower paratracheal,ipsilateral hilar, and subcarinal stations. The second patient hadrecurrence in the aortopulmonary window node station. There wasno significant correlation between dose to nodal stations and nodalrecurrence on univariate Cox regression analysis (Table 2). Figure 3shows the location of mediastinal recurrences observed in the 2patients. A multivariate analysis of multiple variables, includingincidental SBRT doses and patient and primary tumor character-istics, to predict nodal failure status was not feasible with only2 patients with nodal recurrences.

At a median follow-up of 16.8 months (0.6-38.9 months), the 1-and 2-year primary tumor control, LR, OS, and DC rates were100% and 95.5%, 97.4% and 81.7%, 88.1% and 81%, and 96.9%and 86.9%, respectively (Fig. 4). Failure at the primary tumoroccurred in only 1 patient (2.2%) at 12.9 months from treatmentcompletion. LR failure occurred in 5 patients (10.9%) at a meaninterval of 16.2 � 9.7 months. Distant failure occurred in 3 patients(6.5%) at a mean interval of 20.3 � 11.6 months and of these, 1had a synchronous mediastinal node failure (Table 3). There were atotal of 6 deaths, of which 3 were noncancer related (ie, ischemiccardiomyopathy, amyotrophic lateral sclerosis, and end-stage renaldisease complications), and for the remaining 3 patients with un-known cause of death, time of death was ascertained through theSocial Security Death Index.18 In this study, we observed only1 incidence (2.2%) of toxicity in a 69-year-old man with a Kar-nofsky performance score of 70 and a history of chronic obstructivepulmonary disease in whom grade 2 radiation pneumonitis devel-oped 3 months from SBRT completion.

DiscussionThe clinical outcomes observed in the present study are

consistent with those from other SBRT series for stage I NSCLC.Interestingly, mediastinal recurrence rates after SBRT have beenobserved to be on average < 5%, albeit without LN sampling,and distant metastatic recurrence represents the most commonsite of failure.13 Baumann et al reported outcomes of a prospectivephase II trial in patients with medically inoperable stage I NSCLC(n ¼ 57) treated with SBRT; the regional relapse rate in patientswho also had distant metastases was 5%.7 Another SBRT phase IItrial by Timmerman et al reported a 3.6% (2 of 55) mediastinalrecurrence rate.6 A retrospective series of 676 patients by Senthiet al reported a 6.3% overall regional recurrence rate.19 In ourstudy, the mediastinal recurrence rate was 4.9% (2 of 46).Interestingly, our study had a 2.2% grade 2 pneumonitis rate,which is much lower than the 16.3% grade 3/4 pneumonitis rateobserved in the RTOG 0236 study.6 This finding may beexplained by tighter margins used in our study: we used 3-mmaxial and 6-mm craniocaudal gross tumor volume expansionscompared with 5 mm and 10 mm, respectively, used in theRTOG 0236 protocol.

Figure 2 The Distribution of Incidental Stereotactic Body Radiotherapy (SBRT) Doses to all Mediastinal Lymph Node Stations (MLNSs)are Shown in 5-Gy Incremental Dose Groups. MLNSs are Defined as Upper Paratracheal (Level 2 Left [L] and 2 Right [R]),Lower Paratracheal (Level 4R and 4L), Subaortic (Level 5), and Subcarinal (Level 7) and Hilar (Level 10L and 10R)

Jean-Claude M. Rwigema et al

Historical surgical series have shown that subclinical LN diseasecan be present in a significant proportion of patients presentingwith stage T1/T2 disease, thereby justifying the need for patho-logic LN evaluation at the time of surgery. Asamura et al reportedoutcomes of LN involvement and recurrence in 337 patients withstage T1 (94%) and stage T2 (6%) tumors who underwent lo-bectomy (97%) or pneumonectomy (3%) and lymphadenectomy.9

Of 305 patients with clinical N0 status, 68 (22.3%) were found tohave mediastinal and hilar LN involvement after mediastinoscopy.With a follow-up of at least 5 years, the main pattern of recurrencewas distant (68%), and among those with LR recurrence, therewas a 5.3% (1 of 213) incidence of mediastinal recurrence. In aphase III trial comparing postoperative radiotherapy to surgeryalone in completely resected stage I NSCLC, there was a 9.4%mediastinal recurrence in the control arm.10 Takizawa et al foundthat 14% of patients with clinical stage I peripheral NSCLC hadmediastinal LN metastasis.20 In the Lung Cancer Study Group,Ginsberg et al reported on 247 patients who underwent resectionfor stage I lung cancer and the LR recurrence was 6.4% (8 of 125)for lobectomy and 17.2% (21 of 122) for limited (sublobar)resection.11 LR recurrence was defined as ipsilateral lung ormediastinal recurrence. In another study by Feld et al, they

Table 3 Characteristics of Patients With Recurrence to MLNSs

Patient Age KPS Stage Histologic Type Tumor

1 80 80 IA Adenocarcinoma Periphe

2 84 90 IA Adenocarcinoma Periph

Abbreviations: KPS ¼ Karnofsky performance score; LUL ¼ left upper lobe; R ¼ right; MLNSs ¼ m

analyzed sites of recurrence in patients with resected stage INSCLC from a multicenter Lung Cancer Study Group trial andshowed a 7% mediastinal recurrence.12 Martini et al reported on598 patients who underwent resection for stage I NSCLC with anoverall recurrence of 27% (local or regional, 7%; systemic, 20%)in which lobectomy was performed in 85% and mediastinal LNdissection was performed in 94%.2 Regional recurrence wasdefined as clinically or radiologically manifested disease in themediastinum or supraclavicular nodes. Cerfolio et al reported on504 patients who underwent preoperative staging with both PET/CT and complete thoracic lymphadenectomy followed by resec-tion (80% lobectomy) for stage I NSCLC, with an overall recur-rence rate of 10.9% (local, 1.2%; regional, 1.4%; distant, 8.3%).21

However, they defined regional recurrence as failure in the ipsi-lateral mediastinum only, unlike in our study in which we definedmediastinal recurrence to include the ipsilateral and contralateralmediastinum as well as the supraclavicular node region. Moreover,it is important to note that mediastinal recurrence rates in theaforementioned surgical series may have been higher without apreoperative lymphadenectomy.

Our findings fail to support the hypothesis that incidental scatterdoses to the mediastinum may have a direct effect for sterilizing

LocationMLNSs of Recurrence(Incidental Dose)

Interval ToRecurrence

(mo)Other

Recurrence

ral RUL 4R (8 Gy), 10R (0.84 Gy),Level 7 (0.58 Gy)

19.3 None

eral LUL Level 5 (0.71 Gy) 17.4 Distant

ediastinal lymph node station levels; RUL ¼ right upper lobe.

Clinical Lung Cancer Month 2014 - 5

Figure 3 A Coronal Computed Tomographic Scan Showing theLocation of Mediastinal Recurrences in the Cohort.Red Represents a Patient With Nodal Failure at Level5. Blue Represents a Patient With Nodal Failures atLevels 4 Right (R), 7, and 10R

Impact of SBRT on Mediastinal Recurrence

6 - Cli

subclinical nodal disease. The observation that the majority of pa-tients received < 5 Gy to mediastinal LNSs suggests that the bio-logical equivalent dose may not be adequate to sterilize the cancercells in the mediastinum (Fig. 2). The 2 patients with mediastinalnode failure had incidental doses to the MLNSs that were compa-rable to the remainder of patients (P > .05). Our results areconsistent with findings reported by Martin et al in 38 patients;

Figure 4 Actuarial Kaplan-Meier Probability Curves for Local Prima(OS), and Distant Control (DC) for Consecutive Patients WStereotactic Body Radiotherapy (SBRT). The Number of Pa

nical Lung Cancer Month 2014

however, the authors performed a limited dosimetric evaluationwithout including subaortic and subcarinal node stations, and,importantly, they did not correlate incidental SBRT nodal dose withclinical outcomes as done in our study.22

It appears that these results favor an alternative scenario in whichan indirect effect could be implicated in targeting subclinicalnodal disease. For instance, some studies have suggested thatimmune-mediated mechanisms may explain the increased efficacyof ablative doses to the primary tumor in SBRT leading toimproved survival when compared with conventionally fraction-ated radiotherapy. Lee et al used an animal model to demonstratethat ablative radiotherapy alone generates a strong CD8 T-celledependent immunity through T-cell priming in draininglymphoid tissues to cause primary tumor reduction and relapse, aswell as eradication of metastasis in some settings.23 The oppositeeffect was observed with conventionally fractionated radiationtherapy.

Similarly, in a review on systemic effects of radiotherapy, For-menti and Demaria proposed that immune-modulating effects ofradiation are dose dependent, wherein high radiation doses are morelikely to activate an adaptive antitumor immune response ratherthan a tumorigenic effect at low doses.24 This antitumor immuneresponse is thought to be initiated when dying cancer cells in theprimary tumor produce tumor antigens that are picked up byactivated dendritic cells, which then migrate to regional drainingLNs where they interact with naive T cells to acquire effectorfunctions. The latter can infiltrate sites distant from the primarytumor to propagate antitumor activity, a process that is alsoamplified by T-cell release of interferon-g. Other preclinical studieshave also described abscopal effects of radiation (ie, effects at sites

ry Tumor Control, Locoregional (LR) Control, Overall Survivalith Stage I Nonesmall-lung Cell Cancer (NSCLC) Treated Withtients at Risk Is Shown on the Time Interval Axis of Each Curve

Jean-Claude M. Rwigema et al

away from the locally irradiated target), which can be either anti-tumorigenic or protumorigenic depending on factors such asimmunocompetence of the host and type of cancer.24 A few clinicalstudies have reported abscopal effects in different tumor types,including carcinomas, melanoma, and lymphoma. Thus, lowmediastinal node recurrences seen after SBRT as in our series maybe immune mediated.

Although limited by its retrospective nature, this study is clini-cally significant because the results emphasize the dual therapeuticadvantage of SBRT as a noninvasive treatment strategy in stage INSCLC. Specifically, SBRT can target the primary tumor withboth excellent local control and minimal toxicity and may preventprogression of microscopic (subclinical) MLN metastases given therelatively low incidence of mediastinal recurrence after SBRT.Because there was an absence of pathologic mediastinal staging, andPET/CT understages disease in up to 32% of patients with clinicalstage I NSCLC,21 one would expect high recurrence rates in themediastinum after SBRT. Yet, the observed low incidence ofmediastinal recurrence after SBRT with PET/CT staging onlylends further support to the potential therapeutic effect of SBRT onthe mediastinum. Thus, we conclude that prophylactic irradiationto the mediastinum is not necessary in medically inoperable pa-tients treated with SBRT without a staging mediastinoscopy.Additionally, because distant metastases represent a predominantpattern of disease recurrence after SBRT, and adjuvant conven-tional chemotherapy is thought to abrogate the radiation-mediatedimmunity, development of effective immunotherapy that canfurther amplify radiation-induced antitumor immune response mayenhance long-term survival outcomes in these patients.23 With thatin mind, our results set a precedent for an upcoming prospectivephase II trial at our institution to test the efficacy of dendritic cellvaccination in patients with stage I NSCLC treated with SBRT.25

Finally, a better understanding of underlying mechanisms forablative radiation therapy to control disease outside the primarytarget deserves further investigation. The radiobiological impact ofSBRT in lung cancer may have a regional effect beyond its localtherapeutic efficacy that may be advantageous compared with otherlocal therapies.

ConclusionThe rate of MLN recurrence after SBRT in clinical stage I

NSCLC is low compared with the observed incidence in surgicalseries after lobectomy. Our results suggest that with a negative resulton staging PET/CT in the mediastinum, a staging mediastinoscopycan be omitted in medically inoperable patients who are treatedwith SBRT. The observed incidental dose does not seem to explainthe low mediastinal recurrence seen in the majority of our patients.These results also suggest that prophylactic irradiation to themediastinum is not necessary and support the findings from otherstudies suggesting that local ablation of the primary tumor may havean indirect effect on subclinical micrometastatic mediastinal nodedisease through immune-mediated mechanisms that have a regionalantitumor efficacy. Additional studies are warranted to fully eluci-date the advantages of SBRT beyond its local therapeutic effect,which may contribute to further defining the role of SBRT in thetreatment of early-stage lung cancer beyond patients with medicallyinoperable disease.

Clinical Practice Points

� Mediastinal nodal recurrence rates, in patients with early stageNSCLC treated with stereotactic body radiation therapy (SBRT),are comparable or lower to those undergoing definitive surgery.

� Currently, there is no published data describing the correlationbetween mediastinal nodal recurrences to incidental radiationdoses to the mediastinum.

� We tested the hypothesis that incidental radiation doses to thesemediastinal node stations is sufficient to sterilize subclinicaldisease when treating a NSCLC patient with definitive SBRT.

� In the majority of the cases the total dose to each of the medi-astinal nodal stations was <5 Gy, which is unlikely to sterilizemicroscopic disease. Thus, these results suggest that alternatemechanisms for low mediastinal recurrence rates are likely.

� In the treatment of early stage NSCLC by definitive SBRT, thisstudy also validates the current practice that neither a stagingmediastinoscopy nor prophylactic mediastinal irradiation is re-quired when the mediastinum is negative by a staging PET/CT.

DisclosureThe authors have stated that they have no conflicts of interest.

References1. Chang MY, Sugarbaker DJ. Surgery for early stage non-small cell lung cancer.

Semin Surg Oncol 2003; 21:74-84.2. Martini N, Bains MS, Burt ME, et al. Incidence of local recurrence and second

primary tumors in resected stage I lung cancer. J Thorac Cardiovasc Surg 1995; 109:120-9.

3. Dosoretz DE, Galmarini D, Rubenstein JH, et al. Local control in medicallyinoperable lung cancer. An analysis of its importance in outcome and factorsdetermining the probability of tumor eradication. Int J Radiat Oncol Biol Phys1993; 27:507-16.

4. Qiao X, Tullgren O, Lax I, et al. The role of radiotherapy in treatment of stage Inon-small cell lung cancer. Lung Cancer 2003; 41:1-11.

5. Haasbeek CJ, Slotman BJ, Senan S. Radiotherapy for lung cancer: clinical impactof recent technical advances. Lung Cancer 2009; 64:1-8.

6. Timmerman R, Paulus R, Galvin J, et al. Stereotactic body radiation therapy forinoperable early stage lung cancer. JAMA 2010; 303:1070-6.

7. Baumann P, Nyman J, Hoyer M, et al. Outcome in a prospective phase II trial ofmedically inoperable stage I non-small-cell lung cancer patients treated with ste-reotactic body radiotherapy. J Clin Oncol 2009; 27:3290-6.

8. Grutters JP, Kessels AG, Pijls-Johannesma M, De Ruysscher D, Joore MA,Lambin P. Comparison of the effectiveness of radiotherapy with photons, protonsand carbon-ions for non-small cell lung cancer: a meta-analysis. Radiother Oncol2010; 95:32-40.

9. Asamura H, Nakayama H, Kondo H, et al. Lymph node involvement, recurrenceand prognosis in resected small peripheral non-small cell lung carcinomas: are thesecarcinomas candidates for video assisted lobectomy? J Thorac Cardiovasc Surg 1996;111:1125-34.

10. Trodella L, Granone P, Valente S, et al. Adjuvant radiotherapy in non-small celllung cancer with pathological stage I: definitive results of a phase III randomizedtrial. Radiother Oncol 2002; 62:11-9.

11. Ginsberg RJ, Rubinstein LV. Lung Cancer Study Group. Randomized trial oflobectomy versus limited resection for T1N0 non-small cell lung cancer. AnnThorac Surg 1995; 60:615-23.

12. Feld R, Rubinstein LV, Weisenberger, et al. Sites of recurrence in resected stage Inon-small cell lung cancer: a guide for future studies. J Clin Oncol 1984; 2:1352-8.

13. Stephans K. Stereotactic body radiotherapy for stage I non-small cell lung cancer.Cleve Clin J Med 2012; 79(suppl):eS26-31.

14. Zhuang T, Djemil T, Qi P, et al. Dose calculation differences between MonteCarlo and pencil beam depend on the tumor locations and volumes for lungstereotactic body radiation therapy. J Appl Clin Med Phys 2013; 14:4011.

15. Therasse P, Arbuck SG, Eisenhauer EA. New guidelines to evaluate the response totreatment in solid tumors. European Organization for Research and Treatment ofCancer, National Cancer Institute of the United States, National Cancer Instituteof Canada. Natl Cancer Inst 2000; 92:205-16.

16. Rusch VW, Asamura H, Watanabe H. The IASLC lung cancer staging project: aproposal for a new international lymph node map in the forthcoming seventhedition of the TNM classification for lung cancer. J Thorac Oncol 2009; 4:568-77.

17. Huang K, Senthi S, Palma DA, et al. High-risk CT features for detection of localrecurrence after stereotactic ablative radiotherapy for lung cancer. Radiother Oncol2013; 109:51-7.

Clinical Lung Cancer Month 2014 - 7

Impact of SBRT on Mediastinal Recurrence

8 - Cli

18. Social Security Death Index. Available at: http://www.deathindexes.com/ssdi.html.Accessed on February 8, 2014.

19. Senthi S, Lagerwaard FJ, Haasbeek CJ, et al. Patterns of disease recurrence afterstereotactic ablative radiotherapy for early stage non-small-cell lung cancer: aretrospective analysis. Lancet Oncol 2012; 13:802-9.

20. Takizawa T, Terashima M, Koike T, et al. Mediastinal lymph node metastasis inpatients with clinical stage I peripheral non-small-cell lung cancer. J Thorac Car-diovasc Surg 1997; 113:248-52.

21. Cerfolio RJ, Bryant AS. Survival of patients with true pathologic stage I non-smallcell lung cancer. Ann Thorac Surg 2009; 88:917-22.

nical Lung Cancer Month 2014

22. Martin KL, Gomez J, Nazareth DP, et al. Quantification of incidental mediastinaland hilar irradiation delivered during definitive stereotactic body radiation therapyfor peripheral non-small cell lung cancer. Med Dosim 2012; 37:182-5.

23. Lee Y, Auh SL, Wang Y, et al. Therapeutic effects of ablative radiation on localtumor require CD8þ T cells: changing strategies for cancer treatment. Blood 2009;114:589-95.

24. Formenti SC, Demaria S. Systemic effects of local radiotherapy. Lancet Oncol2009; 10:718-26.

25. Um SJ, Choi YJ, Shin HJ, et al. Phase I study of autologous dendritic cell tumorvaccine in patients with non-small cell lung cancer. Lung Cancer 2010; 70:188-94.